d/freebsd-skq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
df28e5cc57
freebsd-skq/release/Makefile

1320 lines
41 KiB
Makefile
Raw Normal View History

$Id$ -> $FreeBSD$
1999-08-28 01:35:59 +00:00
# $FreeBSD$
This is the very small start of moving all the make release related stuff from etc/Makefile to release/Makefile. Right now this has one target (hieriarchy:) that replaces distrib-dirs:.
1994-09-23 08:28:25 +00:00
#
Fixed synopsis. Mention that md(4) should be active (from release(7)). Yet, this has nothing to do with bsd.prog.mk.
2003-02-21 22:37:18 +00:00
# make release [BUILDNAME=somename] CHROOTDIR=/some/dir CVSROOT=/cvs/dir \
# [RELEASETAG=tag]
Actually make sure the make release line at the top isn't misleading. CVSROOT is mandatory. Also note what CVSROOT has to be set to. Found by: local make release hacking
2001-02-19 14:23:49 +00:00
#
# Where "/some/dir" is the pathname of a directory on a some filesystem with
# at least 1000MB of free space, "somename" is what you want the release to
# call itself, "/cvs/dir" is where our source repo resides and, optionally,
# which CVS "tag" name should be used when checking out the sources to build
# the release (default is HEAD).
New and better "make release" rules.
1995-02-25 22:08:48 +00:00
#
Fixed synopsis. Mention that md(4) should be active (from release(7)). Yet, this has nothing to do with bsd.prog.mk.
2003-02-21 22:37:18 +00:00
# Please note: the md(4) driver must be present in the kernel
s/loaded as a module/available as a module/g Submitted by: ru
2003-03-23 04:23:10 +00:00
# (either by being compiled in or available as a kld(4) module),
Crypto is now part of the "base" distribution, and the old "release.3" target no longer exists. Renumber the old release.[4-9] targets to release.[3-8]. Reminded by: ru
2004-08-06 08:49:51 +00:00
# otherwise the target 'release.8' and possibly others will fail.
Add a note, that vn driver has to be compiled into the kernel.
1996-11-14 14:23:53 +00:00
#
Remind developers to update release(7) if they make substantial changes to this file.
2002-03-17 07:34:12 +00:00
# Note: If you add options to this file, please keep release(7) updated!
#
Add rules for building doc back into releases, conditionalized on NODOC. Also clean up some of the comments at the top. It's 12 o'clock, and that is all.
1997-05-24 20:15:13 +00:00
# Set these, release builder!
Create the release name of SNAPs automagically.
1997-06-01 08:06:48 +00:00
#
# Fixed version:
To honor the birth of RELENG_7 bump HEAD to 8.0-CURRENT. Approved by: re (implicit)
2007-10-11 04:28:08 +00:00
#BUILDNAME=8.0-CURRENT
Create the release name of SNAPs automagically.
1997-06-01 08:06:48 +00:00
#
# Automatic SNAP versioning:
Use four-digit year in SNAP name. We're eighteen months from the next millenium, fer cryin' out loud...
1998-07-09 08:26:38 +00:00
DATE != date +%Y%m%d
To honor the birth of RELENG_7 bump HEAD to 8.0-CURRENT. Approved by: re (implicit)
2007-10-11 04:28:08 +00:00
BASE = 8.0
Yikes! Whomever put the non-overridable setting of BUILDNAME in here, don't do that!
1997-09-04 11:54:01 +00:00
BUILDNAME?=${BASE}-${DATE}-SNAP
Create the release name of SNAPs automagically.
1997-06-01 08:06:48 +00:00
#
Simply comment-out BUILDNAME and CHROOTDIR - they should be set at the time of make's invocation.
1996-05-16 03:20:52 +00:00
#CHROOTDIR=/junk/release
Use RELENG_3 as an example, not RELENG_2_2.
1999-05-23 20:01:20 +00:00
# If this is a -stable snapshot, then set
To honor the birth of RELENG_7 bump HEAD to 8.0-CURRENT. Approved by: re (implicit)
2007-10-11 04:28:08 +00:00
#RELEASETAG=RELENG_7
Add CVSARGS for cvs arguments (vs. cvs subcommand arguments). Allow the cvs command to be prefixed by using CVSPREFIX.
2002-11-02 06:00:06 +00:00
#
Get the correct set of cryptographically-infected binaries automagically from the (secure|kerberos5)/Makefile:[KS]PROG.
2003-07-24 17:28:52 +00:00
# To test a release with a source tree containing patches and
# other work. This tree will get copied instead of getting the
# src/ tree from a CVS checkout. For "rerelease", this will NOT
# be copied; cvs update will be used instead.
#EXTSRCDIR=/usr/src
#
Add EXTPORTSDIR like EXTSRCDIR and EXTDOCDIR. PR: misc/83287 Submitted by: Gabor Kovesdan MFC after: 1 week
2006-01-02 13:29:58 +00:00
# To use a checked-out ports collection directory instead of
# checking out from a local CVS repository, set this option.
#EXTPORTSDIR=/usr/ports
#
Add CVSARGS for cvs arguments (vs. cvs subcommand arguments). Allow the cvs command to be prefixed by using CVSPREFIX.
2002-11-02 06:00:06 +00:00
# To add other options to the CVS subcommands (co,up), set
Introduce CVSCMDARGS make variable to set command-line options for cvs. You may want set '-D data-spec' to this variable. PR: 31218 Submitted by: Alexandr Listopad <laa@laa.zp.ua> MFC after: 5 days (if re@ permits)
2002-04-27 14:23:32 +00:00
#CVSCMDARGS="-D '01/01/2002 00:00:00 UTC'"
Convert "RELENG_ver_RELEASE" to "RELEASE_ver" and use it to check out the ports and doc trees. There is no change if we are not building a release (i.e., we are not using a tag of the form RELENG_ver_RELEASE --e.g., a RELENG_ver tag used to build a snapshot). This should allow the ports and doc freezes during the release cycles to be somewhat shorter, as commits to those (non-branched) trees can resume as soon as the tags are laid down. Reviewed by: nik Approved by: jkh
2000-04-06 10:00:20 +00:00
#
Add CVSARGS for cvs arguments (vs. cvs subcommand arguments). Allow the cvs command to be prefixed by using CVSPREFIX.
2002-11-02 06:00:06 +00:00
# To add other options to the CVS command, set
#CVSARGS="-lfq"
#
# To prefix the cvs command
#CVSPREFIX="/usr/bin/time"
#
Get the correct set of cryptographically-infected binaries automagically from the (secure|kerberos5)/Makefile:[KS]PROG.
2003-07-24 17:28:52 +00:00
# Where the CVS repository is
#CVSROOT="/home/ncvs"
#
Convert "RELENG_ver_RELEASE" to "RELEASE_ver" and use it to check out the ports and doc trees. There is no change if we are not building a release (i.e., we are not using a tag of the form RELENG_ver_RELEASE --e.g., a RELENG_ver tag used to build a snapshot). This should allow the ports and doc freezes during the release cycles to be somewhat shorter, as commits to those (non-branched) trees can resume as soon as the tags are laid down. Reviewed by: nik Approved by: jkh
2000-04-06 10:00:20 +00:00
# Non-zero if ${RELEASETAG} is in the form "RELENG_ver_RELEASE"; we
# are building an official release. Otherwise, we are building for
# a branch.
.if defined(RELEASETAG)
Use quotes around ${RELEASETAG} to allow values like -r RELENG_4 -D "2001-07-17 12:00"
2001-07-18 05:40:30 +00:00
ISRELEASE!= expr "${RELEASETAG}" : '^RELENG_.*_RELEASE$$' || true
Convert "RELENG_ver_RELEASE" to "RELEASE_ver" and use it to check out the ports and doc trees. There is no change if we are not building a release (i.e., we are not using a tag of the form RELENG_ver_RELEASE --e.g., a RELENG_ver tag used to build a snapshot). This should allow the ports and doc freezes during the release cycles to be somewhat shorter, as commits to those (non-branched) trees can resume as soon as the tags are laid down. Reviewed by: nik Approved by: jkh
2000-04-06 10:00:20 +00:00
.if ${ISRELEASE} != 0
# Convert "RELENG_ver_RELEASE" to "RELEASE_ver" for ports and doc trees.
AUXRELEASETAG!= echo ${RELEASETAG} | sed -e 's/^RELENG_/RELEASE_/' -e 's/_RELEASE$$//'
Stop using AUXRELEASETAG for everything and use PORTSRELEASETAG and DOCRELEASETAG in the appropriate places instead. Submitted by: asami
2000-11-12 11:04:11 +00:00
DOCRELEASETAG?= ${AUXRELEASETAG}
PORTSRELEASETAG?= ${AUXRELEASETAG}
Convert "RELENG_ver_RELEASE" to "RELEASE_ver" and use it to check out the ports and doc trees. There is no change if we are not building a release (i.e., we are not using a tag of the form RELENG_ver_RELEASE --e.g., a RELENG_ver tag used to build a snapshot). This should allow the ports and doc freezes during the release cycles to be somewhat shorter, as commits to those (non-branched) trees can resume as soon as the tags are laid down. Reviewed by: nik Approved by: jkh
2000-04-06 10:00:20 +00:00
.endif
.endif
Add the stuff to munge the name of the release/snap/whatever we build.
1995-04-23 23:06:36 +00:00
- Don't use quotes for the MD_ROOT_SIZE option when generating the BOOTMFS kernel config file. - Add WORLD_FLAGS and KERNEL_FLAGS so you can build world and kernel with extra make options such as -jX to speed up release builds. - When building kernels, allow their hints to tag along, and use this to install hints onto the boot floppy. Since the boot floppy doesn't load loader.4th, we have to change device.hints to strip comments and change each line to an explicit set command.
2000-07-20 03:07:29 +00:00
# If you want to pass flags to the world build such as -j X, use
# WORLD_FLAGS. Similarly, you can specify make flags for kernel
# builds via KERNEL_FLAGS.
Bah, I go and get all preachy, and then commit from the wrong branch. Revert the use of '-j8'. Apologies to: ru
2003-07-26 06:47:40 +00:00
#WORLD_FLAGS=-j4
#KERNEL_FLAGS=-j4
- Don't use quotes for the MD_ROOT_SIZE option when generating the BOOTMFS kernel config file. - Add WORLD_FLAGS and KERNEL_FLAGS so you can build world and kernel with extra make options such as -jX to speed up release builds. - When building kernels, allow their hints to tag along, and use this to install hints onto the boot floppy. Since the boot floppy doesn't load loader.4th, we have to change device.hints to strip comments and change each line to an explicit set command.
2000-07-20 03:07:29 +00:00
Milestone #2 in cross-arch make releases. They now work! You need to set TARGET_ARCH and possibly TARGET, the same way you normally do it for a cross build(7). Renamed `distribworld' to a more natutal `distributeworld'. Put pwd_mkdb(8) under ${INSTALLTMP}; for `distributeworld'.
2002-04-30 11:50:01 +00:00
TARGET_ARCH?= ${MACHINE_ARCH}
.if ${TARGET_ARCH} == ${MACHINE_ARCH}
TARGET?= ${MACHINE}
.else
TARGET?= ${TARGET_ARCH}
.endif
Fixed style bugs related to parentheses in Makefile.inc1. Submitted by: bde Fixed nearby bug: propagate the root Makefile's idea of the appropriate "make" binary down to release/Makefile.
2003-07-04 14:27:06 +00:00
CROSSENV= TARGET_ARCH=${TARGET_ARCH} TARGET=${TARGET}
CROSSMAKE= ${MAKE} ${CROSSENV}
Make sure the first ${CHROOTDIR} installworld is done for the correct arch. This makes "make release TARGET_ARCH=foo" really work. (I only tested with "make rerelease" before to save some time and did not notice the problem.)
2002-05-14 13:45:50 +00:00
NATIVEMAKE= ${MAKE} TARGET_ARCH=${MACHINE_ARCH} TARGET=${MACHINE}
Milestone #2 in cross-arch make releases. They now work! You need to set TARGET_ARCH and possibly TARGET, the same way you normally do it for a cross build(7). Renamed `distribworld' to a more natutal `distributeworld'. Put pwd_mkdb(8) under ${INSTALLTMP}; for `distributeworld'.
2002-04-30 11:50:01 +00:00
Allow the CVS modules from which the source, doc and ports collections are checked out to be overridden. PR: misc/6309 Submitted by: "Greg A. Woods" <woods@zeus.leitch.com>
1998-04-15 18:28:27 +00:00
# If you are using a local CVS repository with components stored in
# non-standard modules, override these on the make commandline or
# in the environment.
RELEASESRCMODULE?= src
RELEASEDOCMODULE?= doc
RELEASEPORTSMODULE?= ports
Add infrastructure that allows to build docs and RELNOTESng without the need to also create the entire ports tree during a `make release'. The main motivation behind this is that handling the ports tree still takes a huge amount of time due to the large number of directories involved, even on modern disks. The solution is to establish a list of dependent ports that are minimally required. This list needs to be manually maintained in case the doc ports toolchain changes, and has thus been broken out into a separate file Makefile.inc.docports. (release/Makefile has gotten overly lengthy already anyway.) Discussed with: bmah, nik Reviewed by: bmah
2001-05-04 19:50:34 +00:00
# Uncomment this to disable the doc.1 target. Docs normally require
# the ports tree, so NOPORTS can be set together with NODOC in order
# to have neither ports or docs. If only NOPORTS is set to YES, but
# docs are still desired, the DOMINIMALDOCPORTS logic below will only
# install the ports that are minimally required for the docs. This is
# intended as a compromise, less disk space is required than for using
Fix a typo: s/und/and/ Not Reviewed by: re@
2002-07-03 18:04:10 +00:00
# the entire ports collection (and much less time due to the huge number
Add infrastructure that allows to build docs and RELNOTESng without the need to also create the entire ports tree during a `make release'. The main motivation behind this is that handling the ports tree still takes a huge amount of time due to the large number of directories involved, even on modern disks. The solution is to establish a list of dependent ports that are minimally required. This list needs to be manually maintained in case the doc ports toolchain changes, and has thus been broken out into a separate file Makefile.inc.docports. (release/Makefile has gotten overly lengthy already anyway.) Discussed with: bmah, nik Reviewed by: bmah
2001-05-04 19:50:34 +00:00
# of directories it would create), but still quite a bit as well as some
# CPU cycles (some of the programs are C++, and things like ghostscript
# belong to the required ports nevertheless).
#
Fold the functionality of NORELNOTES into NODOC. There is now only one Makefile variable to control the building/installation of both the DocProj documents and RELNOTESng. Suggested by: obrien, dd Tested by: John Hay <jhay@icomtek.csir.co.za>
2001-05-30 17:05:15 +00:00
# Setting this also disables building of release note documentation
# (RELNOTESng).
Turn docs back on - ache was very quick.
1999-08-05 01:28:00 +00:00
#NODOC= YES
Comment out NODOC and NOPORTS again. It got uncommented by mistake a few days ago.
1998-09-15 13:15:18 +00:00
#NOPORTS= YES
First commit of RELNOTESng, the rewrite of the *.TXT documentation files. src/release/doc/README has additional information. Reviewed by: -current, -doc
2001-04-27 21:28:02 +00:00
Document the following knobs: CD_EXTRA_BITS, EXTRA_SRC, HTTP_PROXY and FTP_PROXY. Removed bitrot (BOOT_CONFIG). PR: misc/71465 Submitted by: Jon Passki <cykyc@yahoo.com>
2004-09-07 14:26:02 +00:00
# When retrieving ports using the pkg_add -r method, set your proxies to these.
# src/release/${arch}/mkisoimages.sh can use these.
#HTTP_PROXY?= www.example.com
#FTP_PROXY?= ftp.example.com
What should be the last major changes to the ISO layout: - When a separate livefs ISO is used, change the name of the tree directory to be R/cdrom/livefs and change the name of the iso to livefs.iso rather than using disc2. Instead, disc2 is now always going to be the packages disc regardless of the presence of a separate livefs which makes things simpler. - Build a very simple disc2 tree under R/cdrom that just includes the cdrom.inf file that the packages need. Also, build a disc2.iso image in the iso.1 target. Disc 2's volume label is "FreeBSD_Packages". - Retire CD_EXTRA_BITS and replace it with CD_PACKAGE_TREE. CD_EXTRA_BITS was specific to disc1, but CD_PACKAGE_TREE instead should point to a directory that has disc1 and disc2 subdirectories that contain the extra bits for each image. - Rename the bootonly volume label to "FreeBSD_bootonly" and just hardcode the disc1 and livefs volume names and iso suffixes.
2005-03-23 15:01:54 +00:00
# When creating ISO images, point ${CD_PACKAGE_TREE} to a directory containing
# the package split by an earlier invocation of the 'package-split' target.
#CD_PACKAGE_TREE= /path/to/pkg
Document the following knobs: CD_EXTRA_BITS, EXTRA_SRC, HTTP_PROXY and FTP_PROXY. Removed bitrot (BOOT_CONFIG). PR: misc/71465 Submitted by: Jon Passki <cykyc@yahoo.com>
2004-09-07 14:26:02 +00:00
# Extra source tarballs; each argument is a pair of source dir and
# distribution name. The dist name should not exceed 7 characters
# (another "s" for "source" will be prepended).
#EXTRA_SRC= games/fortune fortune
Collapsed the export list, added CD_EXTRA_BITS to it.
2002-08-14 16:50:50 +00:00
# Modify this definition if you want the release notes
RELNOTESng is now the default for -CURRENT release builds. (Floppy images get TXT renderings only; CDROM and FTP areas get TXT and HTML.) Remove the old *.TXT release documentation files, as they have been subsumed into RELNOTESng. The new layout will greatly facilitate their maintenence and help keep them internally consistent.
2001-05-25 18:01:34 +00:00
# and other release documentation in a language other than English.
Collapsed the export list, added CD_EXTRA_BITS to it.
2002-08-14 16:50:50 +00:00
RELNOTES_LANG?= en_US.ISO8859-1
First commit of RELNOTESng, the rewrite of the *.TXT documentation files. src/release/doc/README has additional information. Reviewed by: -current, -doc
2001-04-27 21:28:02 +00:00
Add infrastructure that allows to build docs and RELNOTESng without the need to also create the entire ports tree during a `make release'. The main motivation behind this is that handling the ports tree still takes a huge amount of time due to the large number of directories involved, even on modern disks. The solution is to establish a list of dependent ports that are minimally required. This list needs to be manually maintained in case the doc ports toolchain changes, and has thus been broken out into a separate file Makefile.inc.docports. (release/Makefile has gotten overly lengthy already anyway.) Discussed with: bmah, nik Reviewed by: bmah
2001-05-04 19:50:34 +00:00
# As an alternative to installing the entire ports collection (which
# can take a huge amount of time, in particular on slower disks),
# setting ${MINIMALDOCPORTS} allows to install and build just those
# ports that are really required for getting the docs up & running.
.if defined(NOPORTS) && !defined(NODOC)
DOMINIMALDOCPORTS= YES
.include "Makefile.inc.docports"
Rework the checking out of the doc, src, and ports trees a bit to make it more manageable. - Add some helper variables (CVS_{SRC,DOC,PORTS}ARGS) to be used when using CVS to checkout files. We stick release tags in these helper variables if they are defined and then use only one cvs command instead of two cvs commands with an .ifdef to choose between them. - rm the old src/doc/ports directories as separate commands from the CVS comands so that the rm commands don't need to be duplicated. - Simplify the DOMINIMALDOCPORTS case by overriding RELEASEPORTSMODULE to be ${MINIMALDOCPORTS} thus removing yet another nearly-duplicate cvs command in an .ifdef. - Add support for grabbing src/ and doc/ from external directories specified via EXTSRCDIR and EXTDOCDIR instead of from CVS. The same is not done for ports/ quite yet as the DOMINIMALDOCPORTS case is a bit tricky. The rerelease target scripts have not been changed to use the helper variables yet, so there is still some room for improvement. Submitted by: kuriyama (4)
2002-11-15 22:03:25 +00:00
RELEASEPORTSMODULE= ${MINIMALDOCPORTS}
Add infrastructure that allows to build docs and RELNOTESng without the need to also create the entire ports tree during a `make release'. The main motivation behind this is that handling the ports tree still takes a huge amount of time due to the large number of directories involved, even on modern disks. The solution is to establish a list of dependent ports that are minimally required. This list needs to be manually maintained in case the doc ports toolchain changes, and has thus been broken out into a separate file Makefile.inc.docports. (release/Makefile has gotten overly lengthy already anyway.) Discussed with: bmah, nik Reviewed by: bmah
2001-05-04 19:50:34 +00:00
.endif
# Helper variable
.if defined(NOPORTS)
.if !defined(DOMINIMALDOCPORTS) || ${DOMINIMALDOCPORTS} != "YES"
NOPORTSATALL= YES
.endif
.endif
Fix-ups for a few problems reported with how ports are handled in the release: - run 'make index' as part of release build, INDEX* files in CVS no longer updated and likely to be removed from CVS soon - don't include README.html files in ports.tgz tarball because they cause cvsup problems for users later - put a copy of /usr/ports on disc2 that does include README.html files so users have access to them Reviewed by: ru (found one bug, several clean-up suggestions) Tested by: 'make release' MFC before: BETA6 build (this isn't likely to be tested in -current anyway...)
2004-09-24 02:10:06 +00:00
#
# Doing 'make index' in /usr/ports requires Perl.
MAKEINDEXPORTS= lang/perl5.8
ALLLANG is long dead, garbage collect it. Instead, make it possible to use DOC_LANG to specify which languages the doc/ is built for. Note: be aware that the DOC_LANG setting in /etc/make.conf will now take the effect on "make release". (This is probably the desired behavior anyway.) PR: docs/42924
2002-09-23 14:33:47 +00:00
# By default, documentation (Handbook, FAQ, etc.) is built for all
# the languages. To speed up building, set the DOC_LANG to just
# the languages you need. (The language for the release notes is
# controlled by the RELNOTES_LANG variable above.)
#DOC_LANG= en_US.ISO8859-1
Now that textproc/docproj works properly, use it instead of the long list of ports.
1998-12-18 05:13:01 +00:00
DOCPORTS= textproc/docproj
Change variable name from DOCDISTFILES to RELEASEDISTFILES. Respect DOCDISTFILES if defined. No impact should happen with this change. DOCDISTFILES is introduced by rev. 1.541, to point the directory that contains the distfiles for ports/textproc/docproj. It's good, but we also build ports/sysutils/mkisofs during the release (if MAKE_ISOS is defined). It is natural that cdrtools tarball is prefetched, then copy it under ${CHROOTDIR}/usr/ports/distfiles. As a result, DOCDISTFILES variable is somewhat abused for non-doc tarballs; *DOC*DISTFILES is no longger appropriate name for that. PR: 16843 PR: 34194 Submitted by: murray MFC after: 6 days
2002-02-03 07:00:37 +00:00
# Set this to wherever the distfiles required by release procedures.
.if defined(DOCDISTFILES)
# Respect DOCDISTFILES which is used before.
RELEASEDISTFILES?= ${DOCDISTFILES}
.else
RELEASEDISTFILES?= ${.CURDIR}/../../ports/distfiles
.endif
Add a new release variable, AUTO_KEYBOARD_DETECT. If you set it to 1, the -P flag is used on the boot floppies. If not, don't. The new default is to not use it.
2000-06-04 04:41:41 +00:00
# Set this to 1 if you want -P to be used for automatic keyboard detection
# on the boot floppy. WARNING: Breaks on some Athlon (K7) motherboards.
AUTO_KEYBOARD_DETECT?= 0
Fold the functionality of NORELNOTES into NODOC. There is now only one Makefile variable to control the building/installation of both the DocProj documents and RELNOTESng. Suggested by: obrien, dd Tested by: John Hay <jhay@icomtek.csir.co.za>
2001-05-30 17:05:15 +00:00
.if !defined(NODOC)
Reflect that the hardware notes is now an MI document.
2007-02-16 02:28:15 +00:00
DIST_DOCS_ARCH_INDEP= hardware readme relnotes errata
Don't copy the installation guide to the release media (i.e. floppies or ISO images). We retain the concept of MD release documentation for now, although it's fairly unlikely that we'll ever do this again. Approved by: re (blanket for installation guide removal)
2007-08-14 20:02:55 +00:00
DIST_DOCS_ARCH_DEP=
First commit of RELNOTESng, the rewrite of the *.TXT documentation files. src/release/doc/README has additional information. Reviewed by: -current, -doc
2001-04-27 21:28:02 +00:00
.endif
Use the new doc locations.
1998-11-03 03:31:22 +00:00
New and better "make release" rules.
1995-02-25 22:08:48 +00:00
# Things which without too much trouble can be considered variables
Change the name of the 'bin' distribution to 'base'. This is done since it contains much more than /bin, and also gets in the way when making a combined install+fixit CD. OK'ed by: jkh
2002-04-23 22:16:41 +00:00
# BASE_DISTS are special in that they get full /etc installation sets.
Change the way distributions are built such that they can be built either in the Makefile or by an external script. Move some of the existing stuff into scripts.
1997-06-21 15:41:55 +00:00
#
Provide 32-bit runtime support on amd64 as a separate distribution, lib32. Prodded by: obrien Nodded by: peter Approved by: re
2005-06-16 18:16:14 +00:00
OTHER_DISTS?= catpages manpages games proflibs dict info doc ${ARCH_DISTS}
Change the name of the 'bin' distribution to 'base'. This is done since it contains much more than /bin, and also gets in the way when making a combined install+fixit CD. OK'ed by: jkh
2002-04-23 22:16:41 +00:00
BASE_DISTS?= base
Provide 32-bit runtime support on amd64 as a separate distribution, lib32. Prodded by: obrien Nodded by: peter Approved by: re
2005-06-16 18:16:14 +00:00
.if ${TARGET_ARCH} == "amd64"
ARCH_DISTS?= lib32
.endif
Start the de-orbital burn of our previous FreeBSD version compat libs living in usr/src. We need to use them from ports to record dependencies. Discussed with: re(scottl)
2004-08-29 18:27:00 +00:00
DISTRIBUTIONS?= ${BASE_DISTS} ${OTHER_DISTS}
Dont install DES bits in cdrom area.
1995-03-17 04:54:13 +00:00
Revamp base system packaging of kernels to enable up/smp selection at runtime and to support distributing additional kernels: o remove kernel from the base tarball o add new kernel tarballs o build + package both SMP and GENERIC kernels when an <arch>/conf/SMP config file is present o add sysinstall support for multiple kernels o update sysinstall to probe for the number of cpus on a system and auto-select smp/up kernel accordingly o add a post-kernels install hook to fixup /boot/kernel o add -ldevinfo to boot crunch for sysinstall's cpu probing logic Notes: 1. On HEAD this code is not currently used because GENERIC kernels include SMP. This work is mainly intended for RELENG_6 where the GENERIC kernel is UP. If HEAD changes to match then just enable WITH_SMP in sysinstall/Makefile. 2. The cpu probing support is done with acpi and MPTable; this means some systems will require work for auto-detection to work. 3. The handling of /boot/kernel may need to be revisited; for now we rename one kernel at the last moment (SMP if installed, otherwise GENERIC). There are other, possibly better, approaches. Lots of help from ru, emaste, scottl, and jhb.
2006-03-08 18:02:32 +00:00
#
# Build and package both GENERIC and SMP kernels if the target
# has both configuration files. Otherwise only GENERIC is done.
#
Fix this to what I've actually submitted.
2006-03-09 17:55:08 +00:00
.if exists(${.CURDIR}/../sys/${TARGET}/conf/SMP)
fix handling of default kernels Reviewed by: ru
2006-03-11 18:59:16 +00:00
KERNELS_BASE?= GENERIC SMP
.else
KERNELS_BASE?= GENERIC
Revamp base system packaging of kernels to enable up/smp selection at runtime and to support distributing additional kernels: o remove kernel from the base tarball o add new kernel tarballs o build + package both SMP and GENERIC kernels when an <arch>/conf/SMP config file is present o add sysinstall support for multiple kernels o update sysinstall to probe for the number of cpus on a system and auto-select smp/up kernel accordingly o add a post-kernels install hook to fixup /boot/kernel o add -ldevinfo to boot crunch for sysinstall's cpu probing logic Notes: 1. On HEAD this code is not currently used because GENERIC kernels include SMP. This work is mainly intended for RELENG_6 where the GENERIC kernel is UP. If HEAD changes to match then just enable WITH_SMP in sysinstall/Makefile. 2. The cpu probing support is done with acpi and MPTable; this means some systems will require work for auto-detection to work. 3. The handling of /boot/kernel may need to be revisited; for now we rename one kernel at the last moment (SMP if installed, otherwise GENERIC). There are other, possibly better, approaches. Lots of help from ru, emaste, scottl, and jhb.
2006-03-08 18:02:32 +00:00
.endif
Thanks to the dynamic duo of John & David, ("meanwhile in the VM-cave"), we can use the vn-driver instead of physical floppies, thanks guys ! Add the fixit-floppy back now that kzip made it fit again. Don't make the gzip'ed *.flp files anymore.
1995-04-19 17:42:32 +00:00
# mountpoint for filesystems.
MNT= /mnt
A major cleanup. Move more over to a "pull" model than a "push" model, this will eventually allow us to add meaningfull dependencies. Add mount_{msdos,nfs,cd9660} and chroot to the fixit.flp.
1995-04-20 06:48:31 +00:00
Remove hard-coded constants being passed to doFS.sh and add in a 386/alpha configuration section.
1998-10-07 03:13:36 +00:00
# Various floppy image parameters.
#
Moved driver modules for some PCI NICs and PCCARD-only NICs to mfsroot.flp. http://people.freebsd.org/~hosokawa/driver-floppy/ for details. Reviewed by: current@FreeBSD.org
2000-10-31 07:39:07 +00:00
Follow 'MAKE_ISOS' in naming and s/FLOPPIES/MAKE_FLOPPIES/g Submitted by: ru
2004-08-16 08:56:50 +00:00
.undef MAKE_FLOPPIES
Milestone #2 in cross-arch make releases. They now work! You need to set TARGET_ARCH and possibly TARGET, the same way you normally do it for a cross build(7). Renamed `distribworld' to a more natutal `distributeworld'. Put pwd_mkdb(8) under ${INSTALLTMP}; for `distributeworld'.
2002-04-30 11:50:01 +00:00
.if ${TARGET_ARCH} == "i386"
Follow 'MAKE_ISOS' in naming and s/FLOPPIES/MAKE_FLOPPIES/g Submitted by: ru
2004-08-16 08:56:50 +00:00
MAKE_FLOPPIES= true
The i386 disc1 size wandered above what fits on 650Mb media which has been our target previously. So i386 joins the others that have switched over to a separate livefs iso image.
2007-11-16 13:14:27 +00:00
SEPARATE_LIVEFS=
Boot floppy for i386 overflowed during Jan. 2007 snapshot build, fix it before Feb. 2007 snapshot build...
2007-02-01 15:12:44 +00:00
SPLIT_MFSROOT=
Milestone #2 in cross-arch make releases. They now work! You need to set TARGET_ARCH and possibly TARGET, the same way you normally do it for a cross build(7). Renamed `distribworld' to a more natutal `distributeworld'. Put pwd_mkdb(8) under ${INSTALLTMP}; for `distributeworld'.
2002-04-30 11:50:01 +00:00
.if ${TARGET} == "pc98"
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
SMALLFLOPPYSIZE= 1200
SMALLFLOPPYSPLITSIZE= 1152
SMALLFLOPPYLABEL= fd1200
- Use a stock loader with forth support on the boot floppies. - Use the regular forth scripts including the stock device.hints file on the boot floppies. - gzip all the various text files on the boot floppy that are larger than one sector to save on space. - Generate a loader.conf to load the kernel, acpi.ko, and mfsroot making use of the 'module_before' variables to prompt for a floppy to be entered when required and use a stock loader.rc. This enables the beastie menu on i386. - PC98 has been changed to split the mfsroot for now so that the floppies fit. When the PC98 bootstrap is updated to work with ELF files (which needs to happen anyway), then SPLIT_MFSROOT can be turned off for PC98.
2004-02-03 22:05:55 +00:00
SPLIT_MFSROOT=
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
.endif
FLOPPYSIZE= 1440
FLOPPYSPLITSIZE= 1392
FLOPPYINODE= 40000
- Tweak order of arch fs options to stick all floppy options together and all MFS root options together and in the same order on all archs. - Drop support for using the no longer existing 2.88 floppy image for CD booting on i386 via the EMUL_BOOT variable. - Use a stock loader on the boot floppy rather than one without forth. - Force the gzip of the mfsroot in release.9 to succeed. This is useful when doing lots of rereleases when testing floppy changes. - If an acpi kernel module exists, stick it on the boot floppy and for now always load it after the kernel.
2004-01-29 19:01:58 +00:00
FLOPPYLABEL= fd1440
Use more rational inode sizes.
1998-10-14 10:58:12 +00:00
BOOTINODE= 80000
- Tweak order of arch fs options to stick all floppy options together and all MFS root options together and in the same order on all archs. - Drop support for using the no longer existing 2.88 floppy image for CD booting on i386 via the EMUL_BOOT variable. - Use a stock loader on the boot floppy rather than one without forth. - Force the gzip of the mfsroot in release.9 to succeed. This is useful when doing lots of rereleases when testing floppy changes. - If an acpi kernel module exists, stick it on the boot floppy and for now always load it after the kernel.
2004-01-29 19:01:58 +00:00
MFSSIZE= 4320
- Remove obsolete PC-card boot.flp hack. It was for making both PC-card boot.flp and plain boot.flp. - Clean up crunchgen related routine. - Add PC-98 support. TODO: o Documentation o Fix some messages for PC-98 o Decrease the size of fixit.flp to 1.2MB o I18N (See: http://www.jp.FreeBSD.org/BootAsia/index.html) No response from jkh
2000-06-05 13:17:23 +00:00
MFSINODE= 8000
Moved driver modules for some PCI NICs and PCCARD-only NICs to mfsroot.flp. http://people.freebsd.org/~hosokawa/driver-floppy/ for details. Reviewed by: current@FreeBSD.org
2000-10-31 07:39:07 +00:00
MFSLABEL= minimum3
Add support for building sparc64 releases. Reviewed by: murray
2002-10-16 05:30:56 +00:00
.elif ${TARGET_ARCH} == "sparc64"
Presumably fix sparc64 "make release". (The breakage was due to sunlabel(8) no longer being linked to disklabel(8).) Reviewed by: phk
2003-04-23 09:01:35 +00:00
DISKLABEL= sunlabel
Add support for building sparc64 releases. Reviewed by: murray
2002-10-16 05:30:56 +00:00
MFSSIZE= 4096
MFSINODE= 8192
MFSLABEL= auto
Added the new knob MINIROOT that, when set, causes the creation of the bootable UFS image, miniroot.ufs, fully identical to the contents of the "bootonly" CD-ROM. The image is made available on FTP as miniroot/miniroot.ufs.gz, for resemblance with NetBSD. This replaces the boot.flp image for sparc64, making the latter as coming with NO_FLOPPIES. Reviewed by: jhb Tested by: hrs
2003-08-08 21:00:57 +00:00
MINIROOT=
Add the necessary tweaks for FreeBSD/ia64 releases. Note that this is in no way final. A typical ia64 wart is that there are no boot blocks. Instead, we need to create disks with EFI partitions if we want auto boot to work. All this functionality is not present yet.
2002-11-04 00:50:01 +00:00
.elif ${TARGET_ARCH} == "ia64"
- Tweak order of arch fs options to stick all floppy options together and all MFS root options together and in the same order on all archs. - Drop support for using the no longer existing 2.88 floppy image for CD booting on i386 via the EMUL_BOOT variable. - Use a stock loader on the boot floppy rather than one without forth. - Force the gzip of the mfsroot in release.9 to succeed. This is useful when doing lots of rereleases when testing floppy changes. - If an acpi kernel module exists, stick it on the boot floppy and for now always load it after the kernel.
2004-01-29 19:01:58 +00:00
MFSSIZE= 8192
Add the necessary tweaks for FreeBSD/ia64 releases. Note that this is in no way final. A typical ia64 wart is that there are no boot blocks. Instead, we need to create disks with EFI partitions if we want auto boot to work. All this functionality is not present yet.
2002-11-04 00:50:01 +00:00
MFSINODE= 8192
MFSLABEL= auto
Always use a separate disc for the livefs on ia64 since the merged image doesn't fit on one disc. Tested by: marcel X-MFC after: 30 seconds
2005-03-04 20:37:20 +00:00
SEPARATE_LIVEFS=
Add amd64 support
2003-06-26 00:55:33 +00:00
.elif ${TARGET_ARCH} == "amd64"
Enable boot floppies for amd64. I just copied the needed variables over from i386. Tested by: Markus Trippelsdorf markus at trippelsdorf dot de Approved by: scottl MFC after: 1 week
2006-01-04 18:02:56 +00:00
MAKE_FLOPPIES= true
FLOPPYSIZE= 1440
FLOPPYSPLITSIZE= 1392
FLOPPYINODE= 40000
FLOPPYLABEL= fd1440
BOOTINODE= 80000
Add amd64 support
2003-06-26 00:55:33 +00:00
MFSSIZE= 4096
MFSINODE= 8192
MFSLABEL= auto
The amd64 monthly snapshots began overflowing disc1 (and that's without any packages...).
2007-05-10 12:04:29 +00:00
SEPARATE_LIVEFS=
Enable splitting up the MFSROOT floppy on amd64, the contents now overflow one floppy's worth of space.
2006-01-30 14:24:31 +00:00
SPLIT_MFSROOT=
PPC definitions required for 'make release' Reviewed by: ru Submitted by: ssouhlal
2004-08-18 11:11:46 +00:00
.elif ${TARGET_ARCH} == "powerpc"
DISKLABEL= ""
MFSSIZE= 4096
MFSINODE= 8192
MFSLABEL= auto
Make the liveCD separate ala ia64 - the combined live/install is too large.
2005-05-23 03:14:56 +00:00
SEPARATE_LIVEFS=
Step #2 of fixing the fixit floppy. Kill duplicates for programs that have been in the boot crunched image as well as on the fixit floppy (pwd, newfs, hostname, test). Our space is really too valuable to have them around there twice. I doubt pwd needs to be there at all since it's a builtin into sh(1) anyway (oh, and the same applies to test(1) IIRC), but heck, leave them by now. Use the new `fixit' target in MAKEDEV to create the /dev nodes on the floppy, instead of including the kitchensink... Finally, tune the values used for creating the floppy. I currently end up with Filesystem 1K-blocks Used Avail Capacity iused ifree %iused /dev/vnn0c 1363 1301 -47 104% 368 14 96% ...which is not quite ideal yet, but at least a working configuration again.
1999-12-14 22:41:09 +00:00
.endif
Bring back 'NO_FLOPPIES', but as a user external knob. The internal logic knob remains MAKE_FLOPPIES. Suggested by: ru
2004-08-16 10:25:43 +00:00
.if defined(NO_FLOPPIES)
.undef MAKE_FLOPPIES
.endif
Remove hard-coded constants being passed to doFS.sh and add in a 386/alpha configuration section.
1998-10-07 03:13:36 +00:00
disklabel(8) is spelled bsdlabel(8) or sunlabel(8) these days. Also make sure bsdlabel(8) (along with the disklabel(8) compat link) still appear on the fixit floppies of platforms that use it natively (alpha, i386, and pc98). Approved by: re (scottl)
2003-05-17 19:16:28 +00:00
.if exists(/sbin/bsdlabel)
DISKLABEL?= bsdlabel
.else
Presumably fix sparc64 "make release". (The breakage was due to sunlabel(8) no longer being linked to disklabel(8).) Reviewed by: phk
2003-04-23 09:01:35 +00:00
DISKLABEL?= disklabel
disklabel(8) is spelled bsdlabel(8) or sunlabel(8) these days. Also make sure bsdlabel(8) (along with the disklabel(8) compat link) still appear on the fixit floppies of platforms that use it natively (alpha, i386, and pc98). Approved by: re (scottl)
2003-05-17 19:16:28 +00:00
.endif
Presumably fix sparc64 "make release". (The breakage was due to sunlabel(8) no longer being linked to disklabel(8).) Reviewed by: phk
2003-04-23 09:01:35 +00:00
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
ZIPNSPLIT= gzip --no-name -9 -c | split -b 1392k -
Remove hard-coded constants being passed to doFS.sh and add in a 386/alpha configuration section.
1998-10-07 03:13:36 +00:00
Change the way distributions are built such that they can be built either in the Makefile or by an external script. Move some of the existing stuff into scripts.
1997-06-21 15:41:55 +00:00
# Things which may get you into trouble if you change them
New and better "make release" rules.
1995-02-25 22:08:48 +00:00
MTREEFILES= ${.CURDIR}/../etc/mtree
+ back out my `iso' target + allows one to set _R so one can more easily make an iso after the fact, in the directory `make release' was issued in + changes name from miniboot to miniinst since we make other bootable ISOs also (and we might confuse someone) Approved by: jkh
2001-08-20 21:22:01 +00:00
_R?= /R
Some minor cosmetics: make mkdir less chatty in the output and stop making /R magic.
1999-09-16 08:44:22 +00:00
RD= ${_R}/stage
Modify the release-building process so the srelease distribution doesn't get spammed by RELNOTESng files. Previously, we built the RELNOTESng files "in place" and then populated the media (e.g. floppy, cdrom, ftp) areas from the build directories. However, the distributions got rolled after the RELNOTESng build, and thus picked up the rendered RELNOTESng files. Now, we install the RELNOTESng documents to R/stage/release.doc, then clean out the original directories in src/release/doc. We use the renditions in the staging area to populate the media areas. PR: 33141 MFC after: 3 days (plus re approval)
2001-12-28 17:21:28 +00:00
RND= ${RD}/release.doc
Some minor cosmetics: make mkdir less chatty in the output and stop making /R magic.
1999-09-16 08:44:22 +00:00
FD= ${_R}/ftp
CD= ${_R}/cdrom
- Boot-only CD-ROM support (bootonly.iso). Submitted by: jhb, ru - Moved the creation of the ports distribution to release.7. - Call MFS root floppy mfsroot.flp even on a small PC98 disk. Submitted by: jhb Reviewed by: ru, jhb
2003-08-02 18:21:02 +00:00
CD_BOOT= ${CD}/bootonly
Sync this up with RELENG_2_2. Du-oh, i've totally spammed rev 1.278, and quite a number of smaller fixes from 2.2 were missing as well.
1997-04-12 21:55:53 +00:00
CD_DISC1= ${CD}/disc1
Bring in split-CDROM changes from 2.2
1997-03-12 02:31:12 +00:00
CD_DISC2= ${CD}/disc2
Move the documentation to its own separate disc to make room for more packages on disc2. This will also let users decide if they want to have a CD of the docs at all - unless they're disconnected from the net they will probably find the Web site more useful. Reviewed by: ru MFC after: 3 days
2006-11-16 23:09:35 +00:00
.if !defined(NODOC)
CD_DOCS= ${CD}/docs
.endif
What should be the last major changes to the ISO layout: - When a separate livefs ISO is used, change the name of the tree directory to be R/cdrom/livefs and change the name of the iso to livefs.iso rather than using disc2. Instead, disc2 is now always going to be the packages disc regardless of the presence of a separate livefs which makes things simpler. - Build a very simple disc2 tree under R/cdrom that just includes the cdrom.inf file that the packages need. Also, build a disc2.iso image in the iso.1 target. Disc 2's volume label is "FreeBSD_Packages". - Retire CD_EXTRA_BITS and replace it with CD_PACKAGE_TREE. CD_EXTRA_BITS was specific to disc1, but CD_PACKAGE_TREE instead should point to a directory that has disc1 and disc2 subdirectories that contain the extra bits for each image. - Rename the bootonly volume label to "FreeBSD_bootonly" and just hardcode the disc1 and livefs volume names and iso suffixes.
2005-03-23 15:01:54 +00:00
.if defined(SEPARATE_LIVEFS)
CD_LIVEFS= ${CD}/livefs
Rework how we layout the CD/DVD ISO images for releases. Note that this is only the first stage and does not yet handle packages (which will move to their own disc(s)). - By default, include the live filesystem on disc1, so that disc1 is now both an install disc and a fixit disc. The images can still be split into two separate discs by specifying SEPARATE_LIVEFS=yes. - Remove /usr/ports from the live filesystem disc as the merged images do not fit on even i386 otherwise. Also, the merged image already contains a tarball of the entire ports tree, so /usr/ports was a duplicate anyway. - Stop building port READMES for the release sinc they are no longer used (/usr/ports in the live filesystem was the last remaining user). - Split cdrom.1 into three targets. cdrom.1 builds the live filesystem image. cdrom.2 builds the disc1 image. Note that if SEPARATE_LIVEFS is not specified these both write to the same area. cdrom.3 builds the bootonly image and the UFS miniroot. - The various distribution directories are now in a <buildname> subdirectory of disc1 rather than being in the root so that the merged image's root is not quite as cluttered. - The disc1 image is now called disc1.iso and we no longer build a miniinst.iso image. If CD_EXTRA_BITS is defined, then the iso.1 target will simply include those bits in the disc1 image that it builds. MFC after: 1 day Requested by: re (scottl, etc.)
2005-03-02 22:46:07 +00:00
.else
CD_LIVEFS= ${CD_DISC1}
It turns out the ia64 architecture's disc2 overflows if we put the ports/ tree on it. This makes putting the ports/ tree on disc2 at all dependent on the NOPORTREADMES knob as well as the NOPORT knob (at the moment NOPORT may be set while NOPORTREADMES isn't, that should probably be revisited). And it hardcodes ia64 for NOPORTREADMES by request the ia64 release builder. While here really get rid of the temp file 'make index' leaves behind. Tested by: i386 and ia64 'make release' Reviewed by: marcel MFC after: 1 day
2004-09-30 22:59:47 +00:00
.endif
Rework how we layout the CD/DVD ISO images for releases. Note that this is only the first stage and does not yet handle packages (which will move to their own disc(s)). - By default, include the live filesystem on disc1, so that disc1 is now both an install disc and a fixit disc. The images can still be split into two separate discs by specifying SEPARATE_LIVEFS=yes. - Remove /usr/ports from the live filesystem disc as the merged images do not fit on even i386 otherwise. Also, the merged image already contains a tarball of the entire ports tree, so /usr/ports was a duplicate anyway. - Stop building port READMES for the release sinc they are no longer used (/usr/ports in the live filesystem was the last remaining user). - Split cdrom.1 into three targets. cdrom.1 builds the live filesystem image. cdrom.2 builds the disc1 image. Note that if SEPARATE_LIVEFS is not specified these both write to the same area. cdrom.3 builds the bootonly image and the UFS miniroot. - The various distribution directories are now in a <buildname> subdirectory of disc1 rather than being in the root so that the merged image's root is not quite as cluttered. - The disc1 image is now called disc1.iso and we no longer build a miniinst.iso image. If CD_EXTRA_BITS is defined, then the iso.1 target will simply include those bits in the disc1 image that it builds. MFC after: 1 day Requested by: re (scottl, etc.)
2005-03-02 22:46:07 +00:00
_MK?= ${CHROOTDIR}/mk
It turns out the ia64 architecture's disc2 overflows if we put the ports/ tree on it. This makes putting the ports/ tree on disc2 at all dependent on the NOPORTREADMES knob as well as the NOPORT knob (at the moment NOPORT may be set while NOPORTREADMES isn't, that should probably be revisited). And it hardcodes ia64 for NOPORTREADMES by request the ia64 release builder. While here really get rid of the temp file 'make index' leaves behind. Tested by: i386 and ia64 'make release' Reviewed by: marcel MFC after: 1 day
2004-09-30 22:59:47 +00:00
Collapse 2 settings of PATH into one (didn't notice the 1st one before).
1998-01-28 04:05:10 +00:00
# Where the bootstrap ports (see DOCPORTS) get installed.
LOCALDIR= /usr/local/bin
Added the new knob MINIROOT that, when set, causes the creation of the bootable UFS image, miniroot.ufs, fully identical to the contents of the "bootonly" CD-ROM. The image is made available on FTP as miniroot/miniroot.ufs.gz, for resemblance with NetBSD. This replaces the boot.flp image for sparc64, making the latter as coming with NO_FLOPPIES. Reviewed by: jhb Tested by: hrs
2003-08-08 21:00:57 +00:00
.if ${TARGET} != ${MACHINE} && ${DISKLABEL} == "bsdlabel"
Presumably fix sparc64 "make release". (The breakage was due to sunlabel(8) no longer being linked to disklabel(8).) Reviewed by: phk
2003-04-23 09:01:35 +00:00
DOFS_SH= ${.CURDIR}/scripts/doFS.sh ${DISKLABEL} ${TARGET}
o Introduce variable DOFS_SH to hold the path of the doFS.sh. o ia64: Set DOFS_SH to point to an alternate version. o ia64: Lower BIGBOOTSIZE from 120M (LS-120) to ~5M. o ia64: Set BIGBOOTLABEL to efi so that we can check for it in the alternate doFS.sh script. If the label is not efi, we can invoke the generic doFS.sh.
2002-11-05 18:06:01 +00:00
.else
Presumably fix sparc64 "make release". (The breakage was due to sunlabel(8) no longer being linked to disklabel(8).) Reviewed by: phk
2003-04-23 09:01:35 +00:00
DOFS_SH= ${.CURDIR}/scripts/doFS.sh ${DISKLABEL} ""
o Introduce variable DOFS_SH to hold the path of the doFS.sh. o ia64: Set DOFS_SH to point to an alternate version. o ia64: Lower BIGBOOTSIZE from 120M (LS-120) to ~5M. o ia64: Set BIGBOOTLABEL to efi so that we can check for it in the alternate doFS.sh script. If the label is not efi, we can invoke the generic doFS.sh.
2002-11-05 18:06:01 +00:00
.endif
- Unbreak Alpha snapshots (we were still attempting to build a no longer existing "fixit" crunched binary). - Simplify the CRUNCH_TARGETS variable's initialization. - Simplify the release.5 logic a bit by accounting for the fact that release/*_crunch.conf files were removed two years ago.
2003-08-06 08:09:40 +00:00
CRUNCH_TARGETS= boot
.if ${TARGET_ARCH} == "i386"
CRUNCH_TARGETS+=fixit
.if ${TARGET} == "pc98"
CRUNCH_TARGETS+=fixit-small
.endif
Add support for building sparc64 releases. Reviewed by: murray
2002-10-16 05:30:56 +00:00
.endif
Now cpio works for me. (it was Mostly confusion about when the chroot happened) Make vsystem crunch aware in an intelligent fashion. make the boot.flp target more specific (and faster).
1995-05-20 23:33:14 +00:00
Follow 'MAKE_ISOS' in naming and s/FLOPPIES/MAKE_FLOPPIES/g Submitted by: ru
2004-08-16 08:56:50 +00:00
.if defined(MAKE_FLOPPIES)
- Only build a fixit floppy on i386 and pc98. - Reenable boot floppies for Alpha.
2004-03-25 20:37:59 +00:00
EXTRAS= floppies.1
.if ${TARGET_ARCH} == "i386"
EXTRAS+= floppies.2
.endif
EXTRAS+= floppies.3
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
.endif
EXTRAS+= ftp.1
Allow a person to specify NOCDROM to skip the cdrom.1 target during a release build. PR: 45300 Submitted by: David Yeske <dyeske@yahoo.com>
2002-11-15 20:51:27 +00:00
.if !defined(NOCDROM)
Rework how we layout the CD/DVD ISO images for releases. Note that this is only the first stage and does not yet handle packages (which will move to their own disc(s)). - By default, include the live filesystem on disc1, so that disc1 is now both an install disc and a fixit disc. The images can still be split into two separate discs by specifying SEPARATE_LIVEFS=yes. - Remove /usr/ports from the live filesystem disc as the merged images do not fit on even i386 otherwise. Also, the merged image already contains a tarball of the entire ports tree, so /usr/ports was a duplicate anyway. - Stop building port READMES for the release sinc they are no longer used (/usr/ports in the live filesystem was the last remaining user). - Split cdrom.1 into three targets. cdrom.1 builds the live filesystem image. cdrom.2 builds the disc1 image. Note that if SEPARATE_LIVEFS is not specified these both write to the same area. cdrom.3 builds the bootonly image and the UFS miniroot. - The various distribution directories are now in a <buildname> subdirectory of disc1 rather than being in the root so that the merged image's root is not quite as cluttered. - The disc1 image is now called disc1.iso and we no longer build a miniinst.iso image. If CD_EXTRA_BITS is defined, then the iso.1 target will simply include those bits in the disc1 image that it builds. MFC after: 1 day Requested by: re (scottl, etc.)
2005-03-02 22:46:07 +00:00
EXTRAS+= cdrom.1 cdrom.2 cdrom.3
Add scripts and Makefile support for building ISO images automatically as part of the release if MAKE_ISOS is set. Will also build the first CD with packages (in addition to the "minimal" CD) if CD_EXTRA_BITS points to them. This probably need a bit more work to get fully useful, but it at least covers the basics for now.
2001-07-23 09:01:46 +00:00
.if defined(MAKE_ISOS)
EXTRAS+= iso.1
Move the setting of the BOOTABLE flag outside of .if MAKE_ISOS so that bootable CDs can still be created if the iso.1 target is called directly. PR: misc/47191 Submitted by: Joost Bekkers <joost@bps.jodocus.org> MFC after: 1 week
2003-05-05 05:29:10 +00:00
.endif
- Tweak order of arch fs options to stick all floppy options together and all MFS root options together and in the same order on all archs. - Drop support for using the no longer existing 2.88 floppy image for CD booting on i386 via the EMUL_BOOT variable. - Use a stock loader on the boot floppy rather than one without forth. - Force the gzip of the mfsroot in release.9 to succeed. This is useful when doing lots of rereleases when testing floppy changes. - If an acpi kernel module exists, stick it on the boot floppy and for now always load it after the kernel.
2004-01-29 19:01:58 +00:00
.if ${TARGET} == "pc98"
Make a bootable CD on pc98.
2006-05-16 14:22:37 +00:00
BOOTABLE="-G"
Added the (undocumented) EMUL_BOOT variable (for TARGET=i386 only) that causes the bootable ISO images to be created using the floppy emulation (the old method) as opposed to the new "cdboot" method. Only copy boot.flp to the 2nd CD-ROM if this variable is defined. Reviewed by: murray
2003-07-23 20:53:37 +00:00
.else
Set a BOOTABLE variable to "-b" if MACHINE is not pc98, and use it for mkisoimages.sh instead of "-b" directly. MFC after: 1 week
2001-11-25 16:43:45 +00:00
BOOTABLE="-b"
.endif
Add scripts and Makefile support for building ISO images automatically as part of the release if MAKE_ISOS is set. Will also build the first CD with packages (in addition to the "minimal" CD) if CD_EXTRA_BITS points to them. This probably need a bit more work to get fully useful, but it at least covers the basics for now.
2001-07-23 09:01:46 +00:00
.endif
The doc dist was being made too early - move it.
1997-06-22 17:21:59 +00:00
.if !defined(NODOC)
Fold the functionality of NORELNOTES into NODOC. There is now only one Makefile variable to control the building/installation of both the DocProj documents and RELNOTESng. Suggested by: obrien, dd Tested by: John Hay <jhay@icomtek.csir.co.za>
2001-05-30 17:05:15 +00:00
DOCREL= doc.1 doc.2
The doc dist was being made too early - move it.
1997-06-22 17:21:59 +00:00
.endif
Add a new package splitting mechanism to replace print-cdrom-packages.sh. The new system tries to be more automated so that there is less work for the re's to do. It also no longer uses a /usr/ports tree as its input, but uses the generated package build including its INDEX file as its input. It parses the INDEX file, determines which packages should go on which ISO images, and then builds full-fledged trees of packages that can be added as an argument to mkisofs along with the tree built by 'make release' to build a full CD image. The INDEX files in the populated trees are generated with volume media number to make use of sysinstall's multiple volume support so that the user is kindly prompted to insert the appropriate disc for a package if it is not on the current disc. There is still some more tweaking to be done here, but this part needs to be committed. This stuff will all be used to prep the 5.4 release as well. Tested by: kensmith, others on re@ Reviewed by: re
2005-03-21 19:17:48 +00:00
.if !make(release) && !make(rerelease) && !make(package-split)
Fixed style bugs related to parentheses in Makefile.inc1. Submitted by: bde Fixed nearby bug: propagate the root Makefile's idea of the appropriate "make" binary down to release/Makefile.
2003-07-04 14:27:06 +00:00
BINMAKE!= cd ${.CURDIR}/..; ${MAKE} -V BINMAKE
WMAKEENV!= cd ${.CURDIR}/..; \
${BINMAKE} ${CROSSENV} -f Makefile.inc1 -V WMAKEENV
WMAKE= ${WMAKEENV} ${BINMAKE}
Initialize the FIXCRYPTO, BINMAKE and related variables (that use the != operator) only when needed. This change allows me to check out the current version of release/ makefiles only (co -l) to /tmp/release, and use that directory to build a release (supplying the correct WORLDDIR). Without this, attempt to "make release" caused an endless fork bomb while trying to evaluate FIXCRYPTO, and the only way I could get away from this on a remote box was to "kill -INT 1", thanks to tcsh(1) and its internal "kill" command.
2003-07-30 14:33:32 +00:00
.endif
Milestone #1 in cross-arch make releases. Do not install games and profiled libraries to the ${CHROOTDIR} with the initial installworld. Eliminate the need in the second installworld. For that, make sure _everything_ is built in the "world" environment, using the right tool chain. Added SUBDIR_OVERRIDE helper stuff to Makefile.inc1. Split the buildworld process into stages, and skip some stages when SUBDIR_OVERRIDE is set (used to build crypto, krb4, and krb5 dists). Added NO_MAKEDB_RUN knob to Makefile.inc1 to avoid running makewhatis(1) at the end of installworld (used when making crypto, krb4, and krb5 dists). In release/scripts/doFS.sh, ensure that the correct boot blocks are used. Moved the creation of the "crypto" dist from release.5 to release.2. In release.3 and doMFSKERN, build kernels in the "world" environment. KERNELS now means "additional" kernels, GENERIC is always built. Ensure we build crunched binaries in the "world" environment. Obfuscate release/Makefile some more (WMAKEENV) to achieve this. Inline createBOOTMFS target. Use already built GENERIC kernel modules to augment mfsfd's /stand/modules. GC doMODULES as such. Assorted fixes: Get rid of the "afterdistribute" target by moving the single use of it from sys/Makefile to etc/Makefile's "distribute". Makefile.inc1: apparently "etc" no longer needs to be last for "distribute" to succeed. gnu/usr.bin/perl/library/Makefile.inc: do not override the "install" and "distribute" targets, do it the "canonical" way. release/scripts/{man,cat}pages-make.sh: make sure Perl manpages and catpages appear in the right dists. Note that because Perl does not respect the MANBUILDCAT (and NOMAN), this results in a loss of /usr/share/perl/man/cat* empty directories. This will be fixed soon. Turn MAKE_KERBEROS4 into a plain boolean variable (if it is set it means "make KerberosIV"), as documented in the make.conf(5) manpage. Most of the userland makefiles did not test it for "YES" anyway. XXX Should specialized kerberized libpam versions be included into the krb4 and krb5 dists? (libpam.a would be incorrect anyway if both krb4 and krb5 dists were choosen.) Make sure "games" dist is made before "catpages", otherwise games catpages settle in the wrong dist. Fast build machine provided by: Igor Kucherenko <kivvy@sunbay.com>
2002-04-26 17:55:27 +00:00
Rework the checking out of the doc, src, and ports trees a bit to make it more manageable. - Add some helper variables (CVS_{SRC,DOC,PORTS}ARGS) to be used when using CVS to checkout files. We stick release tags in these helper variables if they are defined and then use only one cvs command instead of two cvs commands with an .ifdef to choose between them. - rm the old src/doc/ports directories as separate commands from the CVS comands so that the rm commands don't need to be duplicated. - Simplify the DOMINIMALDOCPORTS case by overriding RELEASEPORTSMODULE to be ${MINIMALDOCPORTS} thus removing yet another nearly-duplicate cvs command in an .ifdef. - Add support for grabbing src/ and doc/ from external directories specified via EXTSRCDIR and EXTDOCDIR instead of from CVS. The same is not done for ports/ quite yet as the DOMINIMALDOCPORTS case is a bit tricky. The rerelease target scripts have not been changed to use the helper variables yet, so there is still some room for improvement. Submitted by: kuriyama (4)
2002-11-15 22:03:25 +00:00
CVS_SRCARGS= -P
.if defined(RELEASETAG)
CVS_SRCARGS+= -r ${RELEASETAG}
.endif
CVS_DOCARGS= -P
.if defined(DOCRELEASETAG)
CVS_DOCARGS+= -r ${DOCRELEASETAG}
.endif
CVS_PORTSARGS= -P
.if defined(PORTSRELEASETAG)
CVS_PORTSARGS+= -r ${PORTSRELEASETAG}
.endif
Added the WORLDDIR variable (defaulting to ${.CURDIR}/..) that points to a directory where "make buildworld" was run. Useful for building 5.x snapshots on 4.x.
2003-07-23 14:40:51 +00:00
WORLDDIR?= ${.CURDIR}/..
Fixed synopsis. Mention that md(4) should be active (from release(7)). Yet, this has nothing to do with bsd.prog.mk.
2003-02-21 22:37:18 +00:00
release rerelease:
Protect against CVSROOT not being set. Submitted by: fenner
1997-06-29 18:07:01 +00:00
.if !defined(CHROOTDIR) || !defined(BUILDNAME) || !defined(CVSROOT)
@echo "To make a release you must set CHROOTDIR, BUILDNAME and CVSROOT" && false
Various changes to make this work.
1995-02-26 01:35:32 +00:00
.endif
Add infrastructure that allows to build docs and RELNOTESng without the need to also create the entire ports tree during a `make release'. The main motivation behind this is that handling the ports tree still takes a huge amount of time due to the large number of directories involved, even on modern disks. The solution is to establish a list of dependent ports that are minimally required. This list needs to be manually maintained in case the doc ports toolchain changes, and has thus been broken out into a separate file Makefile.inc.docports. (release/Makefile has gotten overly lengthy already anyway.) Discussed with: bmah, nik Reviewed by: bmah
2001-05-04 19:50:34 +00:00
.if defined(NOPORTSATALL) && !defined(NODOC)
- Add in some sanity checking. If NOPORTS is set but NODOC is not set, print an error message and die immediately rather than an hour later.
2000-07-14 18:03:32 +00:00
@echo "Ports are required for building the docs. Either set NODOC or"
Add infrastructure that allows to build docs and RELNOTESng without the need to also create the entire ports tree during a `make release'. The main motivation behind this is that handling the ports tree still takes a huge amount of time due to the large number of directories involved, even on modern disks. The solution is to establish a list of dependent ports that are minimally required. This list needs to be manually maintained in case the doc ports toolchain changes, and has thus been broken out into a separate file Makefile.inc.docports. (release/Makefile has gotten overly lengthy already anyway.) Discussed with: bmah, nik Reviewed by: bmah
2001-05-04 19:50:34 +00:00
@echo "unset NOPORTS, or set at least DOMINIMALDOCPORTS to YES!"
- Add in some sanity checking. If NOPORTS is set but NODOC is not set, print an error message and die immediately rather than an hour later.
2000-07-14 18:03:32 +00:00
@exit 1
.endif
Remove some serious foot-shooting potential from the release Makefile. For some unknown reason, when LOCAL_PATCHES and LOCAL_SCRIPT were originally added, they were silently ignored if the actual file did not exist. As a result, if one mistyped the pathname to a patch or script, then the release silently succeeded. However, it was not built with the desired changes and no warning was given to inform the builder either. This commit explicitly checks to see that all of the defined patches and scripts exist up front and bails if any of them do not exist. I lost several hours of valuable sleeping time this evening due to this "feature" so I've finally gone and ripped out. I've tripped over this in the past several other times as well. Glanced at by: scottl
2004-07-15 04:52:57 +00:00
.if defined(LOCAL_PATCHES) && !empty(LOCAL_PATCHES)
.for p in ${LOCAL_PATCHES}
.if !exists(${p})
@echo "The patch file ${p} does not exist!"
@exit 1
.endif
.endfor
.endif
.if defined(LOCAL_SCRIPT) && !exists(${LOCAL_SCRIPT})
@echo "The local script ${LOCAL_SCRIPT} does not exist!"
@exit 1
.endif
Moved doFS into shell-script so we can be more smart about it. make two boot floppies, one without help files for 4MB systems. Mine's loading one now. :-) -O2 may be needed.
1996-01-08 22:13:50 +00:00
.if make(release)
Very minor rework to eliminate useage of - operator by checking for existance of ${CHROOTDIR}. Remove -k option from make install for building chroot tree, makes it error out early if you forgot to populate the pre-chroot obj tree. Reviewed by: phk
1995-05-13 19:51:57 +00:00
.if exists(${CHROOTDIR})
Speed up the cleaning of CHROOTDIR by almost a factor of two, by replacing two monstrous file tree walks with one monstrous tree walk and two miniscule ones.
1998-06-07 21:53:01 +00:00
# The first command will fail on a handful of files that have their schg
# flags set. But it greatly speeds up the next two commands.
Correct a typo in a comment. Submitted by: Xin LI
2004-09-08 06:05:19 +00:00
# NB: clear any vestigial devfs mount, just in case
Fixups to get make release going again for i386: o mount (and unmount) devfs in the chroot'd filesystem o handle umounting devfs in case of a keyboard interrupt o remove MAKEDEV from the fixit floppy so things fit again (mount_devfs was added in a separate commit) o explicitly remove mfsroot.gz and loader.conf when building the cd-rom area as otherwice cp -Rp aborts when a previous run left identical files around o increase the number of inodes on the drivers floppy; moving drivers from the kernel there (to shrink it's size) caused things to overflow o while we're increasing the driver's floppy inode config, separate out all parameters from the boot floppy Approved by: re
2002-10-19 16:44:16 +00:00
-umount ${CHROOTDIR}/dev > /dev/null 2>&1
Make trash removals less chatty.
1999-09-19 02:05:52 +00:00
-rm -rf ${CHROOTDIR} 2>/dev/null
Safer version of the chflags command.
2005-01-13 08:29:51 +00:00
-chflags -R 0 ${CHROOTDIR}/.
2.1.0R is over and out. Comment out the tags for it from the release Makefile (so release builds in -current require to set it from the command line). Make the cleanup of ${CHROOTDIR} non-fatal. Otherwise, it stops immediately if the ${CHROTDIR} is on a separately mounted file system.
1995-12-30 21:15:45 +00:00
-rm -rf ${CHROOTDIR}
Very minor rework to eliminate useage of - operator by checking for existance of ${CHROOTDIR}. Remove -k option from make install for building chroot tree, makes it error out early if you forgot to populate the pre-chroot obj tree. Reviewed by: phk
1995-05-13 19:51:57 +00:00
.endif
Some minor cosmetics: make mkdir less chatty in the output and stop making /R magic.
1999-09-16 08:44:22 +00:00
mkdir -p ${CHROOTDIR}
Change the ">>> make release ..." output to also include ${TARGET}.
2002-05-27 10:59:41 +00:00
@echo ">>> make release for ${TARGET} started on `LC_ALL=C TZ=GMT date`"
Convert NO_PROFILE and NO_LIB32 to new style.
2006-03-18 21:37:05 +00:00
cd ${WORLDDIR} && ${NATIVEMAKE} -DWITHOUT_GAMES -DWITHOUT_HTML -DWITHOUT_LIB32 \
-DWITHOUT_MAN -DWITHOUT_NLS -DWITHOUT_PROFILE installworld DESTDIR=${CHROOTDIR}
Use the shiny, new top-level 'distribution' target when populating the chroot area. This fixes make release of HEAD on systems that have a /usr/src with bsd.endian.mk and a matching /usr/obj but haven't installed the world in /usr/obj and thus have no bsd.endian.mk /usr/share/mk.
2005-03-02 14:57:45 +00:00
cd ${WORLDDIR} && ${NATIVEMAKE} distribution DESTDIR=${CHROOTDIR}
Make release would fail for lack of /etc/resolv.conf or $DISTFILES/ PR: misc/9913 Submitted by: YAMAMOTO Shigeru <shigeru@bremen.or.jp>
1999-03-31 11:18:52 +00:00
if [ -f /etc/resolv.conf ]; then \
cp -p /etc/resolv.conf ${CHROOTDIR}/etc; \
fi
Copy from ${EXTLOCALDIR} to ${CHROOTDIR}/usr/local if EXTLOCALDIR is defined.
2002-11-17 13:35:12 +00:00
.if defined(EXTLOCALDIR)
rm -rf ${CHROOTDIR}/usr/local
For EXT*DIR, use -H in addition to -R, so that a symlink may be used in the EXTSRCDIR case, for example, if one finds it useful to symlink /usr/src to an actual source tree, and use /usr/src as the EXTSRCDIR. This has no effects on release builds not using external directories. Reviewed by: ru Approved by: scottl (re)
2003-05-18 19:34:03 +00:00
cd ${CHROOTDIR}/usr && cp -R -H ${EXTLOCALDIR} local
Copy from ${EXTLOCALDIR} to ${CHROOTDIR}/usr/local if EXTLOCALDIR is defined.
2002-11-17 13:35:12 +00:00
.endif
Rework the checking out of the doc, src, and ports trees a bit to make it more manageable. - Add some helper variables (CVS_{SRC,DOC,PORTS}ARGS) to be used when using CVS to checkout files. We stick release tags in these helper variables if they are defined and then use only one cvs command instead of two cvs commands with an .ifdef to choose between them. - rm the old src/doc/ports directories as separate commands from the CVS comands so that the rm commands don't need to be duplicated. - Simplify the DOMINIMALDOCPORTS case by overriding RELEASEPORTSMODULE to be ${MINIMALDOCPORTS} thus removing yet another nearly-duplicate cvs command in an .ifdef. - Add support for grabbing src/ and doc/ from external directories specified via EXTSRCDIR and EXTDOCDIR instead of from CVS. The same is not done for ports/ quite yet as the DOMINIMALDOCPORTS case is a bit tricky. The rerelease target scripts have not been changed to use the helper variables yet, so there is still some room for improvement. Submitted by: kuriyama (4)
2002-11-15 22:03:25 +00:00
rm -rf ${CHROOTDIR}/usr/src
.if defined(EXTSRCDIR)
cd ${CHROOTDIR}/usr && \
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
cp -R -H ${EXTSRCDIR} src
New and better "make release" rules.
1995-02-25 22:08:48 +00:00
.else
Rework the checking out of the doc, src, and ports trees a bit to make it more manageable. - Add some helper variables (CVS_{SRC,DOC,PORTS}ARGS) to be used when using CVS to checkout files. We stick release tags in these helper variables if they are defined and then use only one cvs command instead of two cvs commands with an .ifdef to choose between them. - rm the old src/doc/ports directories as separate commands from the CVS comands so that the rm commands don't need to be duplicated. - Simplify the DOMINIMALDOCPORTS case by overriding RELEASEPORTSMODULE to be ${MINIMALDOCPORTS} thus removing yet another nearly-duplicate cvs command in an .ifdef. - Add support for grabbing src/ and doc/ from external directories specified via EXTSRCDIR and EXTDOCDIR instead of from CVS. The same is not done for ports/ quite yet as the DOMINIMALDOCPORTS case is a bit tricky. The rerelease target scripts have not been changed to use the helper variables yet, so there is still some room for improvement. Submitted by: kuriyama (4)
2002-11-15 22:03:25 +00:00
cd ${CHROOTDIR}/usr && \
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
${CVSPREFIX} cvs -R ${CVSARGS} -d ${CVSROOT} \
co ${CVSCMDARGS} ${CVS_SRCARGS} ${RELEASESRCMODULE}
Moved doFS into shell-script so we can be more smart about it. make two boot floppies, one without help files for 4MB systems. Mine's loading one now. :-) -O2 may be needed.
1996-01-08 22:13:50 +00:00
.endif
Make LOCAL_PATCHES accept multiple patch files, as its name suggests.
2003-09-21 06:33:14 +00:00
.if defined(LOCAL_PATCHES) && !empty(LOCAL_PATCHES)
.for p in ${LOCAL_PATCHES}
Fix ${LOCAL_PATCHES} (chdir was ignored).
2003-11-14 08:21:21 +00:00
patch -d ${CHROOTDIR}/usr/${RELEASESRCMODULE} ${PATCH_FLAGS} < ${p}
Make LOCAL_PATCHES accept multiple patch files, as its name suggests.
2003-09-21 06:33:14 +00:00
.endfor
Add support for CVSROOT/INTCVSROOT as simple makefile variables. Reviewed by: jkh
1996-09-07 01:05:16 +00:00
.endif
Remove some serious foot-shooting potential from the release Makefile. For some unknown reason, when LOCAL_PATCHES and LOCAL_SCRIPT were originally added, they were silently ignored if the actual file did not exist. As a result, if one mistyped the pathname to a patch or script, then the release silently succeeded. However, it was not built with the desired changes and no warning was given to inform the builder either. This commit explicitly checks to see that all of the defined patches and scripts exist up front and bails if any of them do not exist. I lost several hours of valuable sleeping time this evening due to this "feature" so I've finally gone and ripped out. I've tripped over this in the past several other times as well. Glanced at by: scottl
2004-07-15 04:52:57 +00:00
.if defined(LOCAL_SCRIPT)
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
cd ${CHROOTDIR} && env CHROOTDIR=${CHROOTDIR} BUILDNAME=${BUILDNAME} \
RELEASETAG=${RELEASETAG} ${LOCAL_SCRIPT}
Make provisions for a pre-build script that one can use to frob the chroot tree right after everything has been checked out and we're ready to go.
1998-08-16 01:24:40 +00:00
.endif
Rework the checking out of the doc, src, and ports trees a bit to make it more manageable. - Add some helper variables (CVS_{SRC,DOC,PORTS}ARGS) to be used when using CVS to checkout files. We stick release tags in these helper variables if they are defined and then use only one cvs command instead of two cvs commands with an .ifdef to choose between them. - rm the old src/doc/ports directories as separate commands from the CVS comands so that the rm commands don't need to be duplicated. - Simplify the DOMINIMALDOCPORTS case by overriding RELEASEPORTSMODULE to be ${MINIMALDOCPORTS} thus removing yet another nearly-duplicate cvs command in an .ifdef. - Add support for grabbing src/ and doc/ from external directories specified via EXTSRCDIR and EXTDOCDIR instead of from CVS. The same is not done for ports/ quite yet as the DOMINIMALDOCPORTS case is a bit tricky. The rerelease target scripts have not been changed to use the helper variables yet, so there is still some room for improvement. Submitted by: kuriyama (4)
2002-11-15 22:03:25 +00:00
rm -rf ${CHROOTDIR}/usr/ports
.if !defined(NOPORTSATALL)
Add EXTPORTSDIR like EXTSRCDIR and EXTDOCDIR. PR: misc/83287 Submitted by: Gabor Kovesdan MFC after: 1 week
2006-01-02 13:29:58 +00:00
.if defined(EXTPORTSDIR)
cd ${CHROOTDIR}/usr && cp -R -H ${EXTPORTSDIR} ports
# If there are distfiles downloaded removing them
rm -rf ports/distfiles/*
.else
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
cd ${CHROOTDIR}/usr && ${CVSPREFIX} cvs -R ${CVSARGS} -d ${CVSROOT} \
co ${CVSCMDARGS} ${CVS_PORTSARGS} ${RELEASEPORTSMODULE}
Add infrastructure that allows to build docs and RELNOTESng without the need to also create the entire ports tree during a `make release'. The main motivation behind this is that handling the ports tree still takes a huge amount of time due to the large number of directories involved, even on modern disks. The solution is to establish a list of dependent ports that are minimally required. This list needs to be manually maintained in case the doc ports toolchain changes, and has thus been broken out into a separate file Makefile.inc.docports. (release/Makefile has gotten overly lengthy already anyway.) Discussed with: bmah, nik Reviewed by: bmah
2001-05-04 19:50:34 +00:00
.endif
Add EXTPORTSDIR like EXTSRCDIR and EXTDOCDIR. PR: misc/83287 Submitted by: Gabor Kovesdan MFC after: 1 week
2006-01-02 13:29:58 +00:00
.endif
Add rules for building doc back into releases, conditionalized on NODOC. Also clean up some of the comments at the top. It's 12 o'clock, and that is all.
1997-05-24 20:15:13 +00:00
.if !defined(NODOC)
Rework the checking out of the doc, src, and ports trees a bit to make it more manageable. - Add some helper variables (CVS_{SRC,DOC,PORTS}ARGS) to be used when using CVS to checkout files. We stick release tags in these helper variables if they are defined and then use only one cvs command instead of two cvs commands with an .ifdef to choose between them. - rm the old src/doc/ports directories as separate commands from the CVS comands so that the rm commands don't need to be duplicated. - Simplify the DOMINIMALDOCPORTS case by overriding RELEASEPORTSMODULE to be ${MINIMALDOCPORTS} thus removing yet another nearly-duplicate cvs command in an .ifdef. - Add support for grabbing src/ and doc/ from external directories specified via EXTSRCDIR and EXTDOCDIR instead of from CVS. The same is not done for ports/ quite yet as the DOMINIMALDOCPORTS case is a bit tricky. The rerelease target scripts have not been changed to use the helper variables yet, so there is still some room for improvement. Submitted by: kuriyama (4)
2002-11-15 22:03:25 +00:00
rm -rf ${CHROOTDIR}/usr/doc
.if defined(EXTDOCDIR)
For EXT*DIR, use -H in addition to -R, so that a symlink may be used in the EXTSRCDIR case, for example, if one finds it useful to symlink /usr/src to an actual source tree, and use /usr/src as the EXTSRCDIR. This has no effects on release builds not using external directories. Reviewed by: ru Approved by: scottl (re)
2003-05-18 19:34:03 +00:00
cd ${CHROOTDIR}/usr && cp -R -H ${EXTDOCDIR} doc
Convert "RELENG_ver_RELEASE" to "RELEASE_ver" and use it to check out the ports and doc trees. There is no change if we are not building a release (i.e., we are not using a tag of the form RELENG_ver_RELEASE --e.g., a RELENG_ver tag used to build a snapshot). This should allow the ports and doc freezes during the release cycles to be somewhat shorter, as commits to those (non-branched) trees can resume as soon as the tags are laid down. Reviewed by: nik Approved by: jkh
2000-04-06 10:00:20 +00:00
.else
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
cd ${CHROOTDIR}/usr && ${CVSPREFIX} cvs -R ${CVSARGS} -d ${CVSROOT} \
co ${CVSCMDARGS} ${CVS_DOCARGS} ${RELEASEDOCMODULE}
Convert "RELENG_ver_RELEASE" to "RELEASE_ver" and use it to check out the ports and doc trees. There is no change if we are not building a release (i.e., we are not using a tag of the form RELENG_ver_RELEASE --e.g., a RELENG_ver tag used to build a snapshot). This should allow the ports and doc freezes during the release cycles to be somewhat shorter, as commits to those (non-branched) trees can resume as soon as the tags are laid down. Reviewed by: nik Approved by: jkh
2000-04-06 10:00:20 +00:00
.endif
Change variable name from DOCDISTFILES to RELEASEDISTFILES. Respect DOCDISTFILES if defined. No impact should happen with this change. DOCDISTFILES is introduced by rev. 1.541, to point the directory that contains the distfiles for ports/textproc/docproj. It's good, but we also build ports/sysutils/mkisofs during the release (if MAKE_ISOS is defined). It is natural that cdrtools tarball is prefetched, then copy it under ${CHROOTDIR}/usr/ports/distfiles. As a result, DOCDISTFILES variable is somewhat abused for non-doc tarballs; *DOC*DISTFILES is no longger appropriate name for that. PR: 16843 PR: 34194 Submitted by: murray MFC after: 6 days
2002-02-03 07:00:37 +00:00
if [ -d ${RELEASEDISTFILES}/ ]; then \
cp -rp ${RELEASEDISTFILES} ${CHROOTDIR}/usr/ports/distfiles; \
Mkdir ${CHROOTDIR}/usr/ports/distfiles if RELEASEDISTFILES is not set. We ensure that 'ports.tgz' tarball should have ports/distfiles directory. At first, ${CHROOTDIR}/usr/ports is empty (rm -rf ports), then extract ports files from CVS repository. Then if RELEASEDISTFILES is defined, ${CHROOTDIR}/usr/ports/distfiles directory is created by cp(1). However, if you don't specify RELEASEDISTFILES, there is no chance to create ${CHROOTDIR}/usr/ports/distfiles directory. Submitted by: {ushi,tora}.jp.FreeBSD.org MFC after: 5 days
2002-02-12 16:56:06 +00:00
else \
Typo fix: missing ;. Pointed out by: Manfred Antar <null@pozo.com> Pointy hat to: matusita
2002-02-25 05:35:39 +00:00
mkdir -p ${CHROOTDIR}/usr/ports/distfiles; \
Make release would fail for lack of /etc/resolv.conf or $DISTFILES/ PR: misc/9913 Submitted by: YAMAMOTO Shigeru <shigeru@bremen.or.jp>
1999-03-31 11:18:52 +00:00
fi
Make PREFETCHDISTFILES the default. Set NO_PREFETCHDISTFILES if you don't like this. Also make it easier to issue the distfile fetching on command.
2003-03-16 22:59:00 +00:00
.if !defined(NO_PREFETCHDISTFILES)
@cd ${.CURDIR} && ${MAKE} fetch-distfiles
If we are building docs with this release and PREFETCHDISTFILES is defined then download the distfiles for the ports needed to build the docs before we enter the chroot environment. This is useful since often times releases get in a funk trying to download distfiles in the chroot. Approved by: re (murray, bmah) MFC after: 5 days
2002-07-27 19:43:52 +00:00
.endif
Add rules for building doc back into releases, conditionalized on NODOC. Also clean up some of the comments at the top. It's 12 o'clock, and that is all.
1997-05-24 20:15:13 +00:00
.endif
Moved doFS into shell-script so we can be more smart about it. make two boot floppies, one without help files for 4MB systems. Mine's loading one now. :-) -O2 may be needed.
1996-01-08 22:13:50 +00:00
.endif
.if make(rerelease)
Enforce RELEASENOUPDATE if EXTSRCDIR is used. Suggested by: jb
2006-11-11 23:18:28 +00:00
.if !defined(RELEASENOUPDATE) && !defined(EXTSRCDIR)
Moved doFS into shell-script so we can be more smart about it. make two boot floppies, one without help files for 4MB systems. Mine's loading one now. :-) -O2 may be needed.
1996-01-08 22:13:50 +00:00
.if !defined(RELEASETAG)
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
cd ${CHROOTDIR}/usr/src && ${CVSPREFIX} cvs -R ${CVSARGS} -q \
update ${CVSCMDARGS} -P -d -A
Moved doFS into shell-script so we can be more smart about it. make two boot floppies, one without help files for 4MB systems. Mine's loading one now. :-) -O2 may be needed.
1996-01-08 22:13:50 +00:00
.else
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
cd ${CHROOTDIR}/usr/src && ${CVSPREFIX} cvs -R ${CVSARGS} -q \
update ${CVSCMDARGS} -P -d -r ${RELEASETAG}
Moved doFS into shell-script so we can be more smart about it. make two boot floppies, one without help files for 4MB systems. Mine's loading one now. :-) -O2 may be needed.
1996-01-08 22:13:50 +00:00
.endif
If we are told to update the sources in "make rerelease", we also probably want to re-buildworld. If we are not going to rebuild (make rerelease -DRELEASENOUPDATE and the file ${CHROOTDIR}/tmp/.world_done exists), do not touch ${CHROOTDIR}/usr/src/sys/conf/newvers.sh, as the only effect that is has is upsetting release.2. Hey, have you ever seen this "touch: not found" during rerelease? :-)
2003-02-26 15:32:12 +00:00
rm -f ${CHROOTDIR}/tmp/.world_done
Add EXTPORTSDIR like EXTSRCDIR and EXTDOCDIR. PR: misc/83287 Submitted by: Gabor Kovesdan MFC after: 1 week
2006-01-02 13:29:58 +00:00
.if !defined(NOPORTS) && !defined(EXTPORTSDIR)
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
cd ${CHROOTDIR}/usr/ports && ${CVSPREFIX} cvs -R ${CVSARGS} -q \
update ${CVSCMDARGS} -P -d
Rework how we layout the CD/DVD ISO images for releases. Note that this is only the first stage and does not yet handle packages (which will move to their own disc(s)). - By default, include the live filesystem on disc1, so that disc1 is now both an install disc and a fixit disc. The images can still be split into two separate discs by specifying SEPARATE_LIVEFS=yes. - Remove /usr/ports from the live filesystem disc as the merged images do not fit on even i386 otherwise. Also, the merged image already contains a tarball of the entire ports tree, so /usr/ports was a duplicate anyway. - Stop building port READMES for the release sinc they are no longer used (/usr/ports in the live filesystem was the last remaining user). - Split cdrom.1 into three targets. cdrom.1 builds the live filesystem image. cdrom.2 builds the disc1 image. Note that if SEPARATE_LIVEFS is not specified these both write to the same area. cdrom.3 builds the bootonly image and the UFS miniroot. - The various distribution directories are now in a <buildname> subdirectory of disc1 rather than being in the root so that the merged image's root is not quite as cluttered. - The disc1 image is now called disc1.iso and we no longer build a miniinst.iso image. If CD_EXTRA_BITS is defined, then the iso.1 target will simply include those bits in the disc1 image that it builds. MFC after: 1 day Requested by: re (scottl, etc.)
2005-03-02 22:46:07 +00:00
rm -f ${CHROOTDIR}/tmp/.skip_ports_index
Add support for ports collection distribution.
1997-03-11 00:07:11 +00:00
.endif
Add infrastructure that allows to build docs and RELNOTESng without the need to also create the entire ports tree during a `make release'. The main motivation behind this is that handling the ports tree still takes a huge amount of time due to the large number of directories involved, even on modern disks. The solution is to establish a list of dependent ports that are minimally required. This list needs to be manually maintained in case the doc ports toolchain changes, and has thus been broken out into a separate file Makefile.inc.docports. (release/Makefile has gotten overly lengthy already anyway.) Discussed with: bmah, nik Reviewed by: bmah
2001-05-04 19:50:34 +00:00
.if defined(DOMINIMALDOCPORTS) && ${DOMINIMALDOCPORTS} == "YES"
for i in ${MINIMALDOCPORTS}; do \
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
( cd ${CHROOTDIR}/usr/$$i && ${CVSPREFIX} cvs -R ${CVSARGS} -q \
update ${CVSCMDARGS} -P -d ) ; \
Add infrastructure that allows to build docs and RELNOTESng without the need to also create the entire ports tree during a `make release'. The main motivation behind this is that handling the ports tree still takes a huge amount of time due to the large number of directories involved, even on modern disks. The solution is to establish a list of dependent ports that are minimally required. This list needs to be manually maintained in case the doc ports toolchain changes, and has thus been broken out into a separate file Makefile.inc.docports. (release/Makefile has gotten overly lengthy already anyway.) Discussed with: bmah, nik Reviewed by: bmah
2001-05-04 19:50:34 +00:00
done
.endif
Add rules for building doc back into releases, conditionalized on NODOC. Also clean up some of the comments at the top. It's 12 o'clock, and that is all.
1997-05-24 20:15:13 +00:00
.if !defined(NODOC)
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
cd ${CHROOTDIR}/usr/doc && ${CVSPREFIX} cvs -R ${CVSARGS} -q \
update ${CVSCMDARGS} -P -d
Add rules for building doc back into releases, conditionalized on NODOC. Also clean up some of the comments at the top. It's 12 o'clock, and that is all.
1997-05-24 20:15:13 +00:00
.endif
Moved doFS into shell-script so we can be more smart about it. make two boot floppies, one without help files for 4MB systems. Mine's loading one now. :-) -O2 may be needed.
1996-01-08 22:13:50 +00:00
.endif
New and better "make release" rules.
1995-02-25 22:08:48 +00:00
.endif
Change the way distributions are built such that they can be built either in the Makefile or by an external script. Move some of the existing stuff into scripts.
1997-06-21 15:41:55 +00:00
# Add version information to those things that need it.
If we are told to update the sources in "make rerelease", we also probably want to re-buildworld. If we are not going to rebuild (make rerelease -DRELEASENOUPDATE and the file ${CHROOTDIR}/tmp/.world_done exists), do not touch ${CHROOTDIR}/usr/src/sys/conf/newvers.sh, as the only effect that is has is upsetting release.2. Hey, have you ever seen this "touch: not found" during rerelease? :-)
2003-02-26 15:32:12 +00:00
if [ ! -f ${CHROOTDIR}/tmp/.world_done ]; then \
cd ${CHROOTDIR}/usr/src/sys/conf && \
mv newvers.sh foo && \
sed "s/^RELEASE=.*/RELEASE=${BUILDNAME}/" foo > newvers.sh && \
rm foo; \
fi
When doing a rerelease, cvs update from within the source directory so we pick up the old CVSROOT if we don't have the environment variable set. If /usr/src/release/install.cfg is present, put it out onto the root filesystem of the boot floppy. It may optionally be used to pre-configure sysinstall with custom values. (See next batch of commits).
1996-09-26 20:48:03 +00:00
-test -f install.cfg && cp install.cfg ${CHROOTDIR}/usr/src/release
Make the name of the release running script overrideable.
2003-08-10 09:54:22 +00:00
echo "#!/bin/sh" > ${_MK}
echo "set -ex" >> ${_MK}
echo "trap 'umount /dev || true' 0" >> ${_MK}
echo "_RELTARGET=\$${1:-doRELEASE}" >> ${_MK}
Collapsed the export list, added CD_EXTRA_BITS to it.
2002-08-14 16:50:50 +00:00
.for var in \
MFS: Split export list.
2002-12-26 12:20:25 +00:00
AUTO_KEYBOARD_DETECT \
BUILDNAME \
What should be the last major changes to the ISO layout: - When a separate livefs ISO is used, change the name of the tree directory to be R/cdrom/livefs and change the name of the iso to livefs.iso rather than using disc2. Instead, disc2 is now always going to be the packages disc regardless of the presence of a separate livefs which makes things simpler. - Build a very simple disc2 tree under R/cdrom that just includes the cdrom.inf file that the packages need. Also, build a disc2.iso image in the iso.1 target. Disc 2's volume label is "FreeBSD_Packages". - Retire CD_EXTRA_BITS and replace it with CD_PACKAGE_TREE. CD_EXTRA_BITS was specific to disc1, but CD_PACKAGE_TREE instead should point to a directory that has disc1 and disc2 subdirectories that contain the extra bits for each image. - Rename the bootonly volume label to "FreeBSD_bootonly" and just hardcode the disc1 and livefs volume names and iso suffixes.
2005-03-23 15:01:54 +00:00
CD_PACKAGE_TREE \
MFS: Split export list.
2002-12-26 12:20:25 +00:00
DISTRIBUTIONS \
DOC_LANG \
DOMINIMALDOCPORTS \
EXTRA_SRC \
Pass FTP_PASSIVE_MODE, FTP_PROXY, and HTTP_PROXY variables to chroot sandbox. These variables will be used by pkg_add, which installs mkisofs package while MAKE_ISOS=YES. See src/release/${arch}/mkisoimages.sh if you doubt. No objection heard from: current@FreeBSD.org Suggested by: des, nectar (about *_PROXY variables)
2003-02-16 08:34:11 +00:00
FTP_PASSIVE_MODE \
FTP_PROXY \
HTTP_PROXY \
MFS: Split export list.
2002-12-26 12:20:25 +00:00
KERNELS \
fix handling of default kernels Reviewed by: ru
2006-03-11 18:59:16 +00:00
KERNELS_BASE \
MFS: Split export list.
2002-12-26 12:20:25 +00:00
KERNEL_FLAGS \
Follow 'MAKE_ISOS' in naming and s/FLOPPIES/MAKE_FLOPPIES/g Submitted by: ru
2004-08-16 08:56:50 +00:00
MAKE_FLOPPIES \
MFS: Split export list.
2002-12-26 12:20:25 +00:00
MAKE_ISOS \
NOCDROM \
NODOC \
NOPORTS \
NOSHARED -> NO_SHARED
2004-12-21 09:59:45 +00:00
NO_SHARED \
MFS: Split export list.
2002-12-26 12:20:25 +00:00
NOSRC \
NO_CPU_CFLAGS \
NO_CPU_COPTFLAGS \
Bring back 'NO_FLOPPIES', but as a user external knob. The internal logic knob remains MAKE_FLOPPIES. Suggested by: ru
2004-08-16 10:25:43 +00:00
NO_FLOPPIES \
MFS: Split export list.
2002-12-26 12:20:25 +00:00
RELEASETAG \
RELNOTES_LANG \
Rework how we layout the CD/DVD ISO images for releases. Note that this is only the first stage and does not yet handle packages (which will move to their own disc(s)). - By default, include the live filesystem on disc1, so that disc1 is now both an install disc and a fixit disc. The images can still be split into two separate discs by specifying SEPARATE_LIVEFS=yes. - Remove /usr/ports from the live filesystem disc as the merged images do not fit on even i386 otherwise. Also, the merged image already contains a tarball of the entire ports tree, so /usr/ports was a duplicate anyway. - Stop building port READMES for the release sinc they are no longer used (/usr/ports in the live filesystem was the last remaining user). - Split cdrom.1 into three targets. cdrom.1 builds the live filesystem image. cdrom.2 builds the disc1 image. Note that if SEPARATE_LIVEFS is not specified these both write to the same area. cdrom.3 builds the bootonly image and the UFS miniroot. - The various distribution directories are now in a <buildname> subdirectory of disc1 rather than being in the root so that the merged image's root is not quite as cluttered. - The disc1 image is now called disc1.iso and we no longer build a miniinst.iso image. If CD_EXTRA_BITS is defined, then the iso.1 target will simply include those bits in the disc1 image that it builds. MFC after: 1 day Requested by: re (scottl, etc.)
2005-03-02 22:46:07 +00:00
SEPARATE_LIVEFS \
MFS: Split export list.
2002-12-26 12:20:25 +00:00
TARGET \
TARGET_ARCH \
WORLD_FLAGS
Collapsed the export list, added CD_EXTRA_BITS to it.
2002-08-14 16:50:50 +00:00
.if defined(${var})
Make the name of the release running script overrideable.
2003-08-10 09:54:22 +00:00
echo "export ${var}=\"${${var}}\"" >> ${_MK}
Collapsed the export list, added CD_EXTRA_BITS to it.
2002-08-14 16:50:50 +00:00
.endif
.endfor
Reincarnate one occurance of RELEASEDIR, since the build depends on it inside some Makefiles. (Duh-oh, i've once included these references there myself! :) Noticed by akira@silvia.myaw.ei.meisei-u.ac.jp (Akira WATANABE), and by my yesterday's build.
1997-03-06 07:51:40 +00:00
# Don't remove this, or the build will fall over!
Make the name of the release running script overrideable.
2003-08-10 09:54:22 +00:00
echo "export RELEASEDIR=${_R}" >> ${_MK}
echo "export PATH=/bin:/usr/bin:/sbin:/usr/sbin:${LOCALDIR}" >> ${_MK}
echo "export MANBUILDCAT=YES" >> ${_MK}
Fixups to get make release going again for i386: o mount (and unmount) devfs in the chroot'd filesystem o handle umounting devfs in case of a keyboard interrupt o remove MAKEDEV from the fixit floppy so things fit again (mount_devfs was added in a separate commit) o explicitly remove mfsroot.gz and loader.conf when building the cd-rom area as otherwice cp -Rp aborts when a previous run left identical files around o increase the number of inodes on the drivers floppy; moving drivers from the kernel there (to shrink it's size) caused things to overflow o while we're increasing the driver's floppy inode config, separate out all parameters from the boot floppy Approved by: re
2002-10-19 16:44:16 +00:00
# NB: these may fail if the host is running w/o devfs
Make the name of the release running script overrideable.
2003-08-10 09:54:22 +00:00
echo "umount /dev >/dev/null 2>&1 || true" >> ${_MK}
echo "mount -t devfs devfs /dev >/dev/null 2>&1 || true" >> ${_MK}
Assert that /dev/null is a character device in a chroot early. Otherwise, we might get surprises later, like /dev/null having error in the 4th line reported by make(1). :-) Tested by: Dmitriy Kirhlarov (who attempted to make release in a jail)
2006-03-23 07:48:58 +00:00
echo "if [ ! -c /dev/null ]; then" >> ${_MK}
echo " echo /dev/null is not a device!" >> ${_MK}
echo " exit 1" >> ${_MK}
echo "fi" >> ${_MK}
Fix "5.x release on 4.x box" support. It is required to check /etc/rc.d/ldconfig before use, because ${CHROOT} sandbox is sometimes 4.x world. If /etc/rc.d/ldconfig is not there, run ldconfig(8) directly. Discussed with: imp, obrien, ru
2004-01-19 13:35:46 +00:00
echo "if [ -x /etc/rc.d/ldconfig ]; then" >> ${_MK}
echo " /etc/rc.d/ldconfig start" >> ${_MK}
echo "else" >> ${_MK}
echo " ldconfig /lib /usr/lib /usr/local/lib || true" >> ${_MK}
echo "fi" >> ${_MK}
Make the name of the release running script overrideable.
2003-08-10 09:54:22 +00:00
echo "if [ ! -f /tmp/.world_done ]; then" >> ${_MK}
echo " cd /usr/src" >> ${_MK}
NOCLEAN -> NO_CLEAN NOCLEANDIR -> NO_CLEANDIR
2004-12-21 12:21:26 +00:00
echo " ${CROSSMAKE} ${WORLD_FLAGS} -DNO_CLEAN buildworld && \\" >> ${_MK}
Die if make buildworld fails -- don't wait 'till installworld can't find what it needs.
2004-06-10 19:33:26 +00:00
echo " touch /tmp/.world_done || exit 1" >> ${_MK}
Make the name of the release running script overrideable.
2003-08-10 09:54:22 +00:00
echo "fi" >> ${_MK}
Fix-ups for a few problems reported with how ports are handled in the release: - run 'make index' as part of release build, INDEX* files in CVS no longer updated and likely to be removed from CVS soon - don't include README.html files in ports.tgz tarball because they cause cvsup problems for users later - put a copy of /usr/ports on disc2 that does include README.html files so users have access to them Reviewed by: ru (found one bug, several clean-up suggestions) Tested by: 'make release' MFC before: BETA6 build (this isn't likely to be tested in -current anyway...)
2004-09-24 02:10:06 +00:00
echo "if [ ! -f /tmp/.skip_ports_index ]; then" >> ${_MK}
echo " echo \">>> make index started on \`LC_ALL=C TZ=GMT date\`\"" >> ${_MK}
echo " for i in ${MAKEINDEXPORTS}" >> ${_MK}
echo " do" >> ${_MK}
echo " cd /usr/ports/\$${i}" >> ${_MK}
echo " env -i FTP_PASSIVE_MODE=$${FTP_PASSIVE_MODE:-no} PATH=$${PATH} \\" >> ${_MK}
echo " make all install clean BATCH=yes FORCE_PKG_REGISTER=yes" >> ${_MK}
echo " done" >> ${_MK}
echo " cd /usr/ports" >> ${_MK}
echo " rm -f INDEX*" >> ${_MK}
echo " make index -DINDEX_PRISTINE" >> ${_MK}
It turns out the ia64 architecture's disc2 overflows if we put the ports/ tree on it. This makes putting the ports/ tree on disc2 at all dependent on the NOPORTREADMES knob as well as the NOPORT knob (at the moment NOPORT may be set while NOPORTREADMES isn't, that should probably be revisited). And it hardcodes ia64 for NOPORTREADMES by request the ia64 release builder. While here really get rid of the temp file 'make index' leaves behind. Tested by: i386 and ia64 'make release' Reviewed by: marcel MFC after: 1 day
2004-09-30 22:59:47 +00:00
echo " rm -f INDEX*.tmp" >> ${_MK}
Fix-ups for a few problems reported with how ports are handled in the release: - run 'make index' as part of release build, INDEX* files in CVS no longer updated and likely to be removed from CVS soon - don't include README.html files in ports.tgz tarball because they cause cvsup problems for users later - put a copy of /usr/ports on disc2 that does include README.html files so users have access to them Reviewed by: ru (found one bug, several clean-up suggestions) Tested by: 'make release' MFC before: BETA6 build (this isn't likely to be tested in -current anyway...)
2004-09-24 02:10:06 +00:00
echo " touch /tmp/.skip_ports_index" >> ${_MK}
echo " echo \">>> make index finished on \`LC_ALL=C TZ=GMT date\`\"" >> ${_MK}
echo "fi" >> ${_MK}
Make the name of the release running script overrideable.
2003-08-10 09:54:22 +00:00
echo "cd /usr/src/release" >> ${_MK}
echo "make obj" >> ${_MK}
echo "make \$${_RELTARGET}" >> ${_MK}
echo "echo \">>> make ${.TARGET} for ${TARGET} finished on \`LC_ALL=C TZ=GMT date\`\"" >> ${_MK}
chmod 755 ${_MK}
Fix-ups for a few problems reported with how ports are handled in the release: - run 'make index' as part of release build, INDEX* files in CVS no longer updated and likely to be removed from CVS soon - don't include README.html files in ports.tgz tarball because they cause cvsup problems for users later - put a copy of /usr/ports on disc2 that does include README.html files so users have access to them Reviewed by: ru (found one bug, several clean-up suggestions) Tested by: 'make release' MFC before: BETA6 build (this isn't likely to be tested in -current anyway...)
2004-09-24 02:10:06 +00:00
.if defined(NOPORTS)
touch ${CHROOTDIR}/tmp/.skip_ports_index
Have the port readmes follow the same logic as buildworld. We skip doing make readmes if /tmp/.skip_ports exists. This file is created after a complete run, which avoids that we do the lengthy operation again for a rerelease. When NOPORTS or NOPORTREADMES are defined, we create the file prior to starting the release.
2003-06-04 05:17:18 +00:00
.endif
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
# Ensure md.ko is loaded if md(4) is not statically compiled into
# the kernel
In rev 1.603 I should have changed it to mdconfig rather than remove it. I don't recall now, but for some reason I didn't think it was needed. Also place the mdconfig right before the chroot call, to make it stand out this needs to be done in the standard environment, as we've tried doing it in the chroot environment before.
2003-03-17 18:49:46 +00:00
-mdconfig 2>/dev/null
2nd part of making the name of the release running script overrideable.
2003-08-11 04:21:58 +00:00
env -i /usr/sbin/chroot `dirname ${_MK}` /`basename ${_MK}`
Make a less expensive re-release target.
1995-05-09 00:30:00 +00:00
New and better "make release" rules.
1995-02-25 22:08:48 +00:00
clean:
Clean up after mechanical renumbering.
2004-08-06 10:04:06 +00:00
rm -rf ${CRUNCH_TARGETS:S/$/_crunch/} release.[0-8] ${EXTRAS}
Remove extra leading tabs that made this thing blow up.
1995-01-13 08:06:55 +00:00
Make PREFETCHDISTFILES the default. Set NO_PREFETCHDISTFILES if you don't like this. Also make it easier to issue the distfile fetching on command.
2003-03-16 22:59:00 +00:00
fetch-distfiles:
Fix-ups for a few problems reported with how ports are handled in the release: - run 'make index' as part of release build, INDEX* files in CVS no longer updated and likely to be removed from CVS soon - don't include README.html files in ports.tgz tarball because they cause cvsup problems for users later - put a copy of /usr/ports on disc2 that does include README.html files so users have access to them Reviewed by: ru (found one bug, several clean-up suggestions) Tested by: 'make release' MFC before: BETA6 build (this isn't likely to be tested in -current anyway...)
2004-09-24 02:10:06 +00:00
@for i in ${MAKEINDEXPORTS}; do \
cd ${CHROOTDIR}/usr/ports/$$i && \
make PORTSDIR=${CHROOTDIR}/usr/ports BATCH=yes \
checksum-recursive ; \
done
Make PREFETCHDISTFILES the default. Set NO_PREFETCHDISTFILES if you don't like this. Also make it easier to issue the distfile fetching on command.
2003-03-16 22:59:00 +00:00
@for i in ${DOCPORTS}; do \
cd ${CHROOTDIR}/usr/ports/$$i && \
make PORTSDIR=${CHROOTDIR}/usr/ports BATCH=yes \
WITHOUT_X11=yes JADETEX=no WITHOUT_PYTHON=yes \
checksum-recursive ; \
done
Some minor cosmetics: make mkdir less chatty in the output and stop making /R magic.
1999-09-16 08:44:22 +00:00
# Clean out ${_R} and make the directory structure.
New and better "make release" rules.
1995-02-25 22:08:48 +00:00
release.1:
Some minor cosmetics: make mkdir less chatty in the output and stop making /R magic.
1999-09-16 08:44:22 +00:00
mkdir -p ${_R}
Guard against an rm returning non-zero status. Noticed by: Makoto MATSUSHITA
1999-09-23 18:40:54 +00:00
-rm -rf ${_R}/* 2> /dev/null
Some minor cosmetics: make mkdir less chatty in the output and stop making /R magic.
1999-09-16 08:44:22 +00:00
-chflags -R noschg ${_R}/.
rm -rf ${_R}/*
New and better "make release" rules.
1995-02-25 22:08:48 +00:00
mkdir ${RD}
Follow 'MAKE_ISOS' in naming and s/FLOPPIES/MAKE_FLOPPIES/g Submitted by: ru
2004-08-16 08:56:50 +00:00
.if defined(MAKE_FLOPPIES)
New and better "make release" rules.
1995-02-25 22:08:48 +00:00
mkdir ${RD}/floppies
Make NO_FLOPPIES really mean "no floppies", including not even creating the "floppies" directory on FTP and CD-ROMs. Enforce NO_FLOPPIES on amd64. Tested for: amd64, i386 with -DNO_FLOPPIES
2003-07-23 12:12:53 +00:00
.endif
I belive the floppies work as expected now. (incl the fixit)
1995-03-15 06:14:19 +00:00
mkdir ${RD}/trees
- Tweak order of arch fs options to stick all floppy options together and all MFS root options together and in the same order on all archs. - Drop support for using the no longer existing 2.88 floppy image for CD booting on i386 via the EMUL_BOOT variable. - Use a stock loader on the boot floppy rather than one without forth. - Force the gzip of the mfsroot in release.9 to succeed. This is useful when doing lots of rereleases when testing floppy changes. - If an acpi kernel module exists, stick it on the boot floppy and for now always load it after the kernel.
2004-01-29 19:01:58 +00:00
for i in ${DISTRIBUTIONS}; do \
Moved doFS into shell-script so we can be more smart about it. make two boot floppies, one without help files for 4MB systems. Mine's loading one now. :-) -O2 may be needed.
1996-01-08 22:13:50 +00:00
mkdir ${RD}/trees/$$i && \
Backout addition of -L switch to mtree. Using -L breaks the build process in too many cases. Adding mtree to bootstrap-tools to solve this breaks the upgrade path because mtree needs a libc that has strtofflags and fflagstostr.
2000-07-23 16:33:00 +00:00
mtree -deU -f ${MTREEFILES}/BSD.root.dist \
Moved doFS into shell-script so we can be more smart about it. make two boot floppies, one without help files for 4MB systems. Mine's loading one now. :-) -O2 may be needed.
1996-01-08 22:13:50 +00:00
-p ${RD}/trees/$$i > /dev/null && \
Backout addition of -L switch to mtree. Using -L breaks the build process in too many cases. Adding mtree to bootstrap-tools to solve this breaks the upgrade path because mtree needs a libc that has strtofflags and fflagstostr.
2000-07-23 16:33:00 +00:00
mtree -deU -f ${MTREEFILES}/BSD.usr.dist \
Moved doFS into shell-script so we can be more smart about it. make two boot floppies, one without help files for 4MB systems. Mine's loading one now. :-) -O2 may be needed.
1996-01-08 22:13:50 +00:00
-p ${RD}/trees/$$i/usr > /dev/null && \
Backout addition of -L switch to mtree. Using -L breaks the build process in too many cases. Adding mtree to bootstrap-tools to solve this breaks the upgrade path because mtree needs a libc that has strtofflags and fflagstostr.
2000-07-23 16:33:00 +00:00
mtree -deU -f ${MTREEFILES}/BSD.include.dist \
Moved doFS into shell-script so we can be more smart about it. make two boot floppies, one without help files for 4MB systems. Mine's loading one now. :-) -O2 may be needed.
1996-01-08 22:13:50 +00:00
-p ${RD}/trees/$$i/usr/include > /dev/null && \
Backout addition of -L switch to mtree. Using -L breaks the build process in too many cases. Adding mtree to bootstrap-tools to solve this breaks the upgrade path because mtree needs a libc that has strtofflags and fflagstostr.
2000-07-23 16:33:00 +00:00
mtree -deU -f ${MTREEFILES}/BSD.var.dist \
I belive the floppies work as expected now. (incl the fixit)
1995-03-15 06:14:19 +00:00
-p ${RD}/trees/$$i/var > /dev/null ; \
Create directory symlinks neccesary for locale because mtree can't create symlinks. Suggested by Rod.
1995-02-02 08:03:24 +00:00
done
Revamp base system packaging of kernels to enable up/smp selection at runtime and to support distributing additional kernels: o remove kernel from the base tarball o add new kernel tarballs o build + package both SMP and GENERIC kernels when an <arch>/conf/SMP config file is present o add sysinstall support for multiple kernels o update sysinstall to probe for the number of cpus on a system and auto-select smp/up kernel accordingly o add a post-kernels install hook to fixup /boot/kernel o add -ldevinfo to boot crunch for sysinstall's cpu probing logic Notes: 1. On HEAD this code is not currently used because GENERIC kernels include SMP. This work is mainly intended for RELENG_6 where the GENERIC kernel is UP. If HEAD changes to match then just enable WITH_SMP in sysinstall/Makefile. 2. The cpu probing support is done with acpi and MPTable; this means some systems will require work for auto-detection to work. 3. The handling of /boot/kernel may need to be revisited; for now we rename one kernel at the last moment (SMP if installed, otherwise GENERIC). There are other, possibly better, approaches. Lots of help from ru, emaste, scottl, and jhb.
2006-03-08 18:02:32 +00:00
mkdir ${RD}/kernels
fix handling of default kernels Reviewed by: ru
2006-03-11 18:59:16 +00:00
for i in ${KERNELS_BASE} ${KERNELS}; do \
mkdir -p ${RD}/kernels/$${i}; \
Revamp base system packaging of kernels to enable up/smp selection at runtime and to support distributing additional kernels: o remove kernel from the base tarball o add new kernel tarballs o build + package both SMP and GENERIC kernels when an <arch>/conf/SMP config file is present o add sysinstall support for multiple kernels o update sysinstall to probe for the number of cpus on a system and auto-select smp/up kernel accordingly o add a post-kernels install hook to fixup /boot/kernel o add -ldevinfo to boot crunch for sysinstall's cpu probing logic Notes: 1. On HEAD this code is not currently used because GENERIC kernels include SMP. This work is mainly intended for RELENG_6 where the GENERIC kernel is UP. If HEAD changes to match then just enable WITH_SMP in sysinstall/Makefile. 2. The cpu probing support is done with acpi and MPTable; this means some systems will require work for auto-detection to work. 3. The handling of /boot/kernel may need to be revisited; for now we rename one kernel at the last moment (SMP if installed, otherwise GENERIC). There are other, possibly better, approaches. Lots of help from ru, emaste, scottl, and jhb.
2006-03-08 18:02:32 +00:00
done
Replaced the hardcoded target names in the touch(1) commands with ${.TARGET}, to ease maintenance.
2003-07-24 13:45:05 +00:00
touch ${.TARGET}
New and better "make release" rules.
1995-02-25 22:08:48 +00:00
# Install the system into the various distributions.
release.2:
Make it possible again to build -CURRENT snapshots on -STABLE. Reported by: Angelo Turetta
2006-09-11 13:15:09 +00:00
cd ${.CURDIR}/.. && ${CROSSMAKE} distrib-dirs DESTDIR=${RD}/trees/base
Pass WORLD_FLAGS down to the distributeworld target, so it's possible to build releases with customizations such as with NOPROFILE. Problem found and fix tested by Roman Neuhauser.
2004-10-21 08:54:10 +00:00
cd ${.CURDIR}/.. && ${CROSSMAKE} ${WORLD_FLAGS} distributeworld \
DISTDIR=${RD}/trees
Replaced the hardcoded target names in the touch(1) commands with ${.TARGET}, to ease maintenance.
2003-07-24 13:45:05 +00:00
touch ${.TARGET}
Split release.2 into two stages. release.2 now install the base dist, while release.3 makes and installs crypto, krb4, and krb5 dists. This makes release.3 restartable (with "make rerelease").
2002-05-14 13:59:33 +00:00
Sync this up with RELENG_2_2. Du-oh, i've totally spammed rev 1.278, and quite a number of smaller fixes from 2.2 were missing as well.
1997-04-12 21:55:53 +00:00
# Make and install the generic kernel(s).
Crypto is now part of the "base" distribution, and the old "release.3" target no longer exists. Renumber the old release.[4-9] targets to release.[3-8]. Reminded by: ru
2004-08-06 08:49:51 +00:00
release.3:
fix handling of default kernels Reviewed by: ru
2006-03-11 18:59:16 +00:00
.for kernel in ${KERNELS_BASE} ${KERNELS}
Milestone #1 in cross-arch make releases. Do not install games and profiled libraries to the ${CHROOTDIR} with the initial installworld. Eliminate the need in the second installworld. For that, make sure _everything_ is built in the "world" environment, using the right tool chain. Added SUBDIR_OVERRIDE helper stuff to Makefile.inc1. Split the buildworld process into stages, and skip some stages when SUBDIR_OVERRIDE is set (used to build crypto, krb4, and krb5 dists). Added NO_MAKEDB_RUN knob to Makefile.inc1 to avoid running makewhatis(1) at the end of installworld (used when making crypto, krb4, and krb5 dists). In release/scripts/doFS.sh, ensure that the correct boot blocks are used. Moved the creation of the "crypto" dist from release.5 to release.2. In release.3 and doMFSKERN, build kernels in the "world" environment. KERNELS now means "additional" kernels, GENERIC is always built. Ensure we build crunched binaries in the "world" environment. Obfuscate release/Makefile some more (WMAKEENV) to achieve this. Inline createBOOTMFS target. Use already built GENERIC kernel modules to augment mfsfd's /stand/modules. GC doMODULES as such. Assorted fixes: Get rid of the "afterdistribute" target by moving the single use of it from sys/Makefile to etc/Makefile's "distribute". Makefile.inc1: apparently "etc" no longer needs to be last for "distribute" to succeed. gnu/usr.bin/perl/library/Makefile.inc: do not override the "install" and "distribute" targets, do it the "canonical" way. release/scripts/{man,cat}pages-make.sh: make sure Perl manpages and catpages appear in the right dists. Note that because Perl does not respect the MANBUILDCAT (and NOMAN), this results in a loss of /usr/share/perl/man/cat* empty directories. This will be fixed soon. Turn MAKE_KERBEROS4 into a plain boolean variable (if it is set it means "make KerberosIV"), as documented in the make.conf(5) manpage. Most of the userland makefiles did not test it for "YES" anyway. XXX Should specialized kerberized libpam versions be included into the krb4 and krb5 dists? (libpam.a would be incorrect anyway if both krb4 and krb5 dists were choosen.) Make sure "games" dist is made before "catpages", otherwise games catpages settle in the wrong dist. Fast build machine provided by: Igor Kucherenko <kivvy@sunbay.com>
2002-04-26 17:55:27 +00:00
cd ${.CURDIR}/..; \
Fix ${TMAKE} to use the correct .mk files. Removed stray -DNO_WERROR's.
2002-07-12 15:15:35 +00:00
${CROSSMAKE} ${KERNEL_FLAGS} \
Revamp base system packaging of kernels to enable up/smp selection at runtime and to support distributing additional kernels: o remove kernel from the base tarball o add new kernel tarballs o build + package both SMP and GENERIC kernels when an <arch>/conf/SMP config file is present o add sysinstall support for multiple kernels o update sysinstall to probe for the number of cpus on a system and auto-select smp/up kernel accordingly o add a post-kernels install hook to fixup /boot/kernel o add -ldevinfo to boot crunch for sysinstall's cpu probing logic Notes: 1. On HEAD this code is not currently used because GENERIC kernels include SMP. This work is mainly intended for RELENG_6 where the GENERIC kernel is UP. If HEAD changes to match then just enable WITH_SMP in sysinstall/Makefile. 2. The cpu probing support is done with acpi and MPTable; this means some systems will require work for auto-detection to work. 3. The handling of /boot/kernel may need to be revisited; for now we rename one kernel at the last moment (SMP if installed, otherwise GENERIC). There are other, possibly better, approaches. Lots of help from ru, emaste, scottl, and jhb.
2006-03-08 18:02:32 +00:00
KERNCONF=${kernel} kernel \
DESTDIR=${RD}/kernels KODIR=/${kernel}
.endfor
Replaced the hardcoded target names in the touch(1) commands with ${.TARGET}, to ease maintenance.
2003-07-24 13:45:05 +00:00
touch ${.TARGET}
New and better "make release" rules.
1995-02-25 22:08:48 +00:00
Sanitize the making of kernels a bit, and make the BOOTCDROM kernel too. If we can get away with it, it will always just be the GENERIC kernel with a different name.
1995-04-10 08:01:27 +00:00
# Make and install the three crunched binaries which live on the floppies.
New and better "make release" rules.
1995-02-25 22:08:48 +00:00
# You are not supposed to like this :-)
Crypto is now part of the "base" distribution, and the old "release.3" target no longer exists. Renumber the old release.[4-9] targets to release.[3-8]. Reminded by: ru
2004-08-06 08:49:51 +00:00
release.4:
A major cleanup. Move more over to a "pull" model than a "push" model, this will eventually allow us to add meaningfull dependencies. Add mount_{msdos,nfs,cd9660} and chroot to the fixit.flp.
1995-04-20 06:48:31 +00:00
rm -rf ${RD}/crunch
mkdir -p ${RD}/crunch
Some components of the crunched binaries may use build-tools that are to be run on the building host. (Currently, bin/sh and usr.sbin/sysinstall use them.) Make sure we build them in the correct environment, like we do it in the buildworld.
2002-04-29 15:26:37 +00:00
.for j in ${CRUNCH_TARGETS}
- Unbreak Alpha snapshots (we were still attempting to build a no longer existing "fixit" crunched binary). - Simplify the CRUNCH_TARGETS variable's initialization. - Simplify the release.5 logic a bit by accounting for the fact that release/*_crunch.conf files were removed two years ago.
2003-08-06 08:09:40 +00:00
.if exists(${.CURDIR}/${TARGET}/${j}_crunch.conf)
Some components of the crunched binaries may use build-tools that are to be run on the building host. (Currently, bin/sh and usr.sbin/sysinstall use them.) Make sure we build them in the correct environment, like we do it in the buildworld.
2002-04-29 15:26:37 +00:00
rm -rf ${j}_crunch
mkdir ${j}_crunch
Make it possible again to build -CURRENT snapshots on -STABLE. Reported by: Angelo Turetta
2006-09-11 13:15:09 +00:00
cd ${j}_crunch; ${WMAKEENV} MAKEFLAGS="-m ${.CURDIR}/../share/mk" \
When compiled with the combination of flags used to build crunched binaries for the fixit floppy bin/ed/main.c causes a gcc warning message about argc possibly being clobbered by longjmp or vfork. We have threatened to ditch floppies for 8.0 but I don't want to do quite that much rototilling yet so for now turn off -Werror while building ed (and everything else) for the fixit floppy. Thanks to jb for pointing out NO_WERROR.
2007-12-04 02:22:35 +00:00
NO_WERROR= crunchgen -o \
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
${.CURDIR}/${TARGET}/${j}_crunch.conf
Some components of the crunched binaries may use build-tools that are to be run on the building host. (Currently, bin/sh and usr.sbin/sysinstall use them.) Make sure we build them in the correct environment, like we do it in the buildworld.
2002-04-29 15:26:37 +00:00
cd ${j}_crunch; ${WMAKE} -f ${j}_crunch.mk subclean
Fixed the last-minute bug I made in previous commit.
2003-07-04 14:39:17 +00:00
cd ${.CURDIR}/..; ${BINMAKE} -f Makefile.inc1 _build-tools
Now that make(1) passes variables that were given on the command line to subordinate make(1) invocations through MAKEFLAGS, we cannot add CFLAGS onto the make(1) command line. This will conflict with the individual makefiles wanting to append to it, which is not respected when CFLAGS is given on the command line. Hence build breakage. So, put CFLAGS in the environment instead.
2004-08-06 17:07:40 +00:00
cd ${j}_crunch; CFLAGS="-Os -pipe" ${WMAKE} -f ${j}_crunch.mk \
-DNO_CPU_CFLAGS all
Part 2/3 of unbreaking cross releases: Use the right strip(1) binary.
2003-01-29 08:39:28 +00:00
${WMAKEENV} strip -R .comment ${j}_crunch/${j}_crunch
Some components of the crunched binaries may use build-tools that are to be run on the building host. (Currently, bin/sh and usr.sbin/sysinstall use them.) Make sure we build them in the correct environment, like we do it in the buildworld.
2002-04-29 15:26:37 +00:00
mv ${j}_crunch/${j}_crunch ${RD}/crunch/${j}
- Unbreak Alpha snapshots (we were still attempting to build a no longer existing "fixit" crunched binary). - Simplify the CRUNCH_TARGETS variable's initialization. - Simplify the release.5 logic a bit by accounting for the fact that release/*_crunch.conf files were removed two years ago.
2003-08-06 08:09:40 +00:00
.endif
Some components of the crunched binaries may use build-tools that are to be run on the building host. (Currently, bin/sh and usr.sbin/sysinstall use them.) Make sure we build them in the correct environment, like we do it in the buildworld.
2002-04-29 15:26:37 +00:00
.endfor
Replaced the hardcoded target names in the touch(1) commands with ${.TARGET}, to ease maintenance.
2003-07-24 13:45:05 +00:00
touch ${.TARGET}
New and better "make release" rules.
1995-02-25 22:08:48 +00:00
A new cleaned up Makefile.
1996-01-11 20:28:40 +00:00
#
# --==## Fix up the distributions. ##==--
#
Crypto is now part of the "base" distribution, and the old "release.3" target no longer exists. Renumber the old release.[4-9] targets to release.[3-8]. Reminded by: ru
2004-08-06 08:49:51 +00:00
release.5:
Change the way distributions are built such that they can be built either in the Makefile or by an external script. Move some of the existing stuff into scripts.
1997-06-21 15:41:55 +00:00
# Create any "synthetic dists" now.
@for i in ${DISTRIBUTIONS}; do \
if [ -f ${.CURDIR}/scripts/$${i}-make.sh ]; then \
echo -n "Running $$i dist creation script... "; \
Initiate deorbit burn for the i386-only a.out related support. Moves are under way to move the remnants of the a.out toolchain to ports. As the comment in src/Makefile said, this stuff is deprecated and one should not expect this to remain beyond 4.0-REL. It has already lasted WAY beyond that. Notable exceptions: gcc - I have not touched the a.out generation stuff there. ldd/ldconfig - still have some code to interface with a.out rtld. old as/ld/etc - I have not removed these yet, pending their move to ports. some includes - necessary for ldd/ldconfig for now. Tested on: i386 (extensively), alpha
2002-09-17 01:49:00 +00:00
env RD=${RD} sh ${.CURDIR}/scripts/$${i}-make.sh || echo "$$i distribution script returned bad status."; \
Whoops, get the proper rev of this.
1997-06-21 15:45:51 +00:00
echo "Done."; \
Change the way distributions are built such that they can be built either in the Makefile or by an external script. Move some of the existing stuff into scripts.
1997-06-21 15:41:55 +00:00
fi \
done \
A new cleaned up Makefile.
1996-01-11 20:28:40 +00:00
# Remove all the directories we don't need.
Fix the bug which removes all empty dirs from distribution. Sorry Jordan...
1995-04-08 09:45:31 +00:00
-cd ${RD}/trees && \
Provide 32-bit runtime support on amd64 as a separate distribution, lib32. Prodded by: obrien Nodded by: peter Approved by: re
2005-06-16 18:16:14 +00:00
(find ${OTHER_DISTS} -path '*/var/empty' | xargs chflags noschg; \
When deleting empty directories prior to building the release tarballs, pass the "-empty" flag to find(1). This change has no effect other than to get rid of a few pages of "rmdir: ...: Directory not empty" error messages. (Note that the "-empty" flag has been supported by find(1) since 4.3-RELEASE.)
2006-06-09 03:40:04 +00:00
find ${OTHER_DISTS} -depth -type d -empty -print | xargs rmdir)
Replaced the hardcoded target names in the touch(1) commands with ${.TARGET}, to ease maintenance.
2003-07-24 13:45:05 +00:00
touch ${.TARGET}
New and better "make release" rules.
1995-02-25 22:08:48 +00:00
A new cleaned up Makefile.
1996-01-11 20:28:40 +00:00
#
Change the way distributions are built such that they can be built either in the Makefile or by an external script. Move some of the existing stuff into scripts.
1997-06-21 15:41:55 +00:00
# --==## Package up the tarballs from assembled trees ##==--
A new cleaned up Makefile.
1996-01-11 20:28:40 +00:00
#
Crypto is now part of the "base" distribution, and the old "release.3" target no longer exists. Renumber the old release.[4-9] targets to release.[3-8]. Reminded by: ru
2004-08-06 08:49:51 +00:00
release.6:
Add rule to generate the info files. Submitted by: gpalmer
1995-05-24 01:26:23 +00:00
rm -rf ${RD}/dists
mkdir -p ${RD}/dists
@for i in ${DISTRIBUTIONS} ; \
do \
if [ -d ${RD}/trees/$${i} ] ; then \
A new cleaned up Makefile.
1996-01-11 20:28:40 +00:00
cd ${.CURDIR} && $(MAKE) doTARBALL \
SD=${RD}/trees/$${i} \
TN=$$i TD=$$i ARG="." && \
Add rule to generate the info files. Submitted by: gpalmer
1995-05-24 01:26:23 +00:00
echo "$${i} distribution is finished."; \
fi ; \
Remove trailing whitespace.
1995-05-30 08:29:07 +00:00
done
fix handling of default kernels Reviewed by: ru
2006-03-11 18:59:16 +00:00
@for i in ${KERNELS_BASE} ${KERNELS} ; \
Revamp base system packaging of kernels to enable up/smp selection at runtime and to support distributing additional kernels: o remove kernel from the base tarball o add new kernel tarballs o build + package both SMP and GENERIC kernels when an <arch>/conf/SMP config file is present o add sysinstall support for multiple kernels o update sysinstall to probe for the number of cpus on a system and auto-select smp/up kernel accordingly o add a post-kernels install hook to fixup /boot/kernel o add -ldevinfo to boot crunch for sysinstall's cpu probing logic Notes: 1. On HEAD this code is not currently used because GENERIC kernels include SMP. This work is mainly intended for RELENG_6 where the GENERIC kernel is UP. If HEAD changes to match then just enable WITH_SMP in sysinstall/Makefile. 2. The cpu probing support is done with acpi and MPTable; this means some systems will require work for auto-detection to work. 3. The handling of /boot/kernel may need to be revisited; for now we rename one kernel at the last moment (SMP if installed, otherwise GENERIC). There are other, possibly better, approaches. Lots of help from ru, emaste, scottl, and jhb.
2006-03-08 18:02:32 +00:00
do \
if [ -d ${RD}/kernels/$${i} ] ; then \
cd ${.CURDIR} && $(MAKE) doTARBALL \
SD=${RD}/kernels \
TN=$$i TD=kernels ARG="$$i" && \
echo "$${i} distribution is finished."; \
fi ; \
done
- Boot-only CD-ROM support (bootonly.iso). Submitted by: jhb, ru - Moved the creation of the ports distribution to release.7. - Call MFS root floppy mfsroot.flp even on a small PC98 disk. Submitted by: jhb Reviewed by: ru, jhb
2003-08-02 18:21:02 +00:00
.if !defined(NOPORTS)
# XXX: Inline stripped version of doTARBALL
@rm -rf ${RD}/dists/ports/ports*
@mkdir -p ${RD}/dists/ports
@echo rolling ports/ports tarball
Fix-ups for a few problems reported with how ports are handled in the release: - run 'make index' as part of release build, INDEX* files in CVS no longer updated and likely to be removed from CVS soon - don't include README.html files in ports.tgz tarball because they cause cvsup problems for users later - put a copy of /usr/ports on disc2 that does include README.html files so users have access to them Reviewed by: ru (found one bug, several clean-up suggestions) Tested by: 'make release' MFC before: BETA6 build (this isn't likely to be tested in -current anyway...)
2004-09-24 02:10:06 +00:00
@tar --exclude CVS --exclude 'ports/distfiles/*' \
Rework how we layout the CD/DVD ISO images for releases. Note that this is only the first stage and does not yet handle packages (which will move to their own disc(s)). - By default, include the live filesystem on disc1, so that disc1 is now both an install disc and a fixit disc. The images can still be split into two separate discs by specifying SEPARATE_LIVEFS=yes. - Remove /usr/ports from the live filesystem disc as the merged images do not fit on even i386 otherwise. Also, the merged image already contains a tarball of the entire ports tree, so /usr/ports was a duplicate anyway. - Stop building port READMES for the release sinc they are no longer used (/usr/ports in the live filesystem was the last remaining user). - Split cdrom.1 into three targets. cdrom.1 builds the live filesystem image. cdrom.2 builds the disc1 image. Note that if SEPARATE_LIVEFS is not specified these both write to the same area. cdrom.3 builds the bootonly image and the UFS miniroot. - The various distribution directories are now in a <buildname> subdirectory of disc1 rather than being in the root so that the merged image's root is not quite as cluttered. - The disc1 image is now called disc1.iso and we no longer build a miniinst.iso image. If CD_EXTRA_BITS is defined, then the iso.1 target will simply include those bits in the disc1 image that it builds. MFC after: 1 day Requested by: re (scottl, etc.)
2005-03-02 22:46:07 +00:00
-czf ${RD}/dists/ports/ports.tgz -C /usr ports
- Boot-only CD-ROM support (bootonly.iso). Submitted by: jhb, ru - Moved the creation of the ports distribution to release.7. - Call MFS root floppy mfsroot.flp even on a small PC98 disk. Submitted by: jhb Reviewed by: ru, jhb
2003-08-02 18:21:02 +00:00
@cp ${.CURDIR}/scripts/ports-install.sh ${RD}/dists/ports/install.sh
@(cd ${RD}/dists/ports; \
Fix a bug which puts an incorrect checksum of CHECKSUM.MD5 into the CHECKSUM.MD5 file itself. Approved by: re (scottl) MFC after: 3 days
2006-02-08 11:18:29 +00:00
rm -f CHECKSUM.MD5 CHECKSUM.SHA256; \
Add SHA256 checksums to the release build. Requested by: cperciva (so) Reviewed/improved by: ru MFC after: 2 days
2005-09-28 12:39:00 +00:00
md5 * > .CHECKSUM.MD5; \
Fix a bug which puts an incorrect checksum of CHECKSUM.MD5 into the CHECKSUM.MD5 file itself. Approved by: re (scottl) MFC after: 3 days
2006-02-08 11:18:29 +00:00
sha256 * > .CHECKSUM.SHA256; \
mv .CHECKSUM.MD5 CHECKSUM.MD5; \
mv .CHECKSUM.SHA256 CHECKSUM.SHA256)
- Boot-only CD-ROM support (bootonly.iso). Submitted by: jhb, ru - Moved the creation of the ports distribution to release.7. - Call MFS root floppy mfsroot.flp even on a small PC98 disk. Submitted by: jhb Reviewed by: ru, jhb
2003-08-02 18:21:02 +00:00
@echo "ports distribution is finished."
.endif
Replaced the hardcoded target names in the touch(1) commands with ${.TARGET}, to ease maintenance.
2003-07-24 13:45:05 +00:00
touch ${.TARGET}
Add rule to generate the info files. Submitted by: gpalmer
1995-05-24 01:26:23 +00:00
A new cleaned up Makefile.
1996-01-11 20:28:40 +00:00
#
# --==## Make source dists ##==--
#
Crypto is now part of the "base" distribution, and the old "release.3" target no longer exists. Renumber the old release.[4-9] targets to release.[3-8]. Reminded by: ru
2004-08-06 08:49:51 +00:00
release.7:
YAMF22 Submitted by: pst
1997-06-24 23:08:18 +00:00
.if !defined(NOSRC)
Moved doFS into shell-script so we can be more smart about it. make two boot floppies, one without help files for 4MB systems. Mine's loading one now. :-) -O2 may be needed.
1996-01-08 22:13:50 +00:00
@cd ${.CURDIR} && $(MAKE) doTARBALL SD=/usr/src \
Add rule to generate the info files. Submitted by: gpalmer
1995-05-24 01:26:23 +00:00
TD=src TN=sbase ARG="[A-Z]*"
Support crypto and kerberos source dists.
1997-10-04 09:05:06 +00:00
@for i in `cd /usr/src && echo [a-z]*` ; do \
A new cleaned up Makefile.
1996-01-11 20:28:40 +00:00
if [ -d /usr/src/$$i ] ; then \
cd ${.CURDIR} && $(MAKE) doTARBALL \
release glue to support s/des/crypto/ Submitted by: markm
2000-02-29 11:26:24 +00:00
TN=`echo s$$i | tr -d '.' | \
sed -e 's/usr/u/' \
-e 's/kerberos5/krb5/'` \
A new cleaned up Makefile.
1996-01-11 20:28:40 +00:00
SD=/usr/src TD=src ARG="$$i" ; \
Add rule to generate the info files. Submitted by: gpalmer
1995-05-24 01:26:23 +00:00
fi ; \
done
Update the -current sources from the 2.1 branch. Approved (in spirit) by: jkh
1995-12-07 10:34:59 +00:00
.if defined(EXTRA_SRC)
Moved doFS into shell-script so we can be more smart about it. make two boot floppies, one without help files for 4MB systems. Mine's loading one now. :-) -O2 may be needed.
1996-01-08 22:13:50 +00:00
@set ${EXTRA_SRC} && \
Update the -current sources from the 2.1 branch. Approved (in spirit) by: jkh
1995-12-07 10:34:59 +00:00
while [ $$# -ge 2 ] ; do \
A new cleaned up Makefile.
1996-01-11 20:28:40 +00:00
if [ -d /usr/src/$$1 ] ; then \
cd ${.CURDIR} && $(MAKE) doTARBALL \
SD=/usr/src TN="s$$2" TD=src ARG="$$1" ; \
fi && shift && shift ; \
Update the -current sources from the 2.1 branch. Approved (in spirit) by: jkh
1995-12-07 10:34:59 +00:00
done
.endif
Add SHA256 checksums to the release build. Requested by: cperciva (so) Reviewed/improved by: ru MFC after: 2 days
2005-09-28 12:39:00 +00:00
(cd ${RD}/dists/src; \
Fix a bug which puts an incorrect checksum of CHECKSUM.MD5 into the CHECKSUM.MD5 file itself. Approved by: re (scottl) MFC after: 3 days
2006-02-08 11:18:29 +00:00
rm -f CHECKSUM.MD5 CHECKSUM.SHA256; \
Add SHA256 checksums to the release build. Requested by: cperciva (so) Reviewed/improved by: ru MFC after: 2 days
2005-09-28 12:39:00 +00:00
md5 * > .CHECKSUM.MD5; \
Fix a bug which puts an incorrect checksum of CHECKSUM.MD5 into the CHECKSUM.MD5 file itself. Approved by: re (scottl) MFC after: 3 days
2006-02-08 11:18:29 +00:00
sha256 * > .CHECKSUM.SHA256; \
mv .CHECKSUM.MD5 CHECKSUM.MD5; \
mv .CHECKSUM.SHA256 CHECKSUM.SHA256)
Add rule to generate the info files. Submitted by: gpalmer
1995-05-24 01:26:23 +00:00
@echo "src distribution is finished."
YAMF22 Submitted by: pst
1997-06-24 23:08:18 +00:00
.endif
Replaced the hardcoded target names in the touch(1) commands with ${.TARGET}, to ease maintenance.
2003-07-24 13:45:05 +00:00
touch ${.TARGET}
Add rule to generate the info files. Submitted by: gpalmer
1995-05-24 01:26:23 +00:00
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
# Build the memory root filesystem.
Crypto is now part of the "base" distribution, and the old "release.3" target no longer exists. Renumber the old release.[4-9] targets to release.[3-8]. Reminded by: ru
2004-08-06 08:49:51 +00:00
release.8:
Use the target arhitecture's disktab(5) file.
2003-03-21 13:41:17 +00:00
cp ${RD}/trees/base/etc/disktab /etc
Bring in Poul's patch to make the boot floppy more stand-alone in nature.
1995-05-26 21:47:13 +00:00
rm -rf ${RD}/mfsfd
mkdir ${RD}/mfsfd
cd ${RD}/mfsfd && \
Adapt the mfsroot environment of sysinstall to the needs of the new dhclient. This involves having passwd bits available so that seteuid("_dhcp") work, and creating /var/empty so that chroot(_VAR_EMPTY) works. My gut feeling is that the better solution is to make privsep and chroot optional in dhclient, but this works well for now and is low-risk. Approved by: re
2005-07-11 03:59:25 +00:00
mkdir -p etc/defaults dev mnt stand/etc/defaults stand/help \
var/empty
Reduce the size of the MFS filesystem by 250 Kbyte. Use && instead of ; a lot of places.
1995-05-28 07:04:36 +00:00
@cd ${.CURDIR} && $(MAKE) installCRUNCH CRUNCH=boot \
Shrink BOOTMFS kernel further. Make doFS.sh more intelligent. boot4.flp should work fine as far as I can tell.
1996-01-13 23:31:46 +00:00
DIR=${RD}/mfsfd/stand ZIP=false
Take a lesson from picoBSD: symlink /bin and /sbin to /stand in the mfsroot and install dhclient-script in stand/.
2000-07-26 04:28:21 +00:00
( cd ${RD}/mfsfd && \
for dir in bin sbin ; do \
ln -sf /stand $$dir; \
done )
Copy the right dhclient-script(8).
2002-07-24 20:37:21 +00:00
cp ${RD}/trees/base/sbin/dhclient-script ${RD}/mfsfd/stand
Adapt the mfsroot environment of sysinstall to the needs of the new dhclient. This involves having passwd bits available so that seteuid("_dhcp") work, and creating /var/empty so that chroot(_VAR_EMPTY) works. My gut feeling is that the better solution is to make privsep and chroot optional in dhclient, but this works well for now and is low-risk. Approved by: re
2005-07-11 03:59:25 +00:00
cp ${.CURDIR}/../etc/master.passwd ${RD}/mfsfd/etc/master.passwd
Be cross-arch friendly.
2005-11-03 07:27:50 +00:00
cp ${RD}/trees/base/etc/*pwd.db ${RD}/mfsfd/etc/
Fixes to unbreak Alpha releases: Strip ${RD}/mfsfd/stand/etc/defaults/rc.conf. Don't copy loader.help to ${RD}/mfsfd/boot, it serves no useful purpose. Removed the now no-op commands to copy ${RD}/trees/base/dev bits to ${RD}/mfsfd/dev, now that DEVFS is the default. Removed UFS_ACL, PSEUDOFS, and COMPAT_FREEBSD4 options from the BOOTMFS kernel. Moved miibus network module to the drivers floppy. isp(4) and mpt(4) now have corresponding modules, move them to the third floppy as well. Caution! With these changes, mfsroot.flp has only 11K left, and kern.flp has even less -- 5K.
2002-11-05 14:44:46 +00:00
( for F in defaults/rc.conf netconfig protocols ; do \
Strip comments from the MFSROOT's /stand/etc files.
2001-04-07 01:53:14 +00:00
sed -e '/^#.*$$/d' -e 's/[:space:]*#.*$$//g' \
Change the name of the 'bin' distribution to 'base'. This is done since it contains much more than /bin, and also gets in the way when making a combined install+fixit CD. OK'ed by: jkh
2002-04-23 22:16:41 +00:00
${RD}/trees/base/etc/$$F > ${RD}/mfsfd/stand/etc/$$F ; \
Strip comments from the MFSROOT's /stand/etc files.
2001-04-07 01:53:14 +00:00
done )
This adds /etc/netconfig to the list of files that get copied to /stand/etc on the mfsroot, and hence to /etc on the install target disk for use during the install. It also links netconfig and services into /etc on the mfsroot so that they are available to a fixit shell. Submitted by: iedowse
2001-03-21 23:53:40 +00:00
grep -E '^(ftp|nameserver|domain|sunrpc|cmd|nfsd)[^-\w]' \
Milestone #1 in cross-arch make releases. Do not install games and profiled libraries to the ${CHROOTDIR} with the initial installworld. Eliminate the need in the second installworld. For that, make sure _everything_ is built in the "world" environment, using the right tool chain. Added SUBDIR_OVERRIDE helper stuff to Makefile.inc1. Split the buildworld process into stages, and skip some stages when SUBDIR_OVERRIDE is set (used to build crypto, krb4, and krb5 dists). Added NO_MAKEDB_RUN knob to Makefile.inc1 to avoid running makewhatis(1) at the end of installworld (used when making crypto, krb4, and krb5 dists). In release/scripts/doFS.sh, ensure that the correct boot blocks are used. Moved the creation of the "crypto" dist from release.5 to release.2. In release.3 and doMFSKERN, build kernels in the "world" environment. KERNELS now means "additional" kernels, GENERIC is always built. Ensure we build crunched binaries in the "world" environment. Obfuscate release/Makefile some more (WMAKEENV) to achieve this. Inline createBOOTMFS target. Use already built GENERIC kernel modules to augment mfsfd's /stand/modules. GC doMODULES as such. Assorted fixes: Get rid of the "afterdistribute" target by moving the single use of it from sys/Makefile to etc/Makefile's "distribute". Makefile.inc1: apparently "etc" no longer needs to be last for "distribute" to succeed. gnu/usr.bin/perl/library/Makefile.inc: do not override the "install" and "distribute" targets, do it the "canonical" way. release/scripts/{man,cat}pages-make.sh: make sure Perl manpages and catpages appear in the right dists. Note that because Perl does not respect the MANBUILDCAT (and NOMAN), this results in a loss of /usr/share/perl/man/cat* empty directories. This will be fixed soon. Turn MAKE_KERBEROS4 into a plain boolean variable (if it is set it means "make KerberosIV"), as documented in the make.conf(5) manpage. Most of the userland makefiles did not test it for "YES" anyway. XXX Should specialized kerberized libpam versions be included into the krb4 and krb5 dists? (libpam.a would be incorrect anyway if both krb4 and krb5 dists were choosen.) Make sure "games" dist is made before "catpages", otherwise games catpages settle in the wrong dist. Fast build machine provided by: Igor Kucherenko <kivvy@sunbay.com>
2002-04-26 17:55:27 +00:00
${RD}/trees/base/etc/services | \
Strip comments from the MFSROOT's /stand/etc files.
2001-04-07 01:53:14 +00:00
sed -e '/^#.*$$/d' -e 's/[:space:]*#.*$$//g' \
> ${RD}/mfsfd/stand/etc/services
Generate minimal /etc/group file for mfsroot tree. This fixes newfs failures at installation. Reviewed by: re (scottl)
2003-11-15 03:58:12 +00:00
grep 'operator' ${RD}/trees/base/etc/group \
> ${RD}/mfsfd/stand/etc/group
This adds /etc/netconfig to the list of files that get copied to /stand/etc on the mfsroot, and hence to /etc on the install target disk for use during the install. It also links netconfig and services into /etc on the mfsroot so that they are available to a fixit shell. Submitted by: iedowse
2001-03-21 23:53:40 +00:00
ln ${RD}/mfsfd/stand/etc/services ${RD}/mfsfd/etc/services
Generate minimal /etc/group file for mfsroot tree. This fixes newfs failures at installation. Reviewed by: re (scottl)
2003-11-15 03:58:12 +00:00
ln ${RD}/mfsfd/stand/etc/group ${RD}/mfsfd/etc/group
This adds /etc/netconfig to the list of files that get copied to /stand/etc on the mfsroot, and hence to /etc on the install target disk for use during the install. It also links netconfig and services into /etc on the mfsroot so that they are available to a fixit shell. Submitted by: iedowse
2001-03-21 23:53:40 +00:00
ln ${RD}/mfsfd/stand/etc/netconfig ${RD}/mfsfd/etc/netconfig
Add a third floppy to releases. It will only be built if there is a release/${TARGET}/drivers.conf file which list drivers that have to go into the third floppy. Also shuffle i386/drivers.conf so that the floppies don't overflow anymore. Anybody with real/better usage statistics is welcome to shuffle it differently. Reviewed by: ru
2002-08-09 07:45:42 +00:00
cp ${RD}/trees/base/COPYRIGHT ${RD}/mfsfd/stand/help/COPYRIGHT.hlp
Fold the functionality of NORELNOTES into NODOC. There is now only one Makefile variable to control the building/installation of both the DocProj documents and RELNOTESng. Suggested by: obrien, dd Tested by: John Hay <jhay@icomtek.csir.co.za>
2001-05-30 17:05:15 +00:00
.if !defined(NODOC)
First commit of RELNOTESng, the rewrite of the *.TXT documentation files. src/release/doc/README has additional information. Reviewed by: -current, -doc
2001-04-27 21:28:02 +00:00
@for i in ${DIST_DOCS_ARCH_INDEP}; do \
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
cp ${RND}/${RELNOTES_LANG}/$$i/article.txt \
${RD}/mfsfd/stand/help/`echo $${i} | tr 'a-z' 'A-Z'`.TXT; \
First commit of RELNOTESng, the rewrite of the *.TXT documentation files. src/release/doc/README has additional information. Reviewed by: -current, -doc
2001-04-27 21:28:02 +00:00
done
@for i in ${DIST_DOCS_ARCH_DEP}; do \
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
cp ${RND}/${RELNOTES_LANG}/$$i/${TARGET}/article.txt \
${RD}/mfsfd/stand/help/`echo $${i} | tr 'a-z' 'A-Z'`.TXT; \
First commit of RELNOTESng, the rewrite of the *.TXT documentation files. src/release/doc/README has additional information. Reviewed by: -current, -doc
2001-04-27 21:28:02 +00:00
done
.endif
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
-test -f ${.CURDIR}/install.cfg \
&& cp ${.CURDIR}/install.cfg ${RD}/mfsfd
Copy only the boot1 and boot2 (if exists) files into mfs/boot since we don't need the others there.
1999-01-08 05:33:11 +00:00
@mkdir -p ${RD}/mfsfd/boot
PPC definitions required for 'make release' Reviewed by: ru Submitted by: ssouhlal
2004-08-18 11:11:46 +00:00
.if ${TARGET_ARCH} != "ia64" && ${TARGET_ARCH} != "powerpc"
Milestone #1 in cross-arch make releases. Do not install games and profiled libraries to the ${CHROOTDIR} with the initial installworld. Eliminate the need in the second installworld. For that, make sure _everything_ is built in the "world" environment, using the right tool chain. Added SUBDIR_OVERRIDE helper stuff to Makefile.inc1. Split the buildworld process into stages, and skip some stages when SUBDIR_OVERRIDE is set (used to build crypto, krb4, and krb5 dists). Added NO_MAKEDB_RUN knob to Makefile.inc1 to avoid running makewhatis(1) at the end of installworld (used when making crypto, krb4, and krb5 dists). In release/scripts/doFS.sh, ensure that the correct boot blocks are used. Moved the creation of the "crypto" dist from release.5 to release.2. In release.3 and doMFSKERN, build kernels in the "world" environment. KERNELS now means "additional" kernels, GENERIC is always built. Ensure we build crunched binaries in the "world" environment. Obfuscate release/Makefile some more (WMAKEENV) to achieve this. Inline createBOOTMFS target. Use already built GENERIC kernel modules to augment mfsfd's /stand/modules. GC doMODULES as such. Assorted fixes: Get rid of the "afterdistribute" target by moving the single use of it from sys/Makefile to etc/Makefile's "distribute". Makefile.inc1: apparently "etc" no longer needs to be last for "distribute" to succeed. gnu/usr.bin/perl/library/Makefile.inc: do not override the "install" and "distribute" targets, do it the "canonical" way. release/scripts/{man,cat}pages-make.sh: make sure Perl manpages and catpages appear in the right dists. Note that because Perl does not respect the MANBUILDCAT (and NOMAN), this results in a loss of /usr/share/perl/man/cat* empty directories. This will be fixed soon. Turn MAKE_KERBEROS4 into a plain boolean variable (if it is set it means "make KerberosIV"), as documented in the make.conf(5) manpage. Most of the userland makefiles did not test it for "YES" anyway. XXX Should specialized kerberized libpam versions be included into the krb4 and krb5 dists? (libpam.a would be incorrect anyway if both krb4 and krb5 dists were choosen.) Make sure "games" dist is made before "catpages", otherwise games catpages settle in the wrong dist. Fast build machine provided by: Igor Kucherenko <kivvy@sunbay.com>
2002-04-26 17:55:27 +00:00
@cp ${RD}/trees/base/boot/boot* ${RD}/mfsfd/boot
Add the necessary tweaks for FreeBSD/ia64 releases. Note that this is in no way final. A typical ia64 wart is that there are no boot blocks. Instead, we need to create disks with EFI partitions if we want auto boot to work. All this functionality is not present yet.
2002-11-04 00:50:01 +00:00
.endif
TARGET_ARCH == amd64 also has /boot/mbr.
2003-06-04 22:24:43 +00:00
.if ${TARGET} == "i386" || ${TARGET_ARCH} == "amd64"
Milestone #1 in cross-arch make releases. Do not install games and profiled libraries to the ${CHROOTDIR} with the initial installworld. Eliminate the need in the second installworld. For that, make sure _everything_ is built in the "world" environment, using the right tool chain. Added SUBDIR_OVERRIDE helper stuff to Makefile.inc1. Split the buildworld process into stages, and skip some stages when SUBDIR_OVERRIDE is set (used to build crypto, krb4, and krb5 dists). Added NO_MAKEDB_RUN knob to Makefile.inc1 to avoid running makewhatis(1) at the end of installworld (used when making crypto, krb4, and krb5 dists). In release/scripts/doFS.sh, ensure that the correct boot blocks are used. Moved the creation of the "crypto" dist from release.5 to release.2. In release.3 and doMFSKERN, build kernels in the "world" environment. KERNELS now means "additional" kernels, GENERIC is always built. Ensure we build crunched binaries in the "world" environment. Obfuscate release/Makefile some more (WMAKEENV) to achieve this. Inline createBOOTMFS target. Use already built GENERIC kernel modules to augment mfsfd's /stand/modules. GC doMODULES as such. Assorted fixes: Get rid of the "afterdistribute" target by moving the single use of it from sys/Makefile to etc/Makefile's "distribute". Makefile.inc1: apparently "etc" no longer needs to be last for "distribute" to succeed. gnu/usr.bin/perl/library/Makefile.inc: do not override the "install" and "distribute" targets, do it the "canonical" way. release/scripts/{man,cat}pages-make.sh: make sure Perl manpages and catpages appear in the right dists. Note that because Perl does not respect the MANBUILDCAT (and NOMAN), this results in a loss of /usr/share/perl/man/cat* empty directories. This will be fixed soon. Turn MAKE_KERBEROS4 into a plain boolean variable (if it is set it means "make KerberosIV"), as documented in the make.conf(5) manpage. Most of the userland makefiles did not test it for "YES" anyway. XXX Should specialized kerberized libpam versions be included into the krb4 and krb5 dists? (libpam.a would be incorrect anyway if both krb4 and krb5 dists were choosen.) Make sure "games" dist is made before "catpages", otherwise games catpages settle in the wrong dist. Fast build machine provided by: Igor Kucherenko <kivvy@sunbay.com>
2002-04-26 17:55:27 +00:00
@cp ${RD}/trees/base/boot/mbr ${RD}/mfsfd/boot
Teach sysinstall(8) how to read boot managers out of /boot instead of hardcoding them into the binary. This replaces the work-around in usr.sbin/sysinstall/Makefile,v 1.121. Suggested by: jhb MFC in: 1 week
2002-03-28 08:23:33 +00:00
.endif
Milestone #1 in cross-arch make releases. Do not install games and profiled libraries to the ${CHROOTDIR} with the initial installworld. Eliminate the need in the second installworld. For that, make sure _everything_ is built in the "world" environment, using the right tool chain. Added SUBDIR_OVERRIDE helper stuff to Makefile.inc1. Split the buildworld process into stages, and skip some stages when SUBDIR_OVERRIDE is set (used to build crypto, krb4, and krb5 dists). Added NO_MAKEDB_RUN knob to Makefile.inc1 to avoid running makewhatis(1) at the end of installworld (used when making crypto, krb4, and krb5 dists). In release/scripts/doFS.sh, ensure that the correct boot blocks are used. Moved the creation of the "crypto" dist from release.5 to release.2. In release.3 and doMFSKERN, build kernels in the "world" environment. KERNELS now means "additional" kernels, GENERIC is always built. Ensure we build crunched binaries in the "world" environment. Obfuscate release/Makefile some more (WMAKEENV) to achieve this. Inline createBOOTMFS target. Use already built GENERIC kernel modules to augment mfsfd's /stand/modules. GC doMODULES as such. Assorted fixes: Get rid of the "afterdistribute" target by moving the single use of it from sys/Makefile to etc/Makefile's "distribute". Makefile.inc1: apparently "etc" no longer needs to be last for "distribute" to succeed. gnu/usr.bin/perl/library/Makefile.inc: do not override the "install" and "distribute" targets, do it the "canonical" way. release/scripts/{man,cat}pages-make.sh: make sure Perl manpages and catpages appear in the right dists. Note that because Perl does not respect the MANBUILDCAT (and NOMAN), this results in a loss of /usr/share/perl/man/cat* empty directories. This will be fixed soon. Turn MAKE_KERBEROS4 into a plain boolean variable (if it is set it means "make KerberosIV"), as documented in the make.conf(5) manpage. Most of the userland makefiles did not test it for "YES" anyway. XXX Should specialized kerberized libpam versions be included into the krb4 and krb5 dists? (libpam.a would be incorrect anyway if both krb4 and krb5 dists were choosen.) Make sure "games" dist is made before "catpages", otherwise games catpages settle in the wrong dist. Fast build machine provided by: Igor Kucherenko <kivvy@sunbay.com>
2002-04-26 17:55:27 +00:00
@tar --exclude CVS -cf - -C ${.CURDIR}/../usr.sbin/sysinstall help | \
Hammer on the boot floppy generation bits some more. This is looking a little better.
1999-01-05 03:44:08 +00:00
tar xf - -C ${RD}/mfsfd/stand
- Boot-only CD-ROM support (bootonly.iso). Submitted by: jhb, ru - Moved the creation of the ports distribution to release.7. - Call MFS root floppy mfsroot.flp even on a small PC98 disk. Submitted by: jhb Reviewed by: ru, jhb
2003-08-02 18:21:02 +00:00
@mkdir -p ${RD}/mfsroot
- Eliminate the "rm" command that was causing us trouble in rev. 1.799. - Normalize doRELEASE to avoid seeing "`ftp.1' is up to date." warning when doing "make rerelease", and ftp.1 was already built. Reviewed by: jhb
2003-08-07 18:22:34 +00:00
sh -e ${DOFS_SH} ${RD}/mfsroot/mfsroot ${RD} ${MNT} \
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
${MFSSIZE} ${RD}/mfsfd ${MFSINODE} ${MFSLABEL}
- Tweak order of arch fs options to stick all floppy options together and all MFS root options together and in the same order on all archs. - Drop support for using the no longer existing 2.88 floppy image for CD booting on i386 via the EMUL_BOOT variable. - Use a stock loader on the boot floppy rather than one without forth. - Force the gzip of the mfsroot in release.9 to succeed. This is useful when doing lots of rereleases when testing floppy changes. - If an acpi kernel module exists, stick it on the boot floppy and for now always load it after the kernel.
2004-01-29 19:01:58 +00:00
@gzip -9fnv ${RD}/mfsroot/mfsroot
Moved the boot floppy generation code out from release.9 to release.10. The release.9 target is now responsible only for generation of MFS root file systems, that are built for all architectures, even those that do not provide a floppy installation option. The release.10 target is now responsible for creation of a set of boot, MFS root, and fixit floppies, and the NO_FLOPPIES variable now affects only this target. Also, replaced the FIXIT_TARGET variable with a check of whether the *FIXITSIZE variables are present for a given architecture, similar to how this is done when generating boot floppies. Discussed with: jhb
2003-07-25 08:54:41 +00:00
touch ${.TARGET}
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
KERNFLOPPYSET= ${RD}/floppyset/kern/kernel.gz
.if defined(SMALLFLOPPYSIZE)
SMALLKERNFLOPPYSET= ${RD}/floppyset/kern-small/kernel.gz
Add support for building sparc64 releases. Reviewed by: murray
2002-10-16 05:30:56 +00:00
.endif
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
.if defined(SPLIT_MFSROOT)
MFSROOTFLOPPYSET= ${RD}/floppyset/mfsroot/mfsroot.gz
.if defined(SMALLFLOPPYSIZE)
SMALLMFSROOTFLOPPYSET= ${RD}/floppyset/mfsroot-small/mfsroot.gz
Make floppy-making part of release.9 conditional on *BOOTSIZE rather than on "pc98" (rev. 1.710) and "sparc64" (rev. 1.711).
2002-10-21 14:21:08 +00:00
.endif
- Copy the mfsroot.gz file to /R/stage/mfsroot/ so that the CD-ROM layout can fetch it from there instead of looking in the boot floppy's directory. - Add a NO_FLOPPIES variable that disables building of any boot floppies. Tested on: alpha
2002-11-01 20:53:50 +00:00
.endif
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
# Build boot and install floppies.
floppies.1:
Revamp base system packaging of kernels to enable up/smp selection at runtime and to support distributing additional kernels: o remove kernel from the base tarball o add new kernel tarballs o build + package both SMP and GENERIC kernels when an <arch>/conf/SMP config file is present o add sysinstall support for multiple kernels o update sysinstall to probe for the number of cpus on a system and auto-select smp/up kernel accordingly o add a post-kernels install hook to fixup /boot/kernel o add -ldevinfo to boot crunch for sysinstall's cpu probing logic Notes: 1. On HEAD this code is not currently used because GENERIC kernels include SMP. This work is mainly intended for RELENG_6 where the GENERIC kernel is UP. If HEAD changes to match then just enable WITH_SMP in sysinstall/Makefile. 2. The cpu probing support is done with acpi and MPTable; this means some systems will require work for auto-detection to work. 3. The handling of /boot/kernel may need to be revisited; for now we rename one kernel at the last moment (SMP if installed, otherwise GENERIC). There are other, possibly better, approaches. Lots of help from ru, emaste, scottl, and jhb.
2006-03-08 18:02:32 +00:00
@gzip -9nc ${RD}/kernels/GENERIC/kernel > ${RD}/kernels/kernel.gz
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
@echo "Making the kernel boot floppies..."
@cd ${.CURDIR} && ${MAKE} makeFloppySet FLOPPYBASE=kern \
FLOPPYDESC="Kernel" SPLITFILE=${RD}/kernels/kernel.gz
.if defined(SMALLFLOPPYSIZE)
@echo "Making the small kernel boot floppies..."
@cd ${.CURDIR} && ${MAKE} makeFloppySet FLOPPYBASE=kern-small \
FLOPPYDESC="Kernel" SPLITFILE=${RD}/kernels/kernel.gz \
FDSIZE="SMALL"
.endif
.if defined(SPLIT_MFSROOT)
@echo "Making the mfsroot boot floppies..."
@cd ${.CURDIR} && ${MAKE} makeFloppySet FLOPPYBASE=mfsroot \
FLOPPYDESC="Memory Filesystem" SPLITFILE=${RD}/mfsroot/mfsroot.gz
.if defined(SMALLFLOPPYSIZE)
@echo "Making the small mfsroot boot floppies..."
@cd ${.CURDIR} && ${MAKE} makeFloppySet FLOPPYBASE=mfsroot-small \
FLOPPYDESC="Memory Filesystem" SPLITFILE=${RD}/mfsroot/mfsroot.gz \
FDSIZE="SMALL"
.endif
@cd ${.CURDIR} && ${MAKE} buildBootFloppy FSIMAGE="boot" \
KERNFILE="${KERNFLOPPYSET}.split ${KERNFLOPPYSET}.boot" \
MFSROOTFILE="${MFSROOTFLOPPYSET}.split ${MFSROOTFLOPPYSET}.boot"
.if defined(SMALLFLOPPYSIZE)
The name of small boot floppy should be boot-small.flp to avoid conflicting normal size floppy. Reviewed by: jhb
2004-01-29 15:14:59 +00:00
@cd ${.CURDIR} && ${MAKE} buildBootFloppy FSIMAGE="boot-small" \
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
KERNFILE="${SMALLKERNFLOPPYSET}.split ${SMALLKERNFLOPPYSET}.boot" \
MFSROOTFILE="${SMALLMFSROOTFLOPPYSET}.split ${SMALLMFSROOTFLOPPYSET}.boot" \
FDSIZE="SMALL"
.endif
.else # !SPLIT_MFSROOT
@cd ${.CURDIR} && ${MAKE} buildBootFloppy FSIMAGE="boot" \
KERNFILE="${KERNFLOPPYSET}.split ${KERNFLOPPYSET}.boot" \
MFSROOTFILE=${RD}/mfsroot/mfsroot.gz
.if defined(SMALLFLOPPYSIZE)
The name of small boot floppy should be boot-small.flp to avoid conflicting normal size floppy. Reviewed by: jhb
2004-01-29 15:14:59 +00:00
@cd ${.CURDIR} && ${MAKE} buildBootFloppy FSIMAGE="boot-small" \
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
KERNFILE="${SMALLKERNFLOPPYSET}.split ${SMALLKERNFLOPPYSET}.boot" \
MFSROOTFILE=${RD}/mfsroot/mfsroot.gz \
FDSIZE="SMALL"
.endif
.endif # SPLIT_MFSROOT
touch ${.TARGET}
# Build fixit floppy.
floppies.2:
Hammer on the boot floppy generation bits some more. This is looking a little better.
1999-01-05 03:44:08 +00:00
@echo "Making fixit floppy."
@rm -rf ${RD}/fixitfd
@mkdir ${RD}/fixitfd
@cd ${RD}/fixitfd && \
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
mkdir -p dev stand bin sbin etc mnt mnt1 mnt2 mnt3 mnt4 tmp \
usr/share/misc
Change the name of the 'bin' distribution to 'base'. This is done since it contains much more than /bin, and also gets in the way when making a combined install+fixit CD. OK'ed by: jkh
2002-04-23 22:16:41 +00:00
@cp ${RD}/trees/base/etc/spwd.db ${RD}/trees/base/etc/group \
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
${RD}/fixitfd/etc
Removing comment in some text files for fixit floppy overflow. Approved by: re (murray)
2003-04-25 06:08:11 +00:00
@sed -e 's/#.*//' ${RD}/trees/base/etc/protocols \
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
> ${RD}/fixitfd/etc/protocols
Removing comment in some text files for fixit floppy overflow. Approved by: re (murray)
2003-04-25 06:08:11 +00:00
@sed -e 's/#.*//' ${RD}/trees/base/usr/share/misc/scsi_modes \
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
> ${RD}/fixitfd/usr/share/misc/scsi_modes
Hammer on the boot floppy generation bits some more. This is looking a little better.
1999-01-05 03:44:08 +00:00
@cp ${.CURDIR}/fixit.profile ${RD}/fixitfd/.profile
@cp ${.CURDIR}/fixit.services ${RD}/fixitfd/etc/services
@cp ${.CURDIR}/scripts/tar.sh ${RD}/fixitfd/stand/tar
@chmod 555 ${RD}/fixitfd/stand/tar
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
.if defined(SMALLFLOPPYSIZE)
Make a small (1.2M) fixit floppy image for pc98.
2002-11-05 13:54:58 +00:00
@cd ${.CURDIR} && ${MAKE} installCRUNCH CRUNCH=fixit-small \
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
DIR=${RD}/fixitfd/stand ZIP=false
@sh -e ${DOFS_SH} ${RD}/floppies/fixit-small.flp ${RD} ${MNT} \
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
${SMALLFLOPPYSIZE} ${RD}/fixitfd ${FLOPPYINODE} ${SMALLFLOPPYLABEL}
Make a small (1.2M) fixit floppy image for pc98.
2002-11-05 13:54:58 +00:00
@rm -rf ${RD}/fixitfd/stand
@mkdir ${RD}/fixitfd/stand
.endif
@cd ${.CURDIR} && ${MAKE} installCRUNCH CRUNCH=fixit \
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
DIR=${RD}/fixitfd/stand ZIP=false
@sh -e ${DOFS_SH} ${RD}/floppies/fixit.flp ${RD} ${MNT} \
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
${FLOPPYSIZE} ${RD}/fixitfd ${FLOPPYINODE} ${FLOPPYLABEL}
touch ${.TARGET}
# Do our last minute floppies directory setup
floppies.3:
Unbreak make release (didn't completely account for removal of *.TXT files). Another diff will be forthcoming to fold the functionality of NORELNOTES into NODOC. Tested by: John Hay <jhay@icomtek.csir.co.za> Pointy hat to: bmah
2001-05-29 17:14:59 +00:00
@(cd ${RD}/floppies; md5 *.flp > CHECKSUM.MD5)
Add SHA256 checksums to the release build. Requested by: cperciva (so) Reviewed/improved by: ru MFC after: 2 days
2005-09-28 12:39:00 +00:00
@(cd ${RD}/floppies; sha256 *.flp > CHECKSUM.SHA256)
Replaced the hardcoded target names in the touch(1) commands with ${.TARGET}, to ease maintenance.
2003-07-24 13:45:05 +00:00
touch ${.TARGET}
Added a fixit floppy to the rules.
1995-03-13 20:09:38 +00:00
A new cleaned up Makefile.
1996-01-11 20:28:40 +00:00
#
# --==## Setup a suitable ftp-area ##==--
#
I belive the floppies work as expected now. (incl the fixit)
1995-03-15 06:14:19 +00:00
ftp.1:
Hammer on the boot floppy generation bits some more. This is looking a little better.
1999-01-05 03:44:08 +00:00
@echo "Setting up FTP distribution area"
@mkdir -p ${FD}
Ack, that will teach me to grep for "symbolic link" as well as "symlink" in the commit logs - I just found the reason for the self-pointing symlink, as documented in revision 1.517 by phk, who committed the change over a year ago. Accordingly, put the feature back and drop all plans to MFC the previous "fix".
2001-02-06 08:36:00 +00:00
-@ln -s . ${FD}/${BUILDNAME}
Follow 'MAKE_ISOS' in naming and s/FLOPPIES/MAKE_FLOPPIES/g Submitted by: ru
2004-08-16 08:56:50 +00:00
.if defined(MAKE_FLOPPIES)
Hammer on the boot floppy generation bits some more. This is looking a little better.
1999-01-05 03:44:08 +00:00
@cd ${RD} && find floppies -print | cpio -dumpl ${FD}
Make NO_FLOPPIES really mean "no floppies", including not even creating the "floppies" directory on FTP and CD-ROMs. Enforce NO_FLOPPIES on amd64. Tested for: amd64, i386 with -DNO_FLOPPIES
2003-07-23 12:12:53 +00:00
.endif
Hammer on the boot floppy generation bits some more. This is looking a little better.
1999-01-05 03:44:08 +00:00
@cd ${RD}/dists && find . -print | cpio -dumpl ${FD}
Fold the functionality of NORELNOTES into NODOC. There is now only one Makefile variable to control the building/installation of both the DocProj documents and RELNOTESng. Suggested by: obrien, dd Tested by: John Hay <jhay@icomtek.csir.co.za>
2001-05-30 17:05:15 +00:00
.if !defined(NODOC)
First commit of RELNOTESng, the rewrite of the *.TXT documentation files. src/release/doc/README has additional information. Reviewed by: -current, -doc
2001-04-27 21:28:02 +00:00
@for i in ${DIST_DOCS_ARCH_INDEP}; do \
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
cp ${RND}/${RELNOTES_LANG}/$$i/article.txt \
${FD}/`echo $${i} | tr 'a-z' 'A-Z'`.TXT; \
cp ${RND}/${RELNOTES_LANG}/$$i/article.html \
${FD}/`echo $${i} | tr 'a-z' 'A-Z'`.HTM; \
First commit of RELNOTESng, the rewrite of the *.TXT documentation files. src/release/doc/README has additional information. Reviewed by: -current, -doc
2001-04-27 21:28:02 +00:00
done
@for i in ${DIST_DOCS_ARCH_DEP}; do \
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
cp ${RND}/${RELNOTES_LANG}/$$i/${TARGET}/article.txt \
${FD}/`echo $${i} | tr 'a-z' 'A-Z'`.TXT; \
cp ${RND}/${RELNOTES_LANG}/$$i/${TARGET}/article.html \
${FD}/`echo $${i} | tr 'a-z' 'A-Z'`.HTM; \
First commit of RELNOTESng, the rewrite of the *.TXT documentation files. src/release/doc/README has additional information. Reviewed by: -current, -doc
2001-04-27 21:28:02 +00:00
done
Modify the release-building process so the srelease distribution doesn't get spammed by RELNOTESng files. Previously, we built the RELNOTESng files "in place" and then populated the media (e.g. floppy, cdrom, ftp) areas from the build directories. However, the distributions got rolled after the RELNOTESng build, and thus picked up the rendered RELNOTESng files. Now, we install the RELNOTESng documents to R/stage/release.doc, then clean out the original directories in src/release/doc. We use the renditions in the staging area to populate the media areas. PR: 33141 MFC after: 3 days (plus re approval)
2001-12-28 17:21:28 +00:00
@cp ${RND}/${RELNOTES_LANG}/readme/docbook.css ${FD}
First commit of RELNOTESng, the rewrite of the *.TXT documentation files. src/release/doc/README has additional information. Reviewed by: -current, -doc
2001-04-27 21:28:02 +00:00
.endif
Hammer on the boot floppy generation bits some more. This is looking a little better.
1999-01-05 03:44:08 +00:00
@echo "CD_VERSION = ${BUILDNAME}" > ${FD}/cdrom.inf
Replaced the hardcoded target names in the touch(1) commands with ${.TARGET}, to ease maintenance.
2003-07-24 13:45:05 +00:00
touch ${.TARGET}
Added a fixit floppy to the rules.
1995-03-13 20:09:38 +00:00
Rework how we layout the CD/DVD ISO images for releases. Note that this is only the first stage and does not yet handle packages (which will move to their own disc(s)). - By default, include the live filesystem on disc1, so that disc1 is now both an install disc and a fixit disc. The images can still be split into two separate discs by specifying SEPARATE_LIVEFS=yes. - Remove /usr/ports from the live filesystem disc as the merged images do not fit on even i386 otherwise. Also, the merged image already contains a tarball of the entire ports tree, so /usr/ports was a duplicate anyway. - Stop building port READMES for the release sinc they are no longer used (/usr/ports in the live filesystem was the last remaining user). - Split cdrom.1 into three targets. cdrom.1 builds the live filesystem image. cdrom.2 builds the disc1 image. Note that if SEPARATE_LIVEFS is not specified these both write to the same area. cdrom.3 builds the bootonly image and the UFS miniroot. - The various distribution directories are now in a <buildname> subdirectory of disc1 rather than being in the root so that the merged image's root is not quite as cluttered. - The disc1 image is now called disc1.iso and we no longer build a miniinst.iso image. If CD_EXTRA_BITS is defined, then the iso.1 target will simply include those bits in the disc1 image that it builds. MFC after: 1 day Requested by: re (scottl, etc.)
2005-03-02 22:46:07 +00:00
# Build a live filesystem cdrom image
Remove imutable bits, they will not survive anyway. First cut at cdrom make rule.
1995-03-16 22:54:42 +00:00
cdrom.1:
Rework how we layout the CD/DVD ISO images for releases. Note that this is only the first stage and does not yet handle packages (which will move to their own disc(s)). - By default, include the live filesystem on disc1, so that disc1 is now both an install disc and a fixit disc. The images can still be split into two separate discs by specifying SEPARATE_LIVEFS=yes. - Remove /usr/ports from the live filesystem disc as the merged images do not fit on even i386 otherwise. Also, the merged image already contains a tarball of the entire ports tree, so /usr/ports was a duplicate anyway. - Stop building port READMES for the release sinc they are no longer used (/usr/ports in the live filesystem was the last remaining user). - Split cdrom.1 into three targets. cdrom.1 builds the live filesystem image. cdrom.2 builds the disc1 image. Note that if SEPARATE_LIVEFS is not specified these both write to the same area. cdrom.3 builds the bootonly image and the UFS miniroot. - The various distribution directories are now in a <buildname> subdirectory of disc1 rather than being in the root so that the merged image's root is not quite as cluttered. - The disc1 image is now called disc1.iso and we no longer build a miniinst.iso image. If CD_EXTRA_BITS is defined, then the iso.1 target will simply include those bits in the disc1 image that it builds. MFC after: 1 day Requested by: re (scottl, etc.)
2005-03-02 22:46:07 +00:00
@echo "Building CDROM live filesystem image"
@mkdir -p ${CD_LIVEFS}
Hammer on the boot floppy generation bits some more. This is looking a little better.
1999-01-05 03:44:08 +00:00
@for i in ${DISTRIBUTIONS} ; \
Remove imutable bits, they will not survive anyway. First cut at cdrom make rule.
1995-03-16 22:54:42 +00:00
do \
if [ -d ${RD}/trees/$${i} ] ; then \
Ignore errors from chflags. This makes it possible to make installworld with DESTDIR set to an NFS-mounted file system.
1999-02-14 13:56:15 +00:00
chflags -R noschg ${RD}/trees/$${i} || true ; \
Sanitize the making of kernels a bit, and make the BOOTCDROM kernel too. If we can get away with it, it will always just be the GENERIC kernel with a different name.
1995-04-10 08:01:27 +00:00
( cd ${RD}/trees/$${i} && \
Rework how we layout the CD/DVD ISO images for releases. Note that this is only the first stage and does not yet handle packages (which will move to their own disc(s)). - By default, include the live filesystem on disc1, so that disc1 is now both an install disc and a fixit disc. The images can still be split into two separate discs by specifying SEPARATE_LIVEFS=yes. - Remove /usr/ports from the live filesystem disc as the merged images do not fit on even i386 otherwise. Also, the merged image already contains a tarball of the entire ports tree, so /usr/ports was a duplicate anyway. - Stop building port READMES for the release sinc they are no longer used (/usr/ports in the live filesystem was the last remaining user). - Split cdrom.1 into three targets. cdrom.1 builds the live filesystem image. cdrom.2 builds the disc1 image. Note that if SEPARATE_LIVEFS is not specified these both write to the same area. cdrom.3 builds the bootonly image and the UFS miniroot. - The various distribution directories are now in a <buildname> subdirectory of disc1 rather than being in the root so that the merged image's root is not quite as cluttered. - The disc1 image is now called disc1.iso and we no longer build a miniinst.iso image. If CD_EXTRA_BITS is defined, then the iso.1 target will simply include those bits in the disc1 image that it builds. MFC after: 1 day Requested by: re (scottl, etc.)
2005-03-02 22:46:07 +00:00
find . -depth -print | cpio -dumpl ${CD_LIVEFS} ) ; \
Remove imutable bits, they will not survive anyway. First cut at cdrom make rule.
1995-03-16 22:54:42 +00:00
fi \
done
Revamp base system packaging of kernels to enable up/smp selection at runtime and to support distributing additional kernels: o remove kernel from the base tarball o add new kernel tarballs o build + package both SMP and GENERIC kernels when an <arch>/conf/SMP config file is present o add sysinstall support for multiple kernels o update sysinstall to probe for the number of cpus on a system and auto-select smp/up kernel accordingly o add a post-kernels install hook to fixup /boot/kernel o add -ldevinfo to boot crunch for sysinstall's cpu probing logic Notes: 1. On HEAD this code is not currently used because GENERIC kernels include SMP. This work is mainly intended for RELENG_6 where the GENERIC kernel is UP. If HEAD changes to match then just enable WITH_SMP in sysinstall/Makefile. 2. The cpu probing support is done with acpi and MPTable; this means some systems will require work for auto-detection to work. 3. The handling of /boot/kernel may need to be revisited; for now we rename one kernel at the last moment (SMP if installed, otherwise GENERIC). There are other, possibly better, approaches. Lots of help from ru, emaste, scottl, and jhb.
2006-03-08 18:02:32 +00:00
@echo "Copy GENERIC kernel to boot area"
@cp -Rp ${RD}/kernels/GENERIC/ ${CD_LIVEFS}/boot/kernel
Remove kernel and module symbol files from disc1 (and the livecd if it is enabled). This saves around 80MB (for i386) on the ISO images. Note that this is only from the install media, not the installed system where the symbols are still installed, if part of the release. Should anybody want the symbols which match disc1/livecd, they can just be extracted from the kernel distribution files. Reviewed by: kensmith
2007-11-22 20:06:50 +00:00
@rm -f ${CD_LIVEFS}/boot/kernel/*.symbols
Rework how we layout the CD/DVD ISO images for releases. Note that this is only the first stage and does not yet handle packages (which will move to their own disc(s)). - By default, include the live filesystem on disc1, so that disc1 is now both an install disc and a fixit disc. The images can still be split into two separate discs by specifying SEPARATE_LIVEFS=yes. - Remove /usr/ports from the live filesystem disc as the merged images do not fit on even i386 otherwise. Also, the merged image already contains a tarball of the entire ports tree, so /usr/ports was a duplicate anyway. - Stop building port READMES for the release sinc they are no longer used (/usr/ports in the live filesystem was the last remaining user). - Split cdrom.1 into three targets. cdrom.1 builds the live filesystem image. cdrom.2 builds the disc1 image. Note that if SEPARATE_LIVEFS is not specified these both write to the same area. cdrom.3 builds the bootonly image and the UFS miniroot. - The various distribution directories are now in a <buildname> subdirectory of disc1 rather than being in the root so that the merged image's root is not quite as cluttered. - The disc1 image is now called disc1.iso and we no longer build a miniinst.iso image. If CD_EXTRA_BITS is defined, then the iso.1 target will simply include those bits in the disc1 image that it builds. MFC after: 1 day Requested by: re (scottl, etc.)
2005-03-02 22:46:07 +00:00
@rm -f ${CD_LIVEFS}/.profile
@cp ${.CURDIR}/fixit.profile ${CD_LIVEFS}/.profile
The live CD itself no longer has /stand since a long time ago; however, there are still a couple of places under src/release such like fixit.profile assumed that system binary can be retrivied from /stand(ex: boot with the live CD and run "disklabel -e" in the fixit CD shell). Since /stand is still functional in fixit floppy, and there are more than one places in src/release needs to be updated(document for example) if we want to make use of something like /mnt2/rescue/vi. This commit try to deal with aforementioned inconsistency with minimal effort by simply create a symbolic link to /rescue. Reviewed by: jhb, sam (mentor)
2005-03-23 02:51:19 +00:00
@ln -sf /rescue ${CD_LIVEFS}/stand
Rework how we layout the CD/DVD ISO images for releases. Note that this is only the first stage and does not yet handle packages (which will move to their own disc(s)). - By default, include the live filesystem on disc1, so that disc1 is now both an install disc and a fixit disc. The images can still be split into two separate discs by specifying SEPARATE_LIVEFS=yes. - Remove /usr/ports from the live filesystem disc as the merged images do not fit on even i386 otherwise. Also, the merged image already contains a tarball of the entire ports tree, so /usr/ports was a duplicate anyway. - Stop building port READMES for the release sinc they are no longer used (/usr/ports in the live filesystem was the last remaining user). - Split cdrom.1 into three targets. cdrom.1 builds the live filesystem image. cdrom.2 builds the disc1 image. Note that if SEPARATE_LIVEFS is not specified these both write to the same area. cdrom.3 builds the bootonly image and the UFS miniroot. - The various distribution directories are now in a <buildname> subdirectory of disc1 rather than being in the root so that the merged image's root is not quite as cluttered. - The disc1 image is now called disc1.iso and we no longer build a miniinst.iso image. If CD_EXTRA_BITS is defined, then the iso.1 target will simply include those bits in the disc1 image that it builds. MFC after: 1 day Requested by: re (scottl, etc.)
2005-03-02 22:46:07 +00:00
@echo "Setting up CDROM boot area"
@rm -f ${CD_LIVEFS}/boot/loader.conf
@cp ${RD}/mfsroot/mfsroot.gz ${CD_LIVEFS}/boot/mfsroot.gz
@echo 'mfsroot_load="YES"' > ${CD_LIVEFS}/boot/loader.conf
@echo 'mfsroot_type="mfs_root"' >> ${CD_LIVEFS}/boot/loader.conf
@echo 'mfsroot_name="/boot/mfsroot"' >> ${CD_LIVEFS}/boot/loader.conf
.if exists(${RD}/trees/base/boot/device.hints)
# Break the link to device.hints so we can modify it
@rm -f ${CD_LIVEFS}/boot/device.hints
@cp ${RD}/trees/base/boot/device.hints ${CD_LIVEFS}/boot/device.hints
.endif
@echo "CD_VERSION = ${BUILDNAME}" > ${CD_LIVEFS}/cdrom.inf
touch ${.TARGET}
What should be the last major changes to the ISO layout: - When a separate livefs ISO is used, change the name of the tree directory to be R/cdrom/livefs and change the name of the iso to livefs.iso rather than using disc2. Instead, disc2 is now always going to be the packages disc regardless of the presence of a separate livefs which makes things simpler. - Build a very simple disc2 tree under R/cdrom that just includes the cdrom.inf file that the packages need. Also, build a disc2.iso image in the iso.1 target. Disc 2's volume label is "FreeBSD_Packages". - Retire CD_EXTRA_BITS and replace it with CD_PACKAGE_TREE. CD_EXTRA_BITS was specific to disc1, but CD_PACKAGE_TREE instead should point to a directory that has disc1 and disc2 subdirectories that contain the extra bits for each image. - Rename the bootonly volume label to "FreeBSD_bootonly" and just hardcode the disc1 and livefs volume names and iso suffixes.
2005-03-23 15:01:54 +00:00
# Build disc1 and disc2 cdrom images
Rework how we layout the CD/DVD ISO images for releases. Note that this is only the first stage and does not yet handle packages (which will move to their own disc(s)). - By default, include the live filesystem on disc1, so that disc1 is now both an install disc and a fixit disc. The images can still be split into two separate discs by specifying SEPARATE_LIVEFS=yes. - Remove /usr/ports from the live filesystem disc as the merged images do not fit on even i386 otherwise. Also, the merged image already contains a tarball of the entire ports tree, so /usr/ports was a duplicate anyway. - Stop building port READMES for the release sinc they are no longer used (/usr/ports in the live filesystem was the last remaining user). - Split cdrom.1 into three targets. cdrom.1 builds the live filesystem image. cdrom.2 builds the disc1 image. Note that if SEPARATE_LIVEFS is not specified these both write to the same area. cdrom.3 builds the bootonly image and the UFS miniroot. - The various distribution directories are now in a <buildname> subdirectory of disc1 rather than being in the root so that the merged image's root is not quite as cluttered. - The disc1 image is now called disc1.iso and we no longer build a miniinst.iso image. If CD_EXTRA_BITS is defined, then the iso.1 target will simply include those bits in the disc1 image that it builds. MFC after: 1 day Requested by: re (scottl, etc.)
2005-03-02 22:46:07 +00:00
cdrom.2:
@echo "Building CDROM disc1 filesystem image"
@mkdir -p ${CD_DISC1}/${BUILDNAME}
.if defined(MAKE_FLOPPIES)
@cd ${RD} && find floppies -print | cpio -dumpl ${CD_DISC1}
.endif
@cd ${RD}/dists && find . -print | cpio -dumpl ${CD_DISC1}/${BUILDNAME}
Fold the functionality of NORELNOTES into NODOC. There is now only one Makefile variable to control the building/installation of both the DocProj documents and RELNOTESng. Suggested by: obrien, dd Tested by: John Hay <jhay@icomtek.csir.co.za>
2001-05-30 17:05:15 +00:00
.if !defined(NODOC)
First commit of RELNOTESng, the rewrite of the *.TXT documentation files. src/release/doc/README has additional information. Reviewed by: -current, -doc
2001-04-27 21:28:02 +00:00
@for i in ${DIST_DOCS_ARCH_INDEP}; do \
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
cp ${RND}/${RELNOTES_LANG}/$$i/article.txt \
${CD_DISC1}/`echo $${i} | tr 'a-z' 'A-Z'`.TXT; \
cp ${RND}/${RELNOTES_LANG}/$$i/article.html \
${CD_DISC1}/`echo $${i} | tr 'a-z' 'A-Z'`.HTM; \
First commit of RELNOTESng, the rewrite of the *.TXT documentation files. src/release/doc/README has additional information. Reviewed by: -current, -doc
2001-04-27 21:28:02 +00:00
done
@for i in ${DIST_DOCS_ARCH_DEP}; do \
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
cp ${RND}/${RELNOTES_LANG}/$$i/${TARGET}/article.txt \
${CD_DISC1}/`echo $${i} | tr 'a-z' 'A-Z'`.TXT; \
cp ${RND}/${RELNOTES_LANG}/$$i/${TARGET}/article.html \
${CD_DISC1}/`echo $${i} | tr 'a-z' 'A-Z'`.HTM; \
First commit of RELNOTESng, the rewrite of the *.TXT documentation files. src/release/doc/README has additional information. Reviewed by: -current, -doc
2001-04-27 21:28:02 +00:00
done
Modify the release-building process so the srelease distribution doesn't get spammed by RELNOTESng files. Previously, we built the RELNOTESng files "in place" and then populated the media (e.g. floppy, cdrom, ftp) areas from the build directories. However, the distributions got rolled after the RELNOTESng build, and thus picked up the rendered RELNOTESng files. Now, we install the RELNOTESng documents to R/stage/release.doc, then clean out the original directories in src/release/doc. We use the renditions in the staging area to populate the media areas. PR: 33141 MFC after: 3 days (plus re approval)
2001-12-28 17:21:28 +00:00
@cp ${RND}/${RELNOTES_LANG}/readme/docbook.css ${CD_DISC1}
First commit of RELNOTESng, the rewrite of the *.TXT documentation files. src/release/doc/README has additional information. Reviewed by: -current, -doc
2001-04-27 21:28:02 +00:00
.endif
Rework how we layout the CD/DVD ISO images for releases. Note that this is only the first stage and does not yet handle packages (which will move to their own disc(s)). - By default, include the live filesystem on disc1, so that disc1 is now both an install disc and a fixit disc. The images can still be split into two separate discs by specifying SEPARATE_LIVEFS=yes. - Remove /usr/ports from the live filesystem disc as the merged images do not fit on even i386 otherwise. Also, the merged image already contains a tarball of the entire ports tree, so /usr/ports was a duplicate anyway. - Stop building port READMES for the release sinc they are no longer used (/usr/ports in the live filesystem was the last remaining user). - Split cdrom.1 into three targets. cdrom.1 builds the live filesystem image. cdrom.2 builds the disc1 image. Note that if SEPARATE_LIVEFS is not specified these both write to the same area. cdrom.3 builds the bootonly image and the UFS miniroot. - The various distribution directories are now in a <buildname> subdirectory of disc1 rather than being in the root so that the merged image's root is not quite as cluttered. - The disc1 image is now called disc1.iso and we no longer build a miniinst.iso image. If CD_EXTRA_BITS is defined, then the iso.1 target will simply include those bits in the disc1 image that it builds. MFC after: 1 day Requested by: re (scottl, etc.)
2005-03-02 22:46:07 +00:00
.if defined(SEPARATE_LIVEFS)
@cp -Rp ${CD_LIVEFS}/boot ${CD_DISC1}
@echo "CD_VERSION = ${BUILDNAME}" > ${CD_DISC1}/cdrom.inf
Add 'hint.atkbd.0.flags="0x1"' back to /boot/device.hints on the install media on i386 and amd64. While the current default is very useful in server environments, the ability to use USB keyboards is vastly more important then the ability to use PS/2 keyboards the admin forgot to plug in during the install process. PR: kern/71443
2004-09-14 04:01:57 +00:00
.endif
Fix the missing CD_VOLUME=1 in the cdrom.inf file on disc1 for the livecd != disc1 case (i.e. ia64). The line was appended to the non-existing cdrom.inf file, which was created only later. Move the line to after the file is created. MFC after: 1 day
2006-01-29 20:16:25 +00:00
@echo "CD_VOLUME = 1" >> ${CD_DISC1}/cdrom.inf
What should be the last major changes to the ISO layout: - When a separate livefs ISO is used, change the name of the tree directory to be R/cdrom/livefs and change the name of the iso to livefs.iso rather than using disc2. Instead, disc2 is now always going to be the packages disc regardless of the presence of a separate livefs which makes things simpler. - Build a very simple disc2 tree under R/cdrom that just includes the cdrom.inf file that the packages need. Also, build a disc2.iso image in the iso.1 target. Disc 2's volume label is "FreeBSD_Packages". - Retire CD_EXTRA_BITS and replace it with CD_PACKAGE_TREE. CD_EXTRA_BITS was specific to disc1, but CD_PACKAGE_TREE instead should point to a directory that has disc1 and disc2 subdirectories that contain the extra bits for each image. - Rename the bootonly volume label to "FreeBSD_bootonly" and just hardcode the disc1 and livefs volume names and iso suffixes.
2005-03-23 15:01:54 +00:00
@echo "Building CDROM disc2 filesystem image"
@mkdir -p ${CD_DISC2}
@echo "CD_VERSION = ${BUILDNAME}" > ${CD_DISC2}/cdrom.inf
@echo "CD_VOLUME = 2" >> ${CD_DISC2}/cdrom.inf
- Move the doc project docs over to disc2 to free up some room for future bloat on disc1. - Output a message letting the user know that we are generating MD5 sums during the long pause after the last mkisofs invocation in the iso.1 target. MFC after: 3 days
2005-03-31 16:19:26 +00:00
.if !defined(NODOC)
Move the documentation to its own separate disc to make room for more packages on disc2. This will also let users decide if they want to have a CD of the docs at all - unless they're disconnected from the net they will probably find the Web site more useful. Reviewed by: ru MFC after: 3 days
2006-11-16 23:09:35 +00:00
echo "Building CDROM docs filesystem image"
@mkdir -p ${CD_DOCS}
@echo "CD_VERSION = ${BUILDNAME}" > ${CD_DOCS}/cdrom.inf
@mkdir -p ${CD_DOCS}/usr/share/doc
- Move the doc project docs over to disc2 to free up some room for future bloat on disc1. - Output a message letting the user know that we are generating MD5 sums during the long pause after the last mkisofs invocation in the iso.1 target. MFC after: 3 days
2005-03-31 16:19:26 +00:00
@for i in `ls ${CD_LIVEFS}/usr/share/doc`; do \
if [ -L ${CD_LIVEFS}/usr/share/doc/$$i -o \
-d /usr/doc/$$i ]; then \
mv ${CD_LIVEFS}/usr/share/doc/$$i \
Move the documentation to its own separate disc to make room for more packages on disc2. This will also let users decide if they want to have a CD of the docs at all - unless they're disconnected from the net they will probably find the Web site more useful. Reviewed by: ru MFC after: 3 days
2006-11-16 23:09:35 +00:00
${CD_DOCS}/usr/share/doc; \
- Move the doc project docs over to disc2 to free up some room for future bloat on disc1. - Output a message letting the user know that we are generating MD5 sums during the long pause after the last mkisofs invocation in the iso.1 target. MFC after: 3 days
2005-03-31 16:19:26 +00:00
fi \
done
.endif
Rework how we layout the CD/DVD ISO images for releases. Note that this is only the first stage and does not yet handle packages (which will move to their own disc(s)). - By default, include the live filesystem on disc1, so that disc1 is now both an install disc and a fixit disc. The images can still be split into two separate discs by specifying SEPARATE_LIVEFS=yes. - Remove /usr/ports from the live filesystem disc as the merged images do not fit on even i386 otherwise. Also, the merged image already contains a tarball of the entire ports tree, so /usr/ports was a duplicate anyway. - Stop building port READMES for the release sinc they are no longer used (/usr/ports in the live filesystem was the last remaining user). - Split cdrom.1 into three targets. cdrom.1 builds the live filesystem image. cdrom.2 builds the disc1 image. Note that if SEPARATE_LIVEFS is not specified these both write to the same area. cdrom.3 builds the bootonly image and the UFS miniroot. - The various distribution directories are now in a <buildname> subdirectory of disc1 rather than being in the root so that the merged image's root is not quite as cluttered. - The disc1 image is now called disc1.iso and we no longer build a miniinst.iso image. If CD_EXTRA_BITS is defined, then the iso.1 target will simply include those bits in the disc1 image that it builds. MFC after: 1 day Requested by: re (scottl, etc.)
2005-03-02 22:46:07 +00:00
touch ${.TARGET}
#
# --==## Setup a suitable cdrom-area ##==--
#
cdrom.3:
- Boot-only CD-ROM support (bootonly.iso). Submitted by: jhb, ru - Moved the creation of the ports distribution to release.7. - Call MFS root floppy mfsroot.flp even on a small PC98 disk. Submitted by: jhb Reviewed by: ru, jhb
2003-08-02 18:21:02 +00:00
.if defined(CD_BOOT)
Rework how we layout the CD/DVD ISO images for releases. Note that this is only the first stage and does not yet handle packages (which will move to their own disc(s)). - By default, include the live filesystem on disc1, so that disc1 is now both an install disc and a fixit disc. The images can still be split into two separate discs by specifying SEPARATE_LIVEFS=yes. - Remove /usr/ports from the live filesystem disc as the merged images do not fit on even i386 otherwise. Also, the merged image already contains a tarball of the entire ports tree, so /usr/ports was a duplicate anyway. - Stop building port READMES for the release sinc they are no longer used (/usr/ports in the live filesystem was the last remaining user). - Split cdrom.1 into three targets. cdrom.1 builds the live filesystem image. cdrom.2 builds the disc1 image. Note that if SEPARATE_LIVEFS is not specified these both write to the same area. cdrom.3 builds the bootonly image and the UFS miniroot. - The various distribution directories are now in a <buildname> subdirectory of disc1 rather than being in the root so that the merged image's root is not quite as cluttered. - The disc1 image is now called disc1.iso and we no longer build a miniinst.iso image. If CD_EXTRA_BITS is defined, then the iso.1 target will simply include those bits in the disc1 image that it builds. MFC after: 1 day Requested by: re (scottl, etc.)
2005-03-02 22:46:07 +00:00
@echo "Building bootonly CDROM filesystem image"
@mkdir -p ${CD_BOOT}
@cp -Rp ${CD_LIVEFS}/boot ${CD_BOOT}
What should be the last major changes to the ISO layout: - When a separate livefs ISO is used, change the name of the tree directory to be R/cdrom/livefs and change the name of the iso to livefs.iso rather than using disc2. Instead, disc2 is now always going to be the packages disc regardless of the presence of a separate livefs which makes things simpler. - Build a very simple disc2 tree under R/cdrom that just includes the cdrom.inf file that the packages need. Also, build a disc2.iso image in the iso.1 target. Disc 2's volume label is "FreeBSD_Packages". - Retire CD_EXTRA_BITS and replace it with CD_PACKAGE_TREE. CD_EXTRA_BITS was specific to disc1, but CD_PACKAGE_TREE instead should point to a directory that has disc1 and disc2 subdirectories that contain the extra bits for each image. - Rename the bootonly volume label to "FreeBSD_bootonly" and just hardcode the disc1 and livefs volume names and iso suffixes.
2005-03-23 15:01:54 +00:00
@echo "CD_VERSION = ${BUILDNAME}" > ${CD_BOOT}/cdrom.inf
Added the new knob MINIROOT that, when set, causes the creation of the bootable UFS image, miniroot.ufs, fully identical to the contents of the "bootonly" CD-ROM. The image is made available on FTP as miniroot/miniroot.ufs.gz, for resemblance with NetBSD. This replaces the boot.flp image for sparc64, making the latter as coming with NO_FLOPPIES. Reviewed by: jhb Tested by: hrs
2003-08-08 21:00:57 +00:00
.if defined(MINIROOT)
Rework how we layout the CD/DVD ISO images for releases. Note that this is only the first stage and does not yet handle packages (which will move to their own disc(s)). - By default, include the live filesystem on disc1, so that disc1 is now both an install disc and a fixit disc. The images can still be split into two separate discs by specifying SEPARATE_LIVEFS=yes. - Remove /usr/ports from the live filesystem disc as the merged images do not fit on even i386 otherwise. Also, the merged image already contains a tarball of the entire ports tree, so /usr/ports was a duplicate anyway. - Stop building port READMES for the release sinc they are no longer used (/usr/ports in the live filesystem was the last remaining user). - Split cdrom.1 into three targets. cdrom.1 builds the live filesystem image. cdrom.2 builds the disc1 image. Note that if SEPARATE_LIVEFS is not specified these both write to the same area. cdrom.3 builds the bootonly image and the UFS miniroot. - The various distribution directories are now in a <buildname> subdirectory of disc1 rather than being in the root so that the merged image's root is not quite as cluttered. - The disc1 image is now called disc1.iso and we no longer build a miniinst.iso image. If CD_EXTRA_BITS is defined, then the iso.1 target will simply include those bits in the disc1 image that it builds. MFC after: 1 day Requested by: re (scottl, etc.)
2005-03-02 22:46:07 +00:00
@echo "Building bootonly UFS filesystem image"
Added the new knob MINIROOT that, when set, causes the creation of the bootable UFS image, miniroot.ufs, fully identical to the contents of the "bootonly" CD-ROM. The image is made available on FTP as miniroot/miniroot.ufs.gz, for resemblance with NetBSD. This replaces the boot.flp image for sparc64, making the latter as coming with NO_FLOPPIES. Reviewed by: jhb Tested by: hrs
2003-08-08 21:00:57 +00:00
@mkdir -p ${FD}/miniroot
@sh -e ${DOFS_SH} ${FD}/miniroot/miniroot.ufs \
${RD} ${MNT} 0 ${CD_BOOT} 8192 auto
@gzip -9v ${FD}/miniroot/miniroot.ufs
.endif
OK, *this* is how the ports stuff should have been built to begin with. ;-)
1997-07-13 00:54:22 +00:00
.endif
Replaced the hardcoded target names in the touch(1) commands with ${.TARGET}, to ease maintenance.
2003-07-24 13:45:05 +00:00
touch ${.TARGET}
Remove imutable bits, they will not survive anyway. First cut at cdrom make rule.
1995-03-16 22:54:42 +00:00
What should be the last major changes to the ISO layout: - When a separate livefs ISO is used, change the name of the tree directory to be R/cdrom/livefs and change the name of the iso to livefs.iso rather than using disc2. Instead, disc2 is now always going to be the packages disc regardless of the presence of a separate livefs which makes things simpler. - Build a very simple disc2 tree under R/cdrom that just includes the cdrom.inf file that the packages need. Also, build a disc2.iso image in the iso.1 target. Disc 2's volume label is "FreeBSD_Packages". - Retire CD_EXTRA_BITS and replace it with CD_PACKAGE_TREE. CD_EXTRA_BITS was specific to disc1, but CD_PACKAGE_TREE instead should point to a directory that has disc1 and disc2 subdirectories that contain the extra bits for each image. - Rename the bootonly volume label to "FreeBSD_bootonly" and just hardcode the disc1 and livefs volume names and iso suffixes.
2005-03-23 15:01:54 +00:00
.if make(iso.1)
.if defined(CD_PACKAGE_TREE)
.if exists(${CD_PACKAGE_TREE}/disc1)
CD_DISC1_PKGS= ${CD_PACKAGE_TREE}/disc1
.endif
.if exists(${CD_PACKAGE_TREE}/disc2)
CD_DISC2_PKGS= ${CD_PACKAGE_TREE}/disc2
.endif
.endif
.endif
Add scripts and Makefile support for building ISO images automatically as part of the release if MAKE_ISOS is set. Will also build the first CD with packages (in addition to the "minimal" CD) if CD_EXTRA_BITS points to them. This probably need a bit more work to get fully useful, but it at least covers the basics for now.
2001-07-23 09:01:46 +00:00
iso.1:
Unbreak iso.1.
2003-03-17 17:49:57 +00:00
.if exists(${.CURDIR}/${TARGET_ARCH}/mkisoimages.sh)
Use positive logic rather than harder-to-read negative logic.
2003-03-17 05:46:10 +00:00
@echo "Creating ISO images..."
- Boot-only CD-ROM support (bootonly.iso). Submitted by: jhb, ru - Moved the creation of the ports distribution to release.7. - Call MFS root floppy mfsroot.flp even on a small PC98 disk. Submitted by: jhb Reviewed by: ru, jhb
2003-08-02 18:21:02 +00:00
.if defined(CD_BOOT)
@sh ${.CURDIR}/${TARGET_ARCH}/mkisoimages.sh ${BOOTABLE} \
What should be the last major changes to the ISO layout: - When a separate livefs ISO is used, change the name of the tree directory to be R/cdrom/livefs and change the name of the iso to livefs.iso rather than using disc2. Instead, disc2 is now always going to be the packages disc regardless of the presence of a separate livefs which makes things simpler. - Build a very simple disc2 tree under R/cdrom that just includes the cdrom.inf file that the packages need. Also, build a disc2.iso image in the iso.1 target. Disc 2's volume label is "FreeBSD_Packages". - Retire CD_EXTRA_BITS and replace it with CD_PACKAGE_TREE. CD_EXTRA_BITS was specific to disc1, but CD_PACKAGE_TREE instead should point to a directory that has disc1 and disc2 subdirectories that contain the extra bits for each image. - Rename the bootonly volume label to "FreeBSD_bootonly" and just hardcode the disc1 and livefs volume names and iso suffixes.
2005-03-23 15:01:54 +00:00
FreeBSD_bootonly \
- Boot-only CD-ROM support (bootonly.iso). Submitted by: jhb, ru - Moved the creation of the ports distribution to release.7. - Call MFS root floppy mfsroot.flp even on a small PC98 disk. Submitted by: jhb Reviewed by: ru, jhb
2003-08-02 18:21:02 +00:00
${CD}/${BUILDNAME}-${TARGET}-bootonly.iso ${CD_BOOT}
.endif
Use positive logic rather than harder-to-read negative logic.
2003-03-17 05:46:10 +00:00
@sh ${.CURDIR}/${TARGET_ARCH}/mkisoimages.sh ${BOOTABLE} \
What should be the last major changes to the ISO layout: - When a separate livefs ISO is used, change the name of the tree directory to be R/cdrom/livefs and change the name of the iso to livefs.iso rather than using disc2. Instead, disc2 is now always going to be the packages disc regardless of the presence of a separate livefs which makes things simpler. - Build a very simple disc2 tree under R/cdrom that just includes the cdrom.inf file that the packages need. Also, build a disc2.iso image in the iso.1 target. Disc 2's volume label is "FreeBSD_Packages". - Retire CD_EXTRA_BITS and replace it with CD_PACKAGE_TREE. CD_EXTRA_BITS was specific to disc1, but CD_PACKAGE_TREE instead should point to a directory that has disc1 and disc2 subdirectories that contain the extra bits for each image. - Rename the bootonly volume label to "FreeBSD_bootonly" and just hardcode the disc1 and livefs volume names and iso suffixes.
2005-03-23 15:01:54 +00:00
FreeBSD_Install \
${CD}/${BUILDNAME}-${TARGET}-disc1.iso ${CD_DISC1} \
${CD_DISC1_PKGS}
@sh ${.CURDIR}/${TARGET_ARCH}/mkisoimages.sh \
FreeBSD_Packages \
${CD}/${BUILDNAME}-${TARGET}-disc2.iso ${CD_DISC2} \
${CD_DISC2_PKGS}
Move the documentation to its own separate disc to make room for more packages on disc2. This will also let users decide if they want to have a CD of the docs at all - unless they're disconnected from the net they will probably find the Web site more useful. Reviewed by: ru MFC after: 3 days
2006-11-16 23:09:35 +00:00
.if !defined(NODOC)
@sh ${.CURDIR}/${TARGET_ARCH}/mkisoimages.sh \
FreeBSD_Documentation \
${CD}/${BUILDNAME}-${TARGET}-docs.iso ${CD_DOCS}
.endif
Rework how we layout the CD/DVD ISO images for releases. Note that this is only the first stage and does not yet handle packages (which will move to their own disc(s)). - By default, include the live filesystem on disc1, so that disc1 is now both an install disc and a fixit disc. The images can still be split into two separate discs by specifying SEPARATE_LIVEFS=yes. - Remove /usr/ports from the live filesystem disc as the merged images do not fit on even i386 otherwise. Also, the merged image already contains a tarball of the entire ports tree, so /usr/ports was a duplicate anyway. - Stop building port READMES for the release sinc they are no longer used (/usr/ports in the live filesystem was the last remaining user). - Split cdrom.1 into three targets. cdrom.1 builds the live filesystem image. cdrom.2 builds the disc1 image. Note that if SEPARATE_LIVEFS is not specified these both write to the same area. cdrom.3 builds the bootonly image and the UFS miniroot. - The various distribution directories are now in a <buildname> subdirectory of disc1 rather than being in the root so that the merged image's root is not quite as cluttered. - The disc1 image is now called disc1.iso and we no longer build a miniinst.iso image. If CD_EXTRA_BITS is defined, then the iso.1 target will simply include those bits in the disc1 image that it builds. MFC after: 1 day Requested by: re (scottl, etc.)
2005-03-02 22:46:07 +00:00
.if defined(SEPARATE_LIVEFS)
Use positive logic rather than harder-to-read negative logic.
2003-03-17 05:46:10 +00:00
@sh ${.CURDIR}/${TARGET_ARCH}/mkisoimages.sh ${BOOTABLE} \
What should be the last major changes to the ISO layout: - When a separate livefs ISO is used, change the name of the tree directory to be R/cdrom/livefs and change the name of the iso to livefs.iso rather than using disc2. Instead, disc2 is now always going to be the packages disc regardless of the presence of a separate livefs which makes things simpler. - Build a very simple disc2 tree under R/cdrom that just includes the cdrom.inf file that the packages need. Also, build a disc2.iso image in the iso.1 target. Disc 2's volume label is "FreeBSD_Packages". - Retire CD_EXTRA_BITS and replace it with CD_PACKAGE_TREE. CD_EXTRA_BITS was specific to disc1, but CD_PACKAGE_TREE instead should point to a directory that has disc1 and disc2 subdirectories that contain the extra bits for each image. - Rename the bootonly volume label to "FreeBSD_bootonly" and just hardcode the disc1 and livefs volume names and iso suffixes.
2005-03-23 15:01:54 +00:00
FreeBSD_LiveFS \
${CD}/${BUILDNAME}-${TARGET}-livefs.iso ${CD_LIVEFS}
Use positive logic rather than harder-to-read negative logic.
2003-03-17 05:46:10 +00:00
.endif
Add SHA256 checksums to the release build. Requested by: cperciva (so) Reviewed/improved by: ru MFC after: 2 days
2005-09-28 12:39:00 +00:00
@echo "Generating MD5 and SHA256 sums..."
Automatically generate the CHECKSUM.MD5 file for ISO's. Approved by: scottl
2003-12-01 10:46:45 +00:00
@(cd ${CD} && md5 *.iso > ${BUILDNAME}-${TARGET}-iso.CHECKSUM.MD5)
Add SHA256 checksums to the release build. Requested by: cperciva (so) Reviewed/improved by: ru MFC after: 2 days
2005-09-28 12:39:00 +00:00
@(cd ${CD} && sha256 *.iso > ${BUILDNAME}-${TARGET}-iso.CHECKSUM.SHA256)
Replaced the hardcoded target names in the touch(1) commands with ${.TARGET}, to ease maintenance.
2003-07-24 13:45:05 +00:00
touch ${.TARGET}
Use positive logic rather than harder-to-read negative logic.
2003-03-17 05:46:10 +00:00
.else
Unbreak iso.1.
2003-03-17 17:49:57 +00:00
@echo "Do not know how to create an ISO for ${TARGET_ARCH}."
Use positive logic rather than harder-to-read negative logic.
2003-03-17 05:46:10 +00:00
.endif
Add scripts and Makefile support for building ISO images automatically as part of the release if MAKE_ISOS is set. Will also build the first CD with packages (in addition to the "minimal" CD) if CD_EXTRA_BITS points to them. This probably need a bit more work to get fully useful, but it at least covers the basics for now.
2001-07-23 09:01:46 +00:00
First commit of RELNOTESng, the rewrite of the *.TXT documentation files. src/release/doc/README has additional information. Reviewed by: -current, -doc
2001-04-27 21:28:02 +00:00
#
# --==## Documentation Project files such as the Handbook and FAQ ##==--
#
The doc dist was being made too early - move it.
1997-06-22 17:21:59 +00:00
doc.1:
Hammer on the boot floppy generation bits some more. This is looking a little better.
1999-01-05 03:44:08 +00:00
@echo "Making docs..."
@for i in ${DOCPORTS}; do \
Build FDP ports with a clean runtime environment; gmake(1) somehow uses ${TARGET_ARCH}, in a way that is incompatible with our usage. Submitted by: jhay, ru
2002-07-26 11:27:11 +00:00
cd /usr/ports/$$i && \
o Run make readmes (ports) from within CHROOTDIR to avoid pollution and/or breakages due to /etc/make.conf on the host. Move it after make world and put it in an if-block so that we can include the commands in the script even if we're not going to make the readmes. o When building docproj, make sure we propagate the setting of the FTP_PASSIVE_MODE environment variable. Not propagating the envvar may cause fetch failures. This increases the number of cases for which one does not have to prefetch the packages.
2003-03-08 08:51:25 +00:00
env -i FTP_PASSIVE_MODE=$${FTP_PASSIVE_MODE:-no} PATH=$${PATH} \
make all install clean BATCH=yes WITHOUT_X11=yes JADETEX=no \
WITHOUT_PYTHON=yes FORCE_PKG_REGISTER=yes; \
Clean up doc bootstrap ports afterwards. Small stylistic cleanup.
1998-01-24 07:30:55 +00:00
done
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
@cd /usr/doc && make all install 'FORMATS=html html-split txt' \
Add URLS_ABSOLUTE=YES into the doc.1 target.
2004-05-05 13:06:40 +00:00
INSTALL_COMPRESSED='' DOCDIR=${RD}/trees/base/usr/share/doc \
URLS_ABSOLUTE=YES
Replaced the hardcoded target names in the touch(1) commands with ${.TARGET}, to ease maintenance.
2003-07-24 13:45:05 +00:00
touch ${.TARGET}
The doc dist was being made too early - move it.
1997-06-22 17:21:59 +00:00
First commit of RELNOTESng, the rewrite of the *.TXT documentation files. src/release/doc/README has additional information. Reviewed by: -current, -doc
2001-04-27 21:28:02 +00:00
#
# --==## RELNOTESng: Next-generation replacements for *.TXT files ##==--
#
doc.2:
@echo "Making release documentation..."
Add missing '\' in previous change.
2004-01-15 10:17:43 +00:00
@cd ${.CURDIR}/doc && make all install clean 'FORMATS=html txt' \
When building release documentation for release builds or for release documentation snapshots, force URLS_ABSOLUTE. This allows us to relocate HTML renderings anywhere without worrying about breaking relative links to pages on the Web site. This can happen when we use &base; as part of a URL. I wonder if we should do this for all release documentation builds, but I'm not quite sure what all the cases are.
2005-03-20 17:59:17 +00:00
INSTALL_COMPRESSED='' URLS_ABSOLUTE=YES DOCDIR=${RND}
Replaced the hardcoded target names in the touch(1) commands with ${.TARGET}, to ease maintenance.
2003-07-24 13:45:05 +00:00
touch ${.TARGET}
First commit of RELNOTESng, the rewrite of the *.TXT documentation files. src/release/doc/README has additional information. Reviewed by: -current, -doc
2001-04-27 21:28:02 +00:00
Various changes to make this work.
1995-02-26 01:35:32 +00:00
# Various "subroutine" and other supporting targets.
release glue to support s/des/crypto/ Submitted by: markm
2000-02-29 11:26:24 +00:00
# RD=
# SD=
# TD=
# ARG=
New and better "make release" rules.
1995-02-25 22:08:48 +00:00
doTARBALL:
.if !defined(SD)
Moved doFS into shell-script so we can be more smart about it. make two boot floppies, one without help files for 4MB systems. Mine's loading one now. :-) -O2 may be needed.
1996-01-08 22:13:50 +00:00
@echo "SD undefined in doTARBALL" && exit 1
Remove trailing whitespace.
1995-05-30 08:29:07 +00:00
.endif
New and better "make release" rules.
1995-02-25 22:08:48 +00:00
.if !defined(TD)
Fix typo in error message.
2001-04-08 23:02:12 +00:00
@echo "TD undefined in doTARBALL" && exit 1
Remove trailing whitespace.
1995-05-30 08:29:07 +00:00
.endif
New and better "make release" rules.
1995-02-25 22:08:48 +00:00
.if !defined(ARG)
Moved doFS into shell-script so we can be more smart about it. make two boot floppies, one without help files for 4MB systems. Mine's loading one now. :-) -O2 may be needed.
1996-01-08 22:13:50 +00:00
@echo "ARG undefined in doTARBALL" && exit 1
Remove trailing whitespace.
1995-05-30 08:29:07 +00:00
.endif
Hammer on the boot floppy generation bits some more. This is looking a little better.
1999-01-05 03:44:08 +00:00
@rm -rf ${RD}/dists/${TD}/${TN}*
@mkdir -p ${RD}/dists/${TD}
@( cd ${SD} && \
Stop cutting the tarball names to eight symbols. This makes it possible to e.g. distribute kernels with config names larger than eight symbols, without the clash. Previousy, LALALALA and LALALALA-SMP would be the same tarball. (I think this comes from the old days where tarballs were put on the MS-DOS formatted diskettes.) MFC after: 3 days
2006-09-08 19:54:37 +00:00
tn=`echo ${TN} | tr 'A-Z' 'a-z'` && \
Various changes: tip removed from cpio.flp (!) crunch chokes on it after the latest changes. source for secure & eBones moved into securedist.
1995-04-07 01:20:26 +00:00
echo rolling ${TD}/$$tn tarball &&\
Do not spam our distributions accidentally with BOOTMFS files.
1997-03-12 18:09:15 +00:00
tar --exclude CVS --exclude obj --exclude BOOTMFS -cf - ${ARG} | \
Add rule to generate the info files. Submitted by: gpalmer
1995-05-24 01:26:23 +00:00
${ZIPNSPLIT} ${RD}/dists/${TD}/$$tn. && \
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
sh ${.CURDIR}/scripts/info.sh ${RD}/dists/${TD}/$$tn \
> ${RD}/dists/${TD}/$$tn.inf && \
Ok, NOW the install.sh scripts are installed correctly!
1996-06-13 09:35:16 +00:00
if [ -f ${.CURDIR}/scripts/$${TD}-install.sh ]; then \
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
cp -p ${.CURDIR}/scripts/$${TD}-install.sh \
make sure install scripts are executable Submitted by: Ceri Davies
2006-03-21 15:50:08 +00:00
${RD}/dists/${TD}/install.sh && \
chmod +x ${RD}/dists/${TD}/install.sh; \
Ok, NOW the install.sh scripts are installed correctly!
1996-06-13 09:35:16 +00:00
fi && \
Make CHECKSUM.MD5 files automatically.
1997-01-01 03:16:34 +00:00
if [ "${SD}" != "/usr/src" ]; then \
Backout addition of -L switch to mtree. Using -L breaks the build process in too many cases. Adding mtree to bootstrap-tools to solve this breaks the upgrade path because mtree needs a libc that has strtofflags and fflagstostr.
2000-07-23 16:33:00 +00:00
mtree -c -i -p ${SD}/${ARG} \
Allow user to selectively redo any given part of the release in `rerelease'. Create smaller BOOTMFS kernel with more sane sed command rather than fgrep/sed Make it possible to generate multiple kernels with the KERNELS variable. Add mtree generated distribution signatures to dists.
1996-06-06 10:13:32 +00:00
-k gname,md5digest,mode,nlink,uname,size,link,type \
Make CHECKSUM.MD5 files automatically.
1997-01-01 03:16:34 +00:00
> ${RD}/dists/${TD}/$$tn.mtree ; \
else \
true; \
fi; \
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
( cd ${RD}/dists/${TD}; \
Fix a bug which puts an incorrect checksum of CHECKSUM.MD5 into the CHECKSUM.MD5 file itself. Approved by: re (scottl) MFC after: 3 days
2006-02-08 11:18:29 +00:00
rm -f CHECKSUM.MD5 CHECKSUM.SHA256; \
Add SHA256 checksums to the release build. Requested by: cperciva (so) Reviewed/improved by: ru MFC after: 2 days
2005-09-28 12:39:00 +00:00
md5 * > .CHECKSUM.MD5; \
Fix a bug which puts an incorrect checksum of CHECKSUM.MD5 into the CHECKSUM.MD5 file itself. Approved by: re (scottl) MFC after: 3 days
2006-02-08 11:18:29 +00:00
sha256 * > .CHECKSUM.SHA256; \
mv .CHECKSUM.MD5 CHECKSUM.MD5; \
mv .CHECKSUM.SHA256 CHECKSUM.SHA256) \
Make CHECKSUM.MD5 files automatically.
1997-01-01 03:16:34 +00:00
)
Various changes to make this work.
1995-02-26 01:35:32 +00:00
Crypto is now part of the "base" distribution, and the old "release.3" target no longer exists. Renumber the old release.[4-9] targets to release.[3-8]. Reminded by: ru
2004-08-06 08:49:51 +00:00
doRELEASE: release.1 release.2 ${DOCREL} release.3 release.4 \
release.5 release.6 release.7 release.8 ${EXTRAS}
Various changes to make this work.
1995-02-26 01:35:32 +00:00
@echo "Release done"
Make a little convenient "floppies" target for myself.
1995-05-18 13:24:48 +00:00
floppies:
Crypto is now part of the "base" distribution, and the old "release.3" target no longer exists. Renumber the old release.[4-9] targets to release.[3-8]. Reminded by: ru
2004-08-06 08:49:51 +00:00
@rm -f release.4 release.8 floppies.[123]
@cd ${.CURDIR} && ${MAKE} release.4 release.8 floppies.1 floppies.2 \
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
floppies.3
Hammer on the boot floppy generation bits some more. This is looking a little better.
1999-01-05 03:44:08 +00:00
@cd ${RD} && find floppies -print | cpio -dumpl ${FD}
Make a little convenient "floppies" target for myself.
1995-05-18 13:24:48 +00:00
A major cleanup. Move more over to a "pull" model than a "push" model, this will eventually allow us to add meaningfull dependencies. Add mount_{msdos,nfs,cd9660} and chroot to the fixit.flp.
1995-04-20 06:48:31 +00:00
installCRUNCH:
.if !defined(CRUNCH)
Moved doFS into shell-script so we can be more smart about it. make two boot floppies, one without help files for 4MB systems. Mine's loading one now. :-) -O2 may be needed.
1996-01-08 22:13:50 +00:00
@echo "CRUNCH undefined in installCRUNCH" && exit 1
A major cleanup. Move more over to a "pull" model than a "push" model, this will eventually allow us to add meaningfull dependencies. Add mount_{msdos,nfs,cd9660} and chroot to the fixit.flp.
1995-04-20 06:48:31 +00:00
.endif
.if !defined(DIR)
Moved doFS into shell-script so we can be more smart about it. make two boot floppies, one without help files for 4MB systems. Mine's loading one now. :-) -O2 may be needed.
1996-01-08 22:13:50 +00:00
@echo "DIR undefined in installCRUNCH" && exit 1
A major cleanup. Move more over to a "pull" model than a "push" model, this will eventually allow us to add meaningfull dependencies. Add mount_{msdos,nfs,cd9660} and chroot to the fixit.flp.
1995-04-20 06:48:31 +00:00
.endif
.if !defined(ZIP)
Moved doFS into shell-script so we can be more smart about it. make two boot floppies, one without help files for 4MB systems. Mine's loading one now. :-) -O2 may be needed.
1996-01-08 22:13:50 +00:00
@echo "ZIP undefined in installCRUNCH" && exit 1
A major cleanup. Move more over to a "pull" model than a "push" model, this will eventually allow us to add meaningfull dependencies. Add mount_{msdos,nfs,cd9660} and chroot to the fixit.flp.
1995-04-20 06:48:31 +00:00
.endif
Hammer on the boot floppy generation bits some more. This is looking a little better.
1999-01-05 03:44:08 +00:00
@if ${ZIP} ; then \
A major cleanup. Move more over to a "pull" model than a "push" model, this will eventually allow us to add meaningfull dependencies. Add mount_{msdos,nfs,cd9660} and chroot to the fixit.flp.
1995-04-20 06:48:31 +00:00
gzip -9 < ${RD}/crunch/${CRUNCH} > ${DIR}/${CRUNCH}_crunch ; \
else \
ln -f ${RD}/crunch/${CRUNCH} ${DIR}/${CRUNCH}_crunch ; \
fi
Hammer on the boot floppy generation bits some more. This is looking a little better.
1999-01-05 03:44:08 +00:00
@chmod 555 ${DIR}/${CRUNCH}_crunch
Milestone #2 in cross-arch make releases. They now work! You need to set TARGET_ARCH and possibly TARGET, the same way you normally do it for a cross build(7). Renamed `distribworld' to a more natutal `distributeworld'. Put pwd_mkdb(8) under ${INSTALLTMP}; for `distributeworld'.
2002-04-30 11:50:01 +00:00
@if [ -f ${.CURDIR}/${TARGET}/${CRUNCH}_crunch.conf ] ; then \
for i in `crunchgen -l ${.CURDIR}/${TARGET}/${CRUNCH}_crunch.conf` ; do \
- Removed machine dependent directory and command from src/release/{boot,fixit}_crunch.conf. - Added machine specific fixit_crunch.conf for PC/AT and PC-98 to src/release/$MACHINE. - Use config file in src/release/$MACHINE if exist. If it does not exist, use in src/release.
2000-06-29 13:34:54 +00:00
ln -f ${DIR}/${CRUNCH}_crunch ${DIR}/$$i ; \
done \
else \
for i in `crunchgen -l ${.CURDIR}/${CRUNCH}_crunch.conf` ; do \
ln -f ${DIR}/${CRUNCH}_crunch ${DIR}/$$i ; \
done \
fi
Moved doFS into shell-script so we can be more smart about it. make two boot floppies, one without help files for 4MB systems. Mine's loading one now. :-) -O2 may be needed.
1996-01-08 22:13:50 +00:00
#
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
# --==## Build a floppy set for a splitfs file ##==--
Moved doFS into shell-script so we can be more smart about it. make two boot floppies, one without help files for 4MB systems. Mine's loading one now. :-) -O2 may be needed.
1996-01-08 22:13:50 +00:00
#
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
# FLOPPYBASE - basename of floppy image files
# FLOPPYDESC - description of floppy set
# SPLITFILE - filename of the file to split
# FDSIZE - if specified and "small", small floppy is created
.if make(makeFloppySet)
SPLITDIR= ${RD}/floppyset/${FLOPPYBASE}
.if defined(FDSIZE) && ${FDSIZE} == "SMALL"
FLPSPLITSIZE= ${SMALLFLOPPYSPLITSIZE}
FLPSIZE= ${SMALLFLOPPYSIZE}
FLPLABEL= ${SMALLFLOPPYLABEL}
.else
FLPSPLITSIZE= ${FLOPPYSPLITSIZE}
FLPSIZE= ${FLOPPYSIZE}
FLPLABEL= ${FLOPPYLABEL}
PC98 creates different BOOTMFS configs for different floppy sizes. Broken in: rev. 1.672 Spotted by: nyan
2002-05-25 16:13:52 +00:00
.endif
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
.endif
makeFloppySet:
.if !defined(FLOPPYBASE)
@echo "FLOPPYBASE undefined in ${.TARGET}" && exit 1
.endif
.if !defined(FLOPPYDESC)
@echo "FLOPPYDESC undefined in ${.TARGET}" && exit 1
.endif
.if !defined(SPLITFILE)
@echo "SPLITFILE undefined in ${.TARGET}" && exit 1
.endif
Two nits from rev. 1.835: - Fixed spelling of NO_FLOPPIES to unbreak snapshot building on architectures that don't support them (as installation source). - Don't rely on CVS setting execution bits on the shell script. (This is also consistent with the rest of this makefile.) Reviewed by: jhb
2004-01-29 16:21:00 +00:00
sh ${.CURDIR}/scripts/split-file.sh ${SPLITFILE} \
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
${RD}/floppyset/${FLOPPYBASE} ${FLPSPLITSIZE} "${FLOPPYDESC}"
( splitfile=${SPLITDIR}/`basename ${SPLITFILE}`.split ; \
lines=`cat $${splitfile} | wc -l`; \
Fixed the shell arithmetic to work with 4.x /bin/sh. Tested by: make release of HEAD on 4.10-STABLE
2004-08-26 19:15:20 +00:00
lines=$$(($$lines - 1)) ; \
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
for line in `jot $$lines`; do \
file=`head -n $$(($${line} + 1)) $${splitfile} | tail -1 | cut -f 1 -d ' '` ; \
sh -e ${DOFS_SH} ${RD}/floppies/${FLOPPYBASE}$${line}.flp \
${RD} ${MNT} ${FLPSIZE} ${SPLITDIR}/$${file} \
${BOOTINODE} ${FLPLABEL}; \
done )
#
# --==## Build a boot floppy ##==--
#
# FSIMAGE - base floppy image name
# FDSIZE - if specified and "small", small floppy is created
# KERNFILE - path to kernel split file
# MFSROOTFILE - path to mfsroot split file
.if make(buildBootFloppy)
IMAGEDIR= ${RD}/image.${FSIMAGE}
- Use a stock loader with forth support on the boot floppies. - Use the regular forth scripts including the stock device.hints file on the boot floppies. - gzip all the various text files on the boot floppy that are larger than one sector to save on space. - Generate a loader.conf to load the kernel, acpi.ko, and mfsroot making use of the 'module_before' variables to prompt for a floppy to be entered when required and use a stock loader.rc. This enables the beastie menu on i386. - PC98 has been changed to split the mfsroot for now so that the floppies fit. When the PC98 bootstrap is updated to work with ELF files (which needs to happen anyway), then SPLIT_MFSROOT can be turned off for PC98.
2004-02-03 22:05:55 +00:00
BOOTDIR= ${RD}/trees/base/boot
HINTSFILE= ${BOOTDIR}/device.hints
Fix the acpi.ko path for boot.flp. PR: misc/96466 Submitted by: Nobuyuki Koganemaru MFC after: 1 day
2006-04-29 14:03:17 +00:00
ACPI_KO= ${RD}/kernels/GENERIC/acpi.ko
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
IMAGEFILE= ${RD}/floppies/${FSIMAGE}.flp
.if defined(FDSIZE) && ${FDSIZE} == "SMALL"
FLPSIZE= ${SMALLFLOPPYSIZE}
FLPLABEL= ${SMALLFLOPPYLABEL}
.else
FLPSIZE= ${FLOPPYSIZE}
FLPLABEL= ${FLOPPYLABEL}
.endif
.endif
buildBootFloppy:
.if !defined(FSIMAGE)
@echo "FSIMAGE undefined in ${.TARGET}" && exit 1
.endif
.if !defined(KERNFILE)
@echo "KERNFILE undefined in ${.TARGET}" && exit 1
.endif
.if !defined(MFSROOTFILE)
@echo "MFSROOTFILE undefined in ${.TARGET}" && exit 1
.endif
@echo "Running ${.TARGET} for ${FSIMAGE}"
@rm -rf ${IMAGEDIR}
@mkdir ${IMAGEDIR}
Hammer on the boot floppy generation bits some more. This is looking a little better.
1999-01-05 03:44:08 +00:00
@echo "Setting up /boot directory for ${FSIMAGE} floppy"
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
@mkdir -p ${IMAGEDIR}/boot
Milestone #2 in cross-arch make releases. They now work! You need to set TARGET_ARCH and possibly TARGET, the same way you normally do it for a cross build(7). Renamed `distribworld' to a more natutal `distributeworld'. Put pwd_mkdb(8) under ${INSTALLTMP}; for `distributeworld'.
2002-04-30 11:50:01 +00:00
.if ${TARGET} == "i386"
* if-like-constructs now use <TAB>body consistently. * Long statements with '\' continuation now use <4 spaces>continuation consistently. * wraps some long lines * More consistent on how very simular repeated commands are wrapped.
2004-01-13 18:46:09 +00:00
@${WMAKEENV} kgzip -v -l ${RD}/trees/base/usr/lib/kgzldr.o -o \
- Use a stock loader with forth support on the boot floppies. - Use the regular forth scripts including the stock device.hints file on the boot floppies. - gzip all the various text files on the boot floppy that are larger than one sector to save on space. - Generate a loader.conf to load the kernel, acpi.ko, and mfsroot making use of the 'module_before' variables to prompt for a floppy to be entered when required and use a stock loader.rc. This enables the beastie menu on i386. - PC98 has been changed to split the mfsroot for now so that the floppies fit. When the PC98 bootstrap is updated to work with ELF files (which needs to happen anyway), then SPLIT_MFSROOT can be turned off for PC98.
2004-02-03 22:05:55 +00:00
${IMAGEDIR}/boot/loader ${BOOTDIR}/loader
MFS 1.536.2.52 (partial) and 1.536.2.70: kgzip loader if i386 (only). 5.0-CURRENT-20020121-JPSNAP was failed to build since kern.flp has no space. There is already a hack to kgzip loader, and it makes: -r-xr-xr-x 1 root wheel 163840 Jan 21 07:40 loader* -r-xr-xr-x 1 root wheel 90121 Jan 21 08:06 loader.kgz* 73719 bytes of free spaces to us.
2002-01-20 23:17:52 +00:00
.else
- Use a stock loader with forth support on the boot floppies. - Use the regular forth scripts including the stock device.hints file on the boot floppies. - gzip all the various text files on the boot floppy that are larger than one sector to save on space. - Generate a loader.conf to load the kernel, acpi.ko, and mfsroot making use of the 'module_before' variables to prompt for a floppy to be entered when required and use a stock loader.rc. This enables the beastie menu on i386. - PC98 has been changed to split the mfsroot for now so that the floppies fit. When the PC98 bootstrap is updated to work with ELF files (which needs to happen anyway), then SPLIT_MFSROOT can be turned off for PC98.
2004-02-03 22:05:55 +00:00
@cp ${BOOTDIR}/loader ${IMAGEDIR}/boot
.endif
@cp -Rp ${BOOTDIR}/*.4th ${BOOTDIR}/defaults ${BOOTDIR}/loader.help \
${BOOTDIR}/loader.rc ${IMAGEDIR}/boot
.if exists(${HINTSFILE})
Add 'hint.atkbd.0.flags="0x1"' back to /boot/device.hints on the install media on i386 and amd64. While the current default is very useful in server environments, the ability to use USB keyboards is vastly more important then the ability to use PS/2 keyboards the admin forgot to plug in during the install process. PR: kern/71443
2004-09-14 04:01:57 +00:00
@cp ${HINTSFILE} ${IMAGEDIR}/boot/device.hints
@gzip -9n ${IMAGEDIR}/boot/device.hints
- Use a stock loader with forth support on the boot floppies. - Use the regular forth scripts including the stock device.hints file on the boot floppies. - gzip all the various text files on the boot floppy that are larger than one sector to save on space. - Generate a loader.conf to load the kernel, acpi.ko, and mfsroot making use of the 'module_before' variables to prompt for a floppy to be entered when required and use a stock loader.rc. This enables the beastie menu on i386. - PC98 has been changed to split the mfsroot for now so that the floppies fit. When the PC98 bootstrap is updated to work with ELF files (which needs to happen anyway), then SPLIT_MFSROOT can be turned off for PC98.
2004-02-03 22:05:55 +00:00
.endif
@gzip -9n ${IMAGEDIR}/boot/*.4th ${IMAGEDIR}/boot/loader.help \
${IMAGEDIR}/boot/defaults/loader.conf
@echo 'bootfile="/kernel"' > ${IMAGEDIR}/boot/loader.conf
- Tweak order of arch fs options to stick all floppy options together and all MFS root options together and in the same order on all archs. - Drop support for using the no longer existing 2.88 floppy image for CD booting on i386 via the EMUL_BOOT variable. - Use a stock loader on the boot floppy rather than one without forth. - Force the gzip of the mfsroot in release.9 to succeed. This is useful when doing lots of rereleases when testing floppy changes. - If an acpi kernel module exists, stick it on the boot floppy and for now always load it after the kernel.
2004-01-29 19:01:58 +00:00
.if exists(${ACPI_KO})
@gzip -9nc ${ACPI_KO} > ${IMAGEDIR}/acpi.ko.gz
- Use a stock loader with forth support on the boot floppies. - Use the regular forth scripts including the stock device.hints file on the boot floppies. - gzip all the various text files on the boot floppy that are larger than one sector to save on space. - Generate a loader.conf to load the kernel, acpi.ko, and mfsroot making use of the 'module_before' variables to prompt for a floppy to be entered when required and use a stock loader.rc. This enables the beastie menu on i386. - PC98 has been changed to split the mfsroot for now so that the floppies fit. When the PC98 bootstrap is updated to work with ELF files (which needs to happen anyway), then SPLIT_MFSROOT can be turned off for PC98.
2004-02-03 22:05:55 +00:00
@echo 'acpi_load="YES"' >> ${IMAGEDIR}/boot/loader.conf
@echo 'acpi_name="/acpi.ko"' >> ${IMAGEDIR}/boot/loader.conf
@echo 'acpi_before="read -p \"Insert boot floppy and press Enter\""' >> ${IMAGEDIR}/boot/loader.conf
.endif
@echo 'mfsroot_load="YES"' >> ${IMAGEDIR}/boot/loader.conf
@echo 'mfsroot_type="mfs_root"' >> ${IMAGEDIR}/boot/loader.conf
@echo 'mfsroot_name="/mfsroot"' >> ${IMAGEDIR}/boot/loader.conf
.if !exists(${ACPI_KO})
@echo 'mfsroot_before="read -p \"Insert boot floppy and press Enter\""' >> ${IMAGEDIR}/boot/loader.conf
- Tweak order of arch fs options to stick all floppy options together and all MFS root options together and in the same order on all archs. - Drop support for using the no longer existing 2.88 floppy image for CD booting on i386 via the EMUL_BOOT variable. - Use a stock loader on the boot floppy rather than one without forth. - Force the gzip of the mfsroot in release.9 to succeed. This is useful when doing lots of rereleases when testing floppy changes. - If an acpi kernel module exists, stick it on the boot floppy and for now always load it after the kernel.
2004-01-29 19:01:58 +00:00
.endif
Request boot floppy after loading mfsroot floppy because beastie.4th exists in boot floppy. Submitted by: Nobuyuki Koganemaru < n-kogane at syd dot odn dot ne dot jp > MFC after: 1 week
2006-01-02 07:29:22 +00:00
.if defined(SPLIT_MFSROOT)
@echo 'mfsroot_after="read -p \"Insert boot floppy and press Enter\""' >> ${IMAGEDIR}/boot/loader.conf
.endif
The AUTO_KEYBOARD_DETECT is useful on amd64 as well.
2006-10-24 21:00:49 +00:00
.if ${TARGET_ARCH} == "i386" || ${TARGET_ARCH} == "amd64"
.if ${AUTO_KEYBOARD_DETECT}
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
@echo "-P" >> ${IMAGEDIR}/boot.config
The AUTO_KEYBOARD_DETECT is useful on amd64 as well.
2006-10-24 21:00:49 +00:00
.endif
More support for Alpha installs. Submitted by: dfr
1999-01-29 11:39:04 +00:00
.endif
Overhaul our boot floppy generation system so that it scales better and requires minimal care and feeding for future releases. - Consolidate multitude of floppy related constants down to a normal FLOPPY set for 1.44 floppies and on PC98 a SMALLFLOPPY set for 1.2 floppies. Also, cleanup the i386 arch section by not duplicating constants that are the same on both machine types (i386 and pc98). - Update the ZIPNSPLIT macro to generate a file chunks that will actually fit onto 1.44 floppies formatted with UFS1. Unfortunately, split(1) seems to be somewhat buggy, so the files generated are slightly larger than the argument passed to split. - Split the release.10 target into 3 targers: floppies.1, floppies.2 and floppies.3 that are added to EXTRAS only if NOFLOPPIES is defined. floppies.1 builds the install floppies, floppies.2 builds the fixit floppy, and floppies.3 generates the md5 sums and READMEs for the floppies/ directory. - Drop the by now largely obsolete and less useful boot.flp picture. This was more useful when the mfsroot lived inside the kernel rather than being loaded from a separate file by the loader. - Only build a single mfsroot containing no modules that is used for all installation methods. - Use split-file.sh to split up a gzipped GENERIC kernel into however many floppies it takes for the boot kernel. Currently, a stock 5.2 GENERIC kernel including WITNESS, INVARIANTS, DDB, and other assorted bloat fits onto 2 additional floppies besides the boot floppy with some room to spare. - If SPLIT_MFSROOT is defined, the mfsroot.gz file is similar split into however many floppies are needed. Currently it is not defined as the mfsroot.gz fits onto the current boot.flp with room to spare. - Add a 'makeFloppySet' target which builds a floppy set for a file that was split using split-file.sh. - Rename the doMFSKERN target to 'buildBootFloppy' as that more closely matches what it does now. We no longer build a custom BOOTMFS kernel for each boot floppy. - We no longer build a 2.88 boot.flp image to use with emulated CD booting. The non-emulated cdboot works for almost everyone who boots off of CD and if it doesn't work on a particular machine, the user can always boot from the 1.44 floppy images. - We no longer build a driver floppy or stick kernel modules in the mfsroot since we now use a stock kernel when booting from floppy.
2004-01-26 19:58:38 +00:00
@rm -f ${IMAGEFILE}
@cp ${KERNFILE} ${MFSROOTFILE} ${IMAGEDIR}
sh -e ${DOFS_SH} ${IMAGEFILE} ${RD} ${MNT} ${FLPSIZE} ${IMAGEDIR} \
${BOOTINODE} ${FLPLABEL}
Hammer on the boot floppy generation bits some more. This is looking a little better.
1999-01-05 03:44:08 +00:00
@echo "Created ${RD}/floppies/${FSIMAGE}.flp"
mk file needed for 'make obj' done by $CHROOTDIR/mk
2003-02-23 06:27:21 +00:00
Add a new package splitting mechanism to replace print-cdrom-packages.sh. The new system tries to be more automated so that there is less work for the re's to do. It also no longer uses a /usr/ports tree as its input, but uses the generated package build including its INDEX file as its input. It parses the INDEX file, determines which packages should go on which ISO images, and then builds full-fledged trees of packages that can be added as an argument to mkisofs along with the tree built by 'make release' to build a full CD image. The INDEX files in the populated trees are generated with volume media number to make use of sysinstall's multiple volume support so that the user is kindly prompted to insert the appropriate disc for a package if it is not on the current disc. There is still some more tweaking to be done here, but this part needs to be committed. This stuff will all be used to prep the 5.4 release as well. Tested by: kensmith, others on re@ Reviewed by: re
2005-03-21 19:17:48 +00:00
.if make(package-split)
# Targets related to making a package split
#
# PKG_COPY instructs the script to copy the actual package files rather than
# make hard links
# PKG_TREE is the path to the package tree to be split
# PKG_DEST is the path to the destination tree to create the split in
# PKG_VERBOSE asks for verbose output of the layout process
Allow specifying which INDEX file to use by setting PKG_INDEX. By default the INDEX file is taken from the package source tree as defined by the PKG_TREE variable. This change allows using the (possibly incomplete) packages on pointyhat. MFC after: 2 days
2005-05-03 04:11:56 +00:00
# PKG_INDEX is the path to the INDEX file. By default ${PKG_TREE}/INDEX.
Add a new package splitting mechanism to replace print-cdrom-packages.sh. The new system tries to be more automated so that there is less work for the re's to do. It also no longer uses a /usr/ports tree as its input, but uses the generated package build including its INDEX file as its input. It parses the INDEX file, determines which packages should go on which ISO images, and then builds full-fledged trees of packages that can be added as an argument to mkisofs along with the tree built by 'make release' to build a full CD image. The INDEX files in the populated trees are generated with volume media number to make use of sysinstall's multiple volume support so that the user is kindly prompted to insert the appropriate disc for a package if it is not on the current disc. There is still some more tweaking to be done here, but this part needs to be committed. This stuff will all be used to prep the 5.4 release as well. Tested by: kensmith, others on re@ Reviewed by: re
2005-03-21 19:17:48 +00:00
.if defined(PKG_COPY)
PKG_DO_COPY= cp
.else
PKG_DO_COPY= ln
.endif
Allow specifying which INDEX file to use by setting PKG_INDEX. By default the INDEX file is taken from the package source tree as defined by the PKG_TREE variable. This change allows using the (possibly incomplete) packages on pointyhat. MFC after: 2 days
2005-05-03 04:11:56 +00:00
PKG_WRKIDX= ${PKG_DEST}/INDEX.master
Add a new package splitting mechanism to replace print-cdrom-packages.sh. The new system tries to be more automated so that there is less work for the re's to do. It also no longer uses a /usr/ports tree as its input, but uses the generated package build including its INDEX file as its input. It parses the INDEX file, determines which packages should go on which ISO images, and then builds full-fledged trees of packages that can be added as an argument to mkisofs along with the tree built by 'make release' to build a full CD image. The INDEX files in the populated trees are generated with volume media number to make use of sysinstall's multiple volume support so that the user is kindly prompted to insert the appropriate disc for a package if it is not on the current disc. There is still some more tweaking to be done here, but this part needs to be committed. This stuff will all be used to prep the 5.4 release as well. Tested by: kensmith, others on re@ Reviewed by: re
2005-03-21 19:17:48 +00:00
PKG_ENV?=
.if defined(TARGET_ARCH)
PKG_ENV+= PKG_ARCH=${TARGET_ARCH}
.endif
.if defined(PKG_VERBOSE)
PKG_ENV+= PKG_VERBOSE=1
.endif
Allow specifying which INDEX file to use by setting PKG_INDEX. By default the INDEX file is taken from the package source tree as defined by the PKG_TREE variable. This change allows using the (possibly incomplete) packages on pointyhat. MFC after: 2 days
2005-05-03 04:11:56 +00:00
PKG_INDEX?= ${PKG_TREE}/INDEX
Add a new package splitting mechanism to replace print-cdrom-packages.sh. The new system tries to be more automated so that there is less work for the re's to do. It also no longer uses a /usr/ports tree as its input, but uses the generated package build including its INDEX file as its input. It parses the INDEX file, determines which packages should go on which ISO images, and then builds full-fledged trees of packages that can be added as an argument to mkisofs along with the tree built by 'make release' to build a full CD image. The INDEX files in the populated trees are generated with volume media number to make use of sysinstall's multiple volume support so that the user is kindly prompted to insert the appropriate disc for a package if it is not on the current disc. There is still some more tweaking to be done here, but this part needs to be committed. This stuff will all be used to prep the 5.4 release as well. Tested by: kensmith, others on re@ Reviewed by: re
2005-03-21 19:17:48 +00:00
package-split:
.if !defined(PKG_TREE)
@echo "PKG_TREE must be defined" && exit 1
.endif
.if !defined(PKG_DEST)
@echo "PKG_DEST must be defined" && exit 1
.endif
@env ${PKG_ENV} python ${.CURDIR}/scripts/package-split.py \
Allow specifying which INDEX file to use by setting PKG_INDEX. By default the INDEX file is taken from the package source tree as defined by the PKG_TREE variable. This change allows using the (possibly incomplete) packages on pointyhat. MFC after: 2 days
2005-05-03 04:11:56 +00:00
${PKG_INDEX} ${PKG_WRKIDX}
Add a new package splitting mechanism to replace print-cdrom-packages.sh. The new system tries to be more automated so that there is less work for the re's to do. It also no longer uses a /usr/ports tree as its input, but uses the generated package build including its INDEX file as its input. It parses the INDEX file, determines which packages should go on which ISO images, and then builds full-fledged trees of packages that can be added as an argument to mkisofs along with the tree built by 'make release' to build a full CD image. The INDEX files in the populated trees are generated with volume media number to make use of sysinstall's multiple volume support so that the user is kindly prompted to insert the appropriate disc for a package if it is not on the current disc. There is still some more tweaking to be done here, but this part needs to be committed. This stuff will all be used to prep the 5.4 release as well. Tested by: kensmith, others on re@ Reviewed by: re
2005-03-21 19:17:48 +00:00
@env ${PKG_ENV} sh ${.CURDIR}/scripts/package-trees.sh ${PKG_DO_COPY} \
Allow specifying which INDEX file to use by setting PKG_INDEX. By default the INDEX file is taken from the package source tree as defined by the PKG_TREE variable. This change allows using the (possibly incomplete) packages on pointyhat. MFC after: 2 days
2005-05-03 04:11:56 +00:00
${PKG_WRKIDX} ${PKG_TREE} ${PKG_DEST}
Add a new package splitting mechanism to replace print-cdrom-packages.sh. The new system tries to be more automated so that there is less work for the re's to do. It also no longer uses a /usr/ports tree as its input, but uses the generated package build including its INDEX file as its input. It parses the INDEX file, determines which packages should go on which ISO images, and then builds full-fledged trees of packages that can be added as an argument to mkisofs along with the tree built by 'make release' to build a full CD image. The INDEX files in the populated trees are generated with volume media number to make use of sysinstall's multiple volume support so that the user is kindly prompted to insert the appropriate disc for a package if it is not on the current disc. There is still some more tweaking to be done here, but this part needs to be committed. This stuff will all be used to prep the 5.4 release as well. Tested by: kensmith, others on re@ Reviewed by: re
2005-03-21 19:17:48 +00:00
.endif
At least we don't need the whole <bsd.prog.mk> glue. (I did test build the original change, but apparently forgot to include it to LOCAL_PATCHES. It happens.)
2003-02-23 19:15:55 +00:00
.include <bsd.obj.mk>
Reference in New Issue Copy Permalink