Some of the major changes include:
- The SCSI error handling portion of cam_periph_error() has
been broken out into a number of subfunctions to better
modularize the code that handles the hierarchy of SCSI errors.
As a result, the code is now much easier to read.
- String handling and error printing has been significantly
revamped. We now use sbufs to do string formatting instead
of using printfs (for the kernel) and snprintf/strncat (for
userland) as before.
There is a new catchall error printing routine,
cam_error_print() and its string-based counterpart,
cam_error_string() that allow the kernel and userland
applications to pass in a CCB and have errors printed out
properly, whether or not they're SCSI errors. Among other
things, this helped eliminate a fair amount of duplicate code
in camcontrol.
We now print out more information than before, including
the CAM status and SCSI status and the error recovery action
taken to remedy the problem.
- sbufs are now available in userland, via libsbuf. This
change was necessary since most of the error printing code
is shared between libcam and the kernel.
- A new transfer settings interface is included in this checkin.
This code is #ifdef'ed out, and is primarily intended to aid
discussion with HBA driver authors on the final form the
interface should take. There is example code in the ahc(4)
driver that implements the HBA driver side of the new
interface. The new transfer settings code won't be enabled
until we're ready to switch all HBA drivers over to the new
interface.
src/Makefile.inc1,
lib/Makefile: Add libsbuf. It must be built before libcam,
since libcam uses sbuf routines.
libcam/Makefile: libcam now depends on libsbuf.
libsbuf/Makefile: Add a makefile for libsbuf. This pulls in the
sbuf sources from sys/kern.
bsd.libnames.mk: Add LIBSBUF.
camcontrol/Makefile: Add -lsbuf. Since camcontrol is statically
linked, we can't depend on the dynamic linker
to pull in libsbuf.
camcontrol.c: Use cam_error_print() instead of checking for
CAM_SCSI_STATUS_ERROR on every failed CCB.
sbuf.9: Change the prototypes for sbuf_cat() and
sbuf_cpy() so that the source string is now a
const char *. This is more in line wth the
standard system string functions, and helps
eliminate warnings when dealing with a const
source buffer.
Fix a typo.
cam.c: Add description strings for the various CAM
error status values, as well as routines to
look up those strings.
Add new cam_error_string() and
cam_error_print() routines for userland and
the kernel.
cam.h: Add a new CAM flag, CAM_RETRY_SELTO.
Add enumerated types for the various options
available with cam_error_print() and
cam_error_string().
cam_ccb.h: Add new transfer negotiation structures/types.
Change inq_len in the ccb_getdev structure to
be "reserved". This field has never been
filled in, and will be removed when we next
bump the CAM version.
cam_debug.h: Fix typo.
cam_periph.c: Modularize cam_periph_error(). The SCSI error
handling part of cam_periph_error() is now
in camperiphscsistatuserror() and
camperiphscsisenseerror().
In cam_periph_lock(), increase the reference
count on the periph while we wait for our lock
attempt to succeed so that the periph won't go
away while we're sleeping.
cam_xpt.c: Add new transfer negotiation code. (ifdefed
out)
Add a new function, xpt_path_string(). This
is a string/sbuf analog to xpt_print_path().
scsi_all.c: Revamp string handing and error printing code.
We now use sbufs for much of the string
formatting code. More of that code is shared
between userland the kernel.
scsi_all.h: Get rid of SS_TURSTART, it wasn't terribly
useful in the first place.
Add a new error action, SS_REQSENSE. (Send a
request sense and then retry the command.)
This is useful when the controller hasn't
performed autosense for some reason.
Change the default actions around a bit.
scsi_cd.c,
scsi_da.c,
scsi_pt.c,
scsi_ses.c: SF_RETRY_SELTO -> CAM_RETRY_SELTO. Selection
timeouts shouldn't be covered by a sense flag.
scsi_pass.[ch]: SF_RETRY_SELTO -> CAM_RETRY_SELTO.
Get rid of the last vestiges of a read/write
interface.
libkern/bsearch.c,
sys/libkern.h,
conf/files: Add bsearch.c, which is needed for some of the
new table lookup routines.
aic7xxx_freebsd.c: Define AHC_NEW_TRAN_SETTINGS if
CAM_NEW_TRAN_CODE is defined.
sbuf.h,
subr_sbuf.c: Add the appropriate #ifdefs so sbufs can
compile and run in userland.
Change sbuf_printf() to use vsnprintf()
instead of kvprintf(), which is only available
in the kernel.
Change the source string for sbuf_cpy() and
sbuf_cat() to be a const char *.
Add __BEGIN_DECLS and __END_DECLS around
function prototypes since they're now exported
to userland.
kdump/mkioctls: Include stdio.h before cam.h since cam.h now
includes a function with a FILE * argument.
Submitted by: gibbs (mostly)
Reviewed by: jdp, marcel (libsbuf makefile changes)
Reviewed by: des (sbuf changes)
Reviewed by: ken
manual section. If, for example, MANSECT is set to 8, the default
MAN1=${PROG}.1 feature of bsd.prog.mk becomes MAN8=${PROG}.8.
Useful for games, libexec, sbin and usr.sbin subtrees.
Reviewed by: bde
and Pentium II, III and IV processors (p2, p3, p4), as well as 'mmx' and
'3dnow' MACHINE_CPU tags as appropriate. In the near future this will
be used to control various ports which have MMX/3dNow optimizations,
instead of the ad-hoc methods currently used.
Reviewed by: peter
* Rip out MACHINE_CPU stuff from sys.mk and include a new <bsd.cpu.mk>
after we pull in /etc/make.conf. We need to do it afterwards so we can
react to the user setting of the:
* CPUTYPE variable, which contains the CPU type which the user wants to
optimize for. For example, if you want your binaries to only run on an
i686-class machine (or higher), set this to i686. If you want to support
running binaries on a variety of CPU generations, set this to the lowest
common denominator. Supported values are listed in make.conf.
* bsd.cpu.mk does the expansion of CPUTYPE into MACHINE_CPU using the
(hopefully) correct unordered list of CPU types which should be used on
that CPU. For example, an AMD k6 CPU wants any of the following:
k6 k5 i586 i486 i386
This is still an unordered list so the client makefile logic is simple -
client makefiles need to test for the various elements of the set in
decreasing order of priority using ${MACHINE_CPU:M<foo>}, as before.
The various MACHINE_CPU lists are believed to be correct, but should be
checked.
* If NO_CPU_CFLAGS is not defined, add relevant gcc compiler optimization
settings by default (e.g. -karch=k6 for CPUTYPE=k6, etc). Release
builders and developers of third-party software need to make sure not to
enable CPU-specific optimization when generating code intended to be
portable. We probably need to move to an /etc/world.conf to allow the
optimization stuff to be applied separately to world/kernel and external
compilations, but it's not any worse a problem than it was before.
* Add coverage for the ia64/itanium MACHINE_ARCH/CPUTYPE.
* Add CPUTYPE support for all of the CPU types supported by FreeBSD and gcc
(only i386, alpha and ia64 first, since those are the minimally-working
ports. Other architecture porters, please feel free to add the relevant
gunk for your platform).
Reviewed by: jhb, obrien
* Rip out MACHINE_CPU stuff from sys.mk and include a new <bsd.cpu.mk>
after we pull in /etc/make.conf. We need to do it afterwards so we can
react to the user setting of the:
* CPUTYPE variable, which contains the CPU type which the user wants to
optimize for. For example, if you want your binaries to only run on an
i686-class machine (or higher), set this to i686. If you want to support
running binaries on a variety of CPU generations, set this to the lowest
common denominator. Supported values are listed in make.conf.
* bsd.cpu.mk does the expansion of CPUTYPE into MACHINE_CPU using the
(hopefully) correct unordered list of CPU types which should be used on
that CPU. For example, an AMD k6 CPU wants any of the following:
k6 k5 i586 i486 i386
This is still an unordered list so the client makefile logic is simple -
client makefiles need to test for the various elements of the set in
decreasing order of priority using ${MACHINE_CPU:M<foo>}, as before.
The various MACHINE_CPU lists are believed to be correct, but should be
checked.
* If NO_CPU_CFLAGS is not defined, add relevant gcc compiler optimization
settings by default (e.g. -karch=k6 for CPUTYPE=k6, etc). Release
builders and developers of third-party software need to make sure not to
enable CPU-specific optimization when generating code intended to be
portable. We probably need to move to an /etc/world.conf to allow the
optimization stuff to be applied separately to world/kernel and external
compilations, but it's not any worse a problem than it was before.
* Add coverage for the ia64/itanium MACHINE_ARCH/CPUTYPE.
* Add CPUTYPE support for all of the CPU types supported by FreeBSD and gcc
(only i386, alpha and ia64 first, since those are the minimally-working
ports. Other architecture porters, please feel free to add the relevant
gunk for your platform).
Reviewed by: jhb, obrien
set the variable until you rebuild it, and the alternative is to be stuck
playing games with ``.if defined(MACHINE_CPU) && ... '' for all eternity.
We now set up the reasonable default for i386 and alpha here -- given this
it probably makes sense to remove the corresponding code from make(1).
through the use of a new build directive, MACHINE_CPU, which contains a
list of the CPU generations/features for which optimizations are desired.
This feature will be extended to cover the ports tree in the future.
Currently OpenSSL provides optimizations for i386, i586 and i686-class
CPUs. Currently it has not been tested on an i386 or i486.
Teach make(1) to provide sensible defaults for MACHINE_CPU if it is not
defined (namely, the lowest common denominator CPU we support for each
architecture). Currently this is i386 for the i386 architecture and ev4
for the alpha. sys.mk also sets the variable as a last resort for
consistency with MACHINE_ARCH and bootstrapping from very old versions of
make.
Benchmarks show a significant speed increase even in the i386 case, with
additional improvements for i586 and i686 systems. For maximum performance
define MACHINE_CPU=i686 i586 i386 in /etc/make.conf.
Based on a patch submitted by: Mike Silbersack <silby@silby.com>
Reviewed by: current
.PATH to ${.CURDIR}/[...]/kern , the "exists" expression will fail for the
form exists(${.CURDIR}/[...]/kern/). This appears to be happening because
make is searching for the argument to "exists" by using .PATH rather than a
relative search, because .PATH and the argument match at the beginning.
Additionally, make appears to consider a path that starts with ${.CURDIR}
as relative, even though it expands to an absolute path.
The reason that most people aren't seeing this problem is that the absolute
paths of /usr/src/sys and /sys are also searched, so as long as the kernel
source can be found in at least one of those places, no problems surface.
This problem was inadvertently introduced on 1 December 2000, with the
addition of the sysvipc modules.
bind distribution, but until now was not being built as a separate
entity. For documentation, see these man pages:
assertions(3), eventlib(3), heap(3), logging(3), memcluster(3), tree(3).
Reviewed by: jdp
${LIB} library". "standard" tends to imply the one that is normally
used... but by default it is not the case - the .so would be the
"standard" library. Therefore, change this to 'static'. Another option
might be "conventional ${LIB} library".
by -n is nonexistant, then the following -d was misinterpreted with
a strange error. By putting double quotes (") around the argument,
we can be sure there is _something_ there that we can check a zero
length against.
cause the working directory to be used. Make it so.
When we're more convinced that it'll work, we might try this
to avoid a shell invocation:
.if defined(MAKEOBJDIRPREFIX) && !empty(MAKEOBJDIRPREFIX) &&
exists(${CANONICALOBJDIR}/)
Reported by: bde
there happens to be a source file named install.sh. The null rule
for "install" in the NOINFO case must not be completely null, since
then it may be overridden by the implicit .sh rule.
via the MODULE_VERSION() and MODULE_DEPEND() macros that both the loader
and kld system know how to deal with. The old DT_NEEDED tag is still
supported by the loader (and will remain supported for a while) - but the
kernel side presently doesn't know how to deal with DT_NEEDED.
do not have the kernel you wish to compile against in either
/usr/src/sys or /sys, then you will need to set SYSDIR to point to the
sys directory of the source tree that contians the source.
Also, minor tweaks to the load/unload targets from Bruce.
I've had this through several make worlds, as well as using it on a
daily basis for the past couple of weeks to build modules needed for
testing at Timing Solutions.
Reviewed and revised by: bde
Work sponsored by: Timing Solutions
it into a ``shared'' .ko file. This intermediate file can be directly
linked into a static kernel. This isn't all that useful yet but will
become much more interesting shortly.
essentially as in kernel makefiles, so that module sources can include
<stddef.h> and other standard headers. Only add the second path when
the first path can't be found, instead of when DESTDIR is defined.
Adding it used to be just an obfuscation.
Use "${.OBJDIR}" instyead of "." in -I paths. Using "${.OBJDIR}" just
gave more verbose command lines and depend files.
is an application space macro and the applications are supposed to be free
to use it as they please (but cannot). This is consistant with the other
BSD's who made this change quite some time ago. More commits to come.
we use. The .c half is statically compiled into the kernel. It's kinda
silly to generate a .h file on the fly that has inlines to call the
.c stuff when the .c code is fixed.
Also, zap the special treatment for VFS_KLD modules. This treatment
applies to lots of things, not just VFS's.
files (opt_*.h) automatically (if they are in ${SRCS}).
Clean vnode_if.[ch] automatically (if one of them is in ${SRCS}, not just
if VFS_KLD is defined).
There are some complications to avoid using the "@" symlink before it
is built.
All Makefiles now use MACHINE_ARCH for the target architecture.
Unification is required for cross-building.
Tags added to:
sys/boot/Makefile
sys/boot/arc/loader/Makefile
sys/kern/Makefile
usr.bin/cpp/Makefile
usr.bin/gcore/Makefile
usr.bin/truss/Makefile
usr.bin/gcore/Makefile:
fixed typo: MACHINDE -> MACHINE_ARCH
Been in production for 3 years now. Gives Instant Frame relay to if_sr
and if_ar drivers, and PPPOE support soon. See:
ftp://ftp.whistle.com/pub/archie/netgraph/index.html
for on-line manual pages.
Reviewed by: Doug Rabson (dfr@freebsd.org)
Obtained from: Whistle CVS tree
2) s/MODLOAD/KMODLOAD/ to be consistent with the rest of the variables
(KMOD, KMODOWN, KMODGRP, etc) and definition of MODLOAD/UNLOAD in the
Makefile of the ATAPI module
3) textual fixups
Specifically intended for removing -fschg ("INSTALLFLAGS_EDIT=:S/schg/uchg/")
this makes the NOFSCHG flag redundant. NOFSCHG will still be honoured by
bsd.lib.mk but is valid for buildworld only. NOFSCHG is still implemented in
the old way (ie. _not_ ".if NOFSCHG then { INSTALLFLAGS_EDIT+=:S/schg/,/ }"
to emphasize the fact that NOFSCHG is only supported in a limited
fashion and for buildworld.
The interface and implementation are such that future use of flags such
as sappnd can also be easily removed or altered (perhaps to uappnd).
This commit brought to you by the letters B, D, and E, and the numbers six,
one, thirteen, and three.
-DNOFSCHG disables installation of libs with flag schg
GAMEGRP change the group with which games are installed
also organize the binary section into alphebetical order some what..
/usr/sbin/sysctl -> ${DESTDIR}/sbin/sysctl in some versions of 2.2,
and this link was broken if DESTDIR was set.
Added a SYMLINKS macro. This works the same as LINKS, except it
creates symlinks and the linked-to pathname may be relative. This
is more flexible than LN_FLAGS, since it supports installing
symlinks independently of hard links.
Use `ln -f[s] ...' instead of `rm -f ...; ln [-s] ...' for LINKS and
SYMLINKS. This is equivalent if the target is neither a directory nor
a symlink to a directory.
PR: 8279
is not implied by -Wall as claimed by gcc.1. Adding it causes a
measly 7193 new warnings for LINT, mostly for "unused parameter" and
"comparison between signed and unsigned".
There are only skeletons left here; they merely serve as a backup to
include the real versions under ${PORTSDIR}/Mk while we update the ports
tree to include them directly.
<bsd.libnames.mk> is included regardless of the object file format.
This is needed to fix the a.out PAM breakage that manifests itself
when trying to build login.
When linking statically, LIBPAM is augmented with the extra libraries
that the PAM modules require. The idea is to centralize this
information rather than scattering it about in the Makefiles of
all the applications that use (OK, will use) PAM.
There is a new variable MINUSLPAM that should be used instead of
"-lpam". In the static case, it gets -l flags for the extra required
libraries.
This approach was suggested by <bde>, but he didn't actually review
my changes.
building dlopen-able modules, and add features needed to build a
static PAM library. I think I cleaned it up some, too, but beauty
is in the eye of the beholder.
You can now build a shared library without version numbers, by
defining SHLIB_NAME to something like "pam_unix.so". If SHLIB_MAJOR
and/or SHLIB_MINOR are set, SHLIB_NAME gets the usual default value,
but it can be overridden if desired. If none of these symbols are
set, no shared library is built.
SHLIB_LINK controls the name of the symbolic link that points to
the library. If it is unset, no link is made. In the usual case,
it gets the right default: e.g., "libc.so" for ELF, nothing for
a.out. This can be overridden.
STATICOBJS can be set to a list of extra object files that should
be added to the static library but not to the shared library.
These objects are added to the profiled library too.
These changes should make it easy to use <bsd.lib.mk> for building
things such as PAM modules and dynamic linkers, for which <bsd.prog.mk>
has been abused until now.
(as given by "uname -m") by which you can specify which
architectures the port is appropriate for.
Idea borrowed from: NetBSD, OpenBSD
Reviewed by: simokawa
(2) New variable PERL_ARCH (value: ${ARCH}-freebsd) that is also
passed to ${PLIST_SUB}. Use it to simplify PLISTs.
Submitted by: simokawa
(3) Check OSVERSION as well as existence of /usr/bin/perl5 before
assuming USE_PERL5 is to be a no-op. Basically to allow building
of 2.2-INDEX on a 3.0 machine.
(4) Change USE_QT definition to use new shlib version (2) and
directory (qt142).
(5) Uncomment temporary Motif dependency for parallel package
building. We still need to figure out a way to install the pkg
database files, but it's a start.
(6) Move EXTRACT_SUFX up into the pre.mk area so it can be used in
exists() tests.
(7) Add MASTER_SITE_COMP_SOURCES. Note that this is defined like
"/pub/usenet/comp.sources.%SUBDIR%/" so you can specify something
like "SUBDIR=x/volume18" to select the newsgroup as well as
subdirectory name.
Submitted by: "distfiles" fenner
(8) Other misc. master site cleanup.
Submitted by: "distfiles" fenner
(9) New target "maintainer". I intend to use it to auto-mail failure
build failure notices to the maintainer.
than ".so". The old extension conflicted with well-established
naming conventions for dynamically loadable modules.
The "clean" targets continue to remove ".so" files too, to deal with
old systems.
don't recurse in "make describe". The new INDEX target in
ports/Makefile invokes a perl script to recurse and convert them
into package names.
While I'm here, change the name of targets and move them around a
little bit for the sake of consistency.
It is also probably worth noting here that the meaning of the
"build dependency" list in INDEX has been changed slightly
changed. The old list was "build depends and its build depends"
-- not particularly useful if you had things like autoconf, which
run-depend on gm4 (you install all the things listed here and
you'll get an autoconf that won't run).
It is now "build depends and its run depends" -- you install
everything listed here, and you'll be able to build the port.
Submitted by: steve
(0') Fast README.html generation. It uses ports/INDEX to find
dependencies instead of embarking on to a recursive loop.
Submitted by: steve
(1) Remove NO_WRKDIR and NO_EXTRACT. Their functionality are easily
replacable with NO_WRKSUBDIR=t and EXTRACT_ONLY= (nothing on right
side), and they get in the way of read-only port trees.
(2) Surround first few variable definitions with ".if !defined()".
This will make cross-compilation easier and also speed up make
processes.
(3) Call sysctl with absolute path. Prefer the one in /sbin over the
one in /usr/sbin.
(4) Add four new variables
PKGINSTALL?= ${PKGDIR}/INSTALL
PKGDEINSTALL?= ${PKGDIR}/DEINSTALL
PKGREQ?= ${PKGDIR}/REQ
PKGMESSAGE?= ${PKGDIR}/MESSAGE
and use them in PKG_ARGS. Frobbing with PKG_ARGS directly is
strongly discouraged.
(5) Change PKG_SUFX to ".tar" (instead of ".tgz") if PKG_NOCOMPRESS is
defined. This is intended only for our own use.
(6) Add more sites to MASTER_SITE_GNU.
Submitted by: billf
(7) Override MANUAL_PACKAGE_BUILD if PARALLEL_PACKAGE_BUILD is
defined. This is intended only for our own use.
(8) Add new target "ignorelist" which will print out the package name
if the port is not going to be built on this machine. This is
intended only for our own use.
(9) Make mtree a little quieter.
break one way or another. With it goes the package-loop and the
describe loop.
(2) Add new variable MASTERDIR to make it easier to share files between ports.
bsd.port.mk will find things like ${PKGDIR} underneath
${MASTERDIR} (which defaults to ${.CURDIR}).
(3) Do not allow MD5_FILE to be renamed. Funny things can happen if
you do that.
(4) Use a few more absolute paths in the bsd.port.pre.mk part. I
can't use absolute paths for sysctl because it moved recently.
a few variables that could be used in the port Makefile for ".if exists()"
tests. bsd.port.post.mk defines the rest.
Note: if you define USE_X_PREFIX or USE_IMAKE, put it before including
bsd.port.pre.mk. These are the only two variables used in the first part.
In reality, bsd.port.pre.mk and bsd.port.post.mk just include bsd.port.mk
with special variables to turn part of it off.
list. The old MAKE_FLAGS was a little hard to use since it had a weird
default ("-f").
Suggested by: Shigeyuki FUKUSHIMA <shige@kuis.kyoto-u.ac.jp>
(2) Add new targets clean-restricted and clean-for-cdrom, which will
delete RESTRICTED and NO_CDROM packages and distfiles from the top.
Reviewed by: jkh
(3) Add depends to list of things to recurse on. It will help people
who are trying to fetch some ports plus their dependencies.
Requested by: Chris Dillon <cdillon@wolves.k12.mo.us>
packages from a single port. LOOP_VAR is the name of the variable
and LOOP_OPTIONS is a space-separated list of values it should
take. When these are set, the target "package-loop" will go
through a clean and package loop for all the options. The
"package-loop" target is defined as "package" when LOOP_VAR is not
defined, so if you are in the business for building packages, you
should use "package-loop" all the time. (This target is added to
bsd.port.subdir.mk too.)
Also, the "describe" target prints out multiple lines so that all
options will go into the INDEX. (In other words, if you define
these variables, INDEX is going to look real silly if you don't
put ${${LOOP_VAR}} in PKGNAME.)
Seconded by: obrien ("ANYTHING")
(2) Turn off regexp support for LIB_DEPENDS. It is a fixed string of
the form <NAME>.<VER> now.
Tested by: several rounds of complete package builds
(3) Check checksum even if NO_EXTRACT is defined.
(4) Cosmetic fix for message in MANUAL_PACKAGE_BUILD case.
Alpha. This is a minor, but important distinction. Should be a no-op
to the install base. If OBJFORMAT is set elsewhere, things work
exactly as they did before.
name conversion. Use it for binary ports that come with its own private
shlib dirs, ports that install linux compatibility libraries (thus following
their naming conventions and not ours), etc.
Reviewed by: Martin Cracauer <cracauer@FreeBSD.ORG>
version number part (i.e., "<directory>/perl"). Use this to
substitute #! lines in your perl5 scripts.
Requested and reviewed by: ache
(2) Add new variable WRKDIRPREFIX (defaults to ""). The "work"
directories are now in ${WRKDIRPREFIX}${.CURDIR}/work by default.
You can have a read-only ports tree (modulo any broken ports that
write something to places other than ${WRKDIR}) by setting this to
a writable location.
Ports that set WRKDIR explicitly should append this to front so
they will work when the user has WRKDIRPREFIX set.
Reviewed by: Toshihiko Kodama <kodama@ayame.mfd.cs.fujitsu.co.jp>
since ports are not supposed to be broken during the process of
conversion to ELF -- please proceed with caution.)
(2) Support for checking file size before fetching. The essential
parts are commented out for now, so I won't lose the submission
while we discuss how to do it.
Submitted by: se (mostly)
(3) Don't run "fetch" twice. It was due to the change in checksum
target chaining. It used to be fetch -> checksum -> extract,
after 1.285 it was fetch -> checksum and checksum was also
explicitly called from extract. Fix it by not calling fetch from
checksum when it's invoked by extract.
Noticed by: pre-fetch target of lesstif being run twice
(4) Don't try to remove non-existent distfiles and patchfiles in
distclean.
Submitted by: anto@netscape.net
PR: 7988
Submitted by: "Eugene M. Kim" <gene@nttlabs.com>
(2) Check for exact version of perl5 in /usr/bin and exit with error
message if USE_PERL5 is defined and version does not match
${PERL_VERSION}.
(3) Note LIB_DEPENDS should not have any regular expressions. Remove
those in USE_XLIB and USE_QT.
Submitted by: hoek
(2) ELF support.
2a. Variable PORTOBJFORMAT specifies the object format of the system.
It is passed down to configure/make via CONFIGURE_ENV/MAKE_ENV,
and is given to generate-plist via PLIST_SUB.
2b. In PLIST, substitute lines that end with "/libFOO.so.X" with
"/libFOO.so.X.0". (This means PLISTs should only list ELF
libraries.)
Reviewed by: jb, jdp, hoek, jseger, steve
(3) Perl5-in-system support. Basically turns USE_PERL5 into a no-op
if there exists a "/usr/bin/perl5". Also fix prior breakage by
dima (${PREFIX} => ${LOCALBASE} in perl5 path).
Reviewed by: markm (sort of)
(4) Install requirement file as "+REQUIRE" so it will be executed
correctly by pkg_delete.
Reported by: Shigeyuki FUKUSHIMA <shige@kuis.kyoto-u.ac.jp>
(5) Do not disable checksum and makesum when NO_EXTRACT is set.
Pointed out by: hoek and kiri, among other people
src/lkm/Makefile.inc. This fixes broken builds of the syscons LKMs
when OBJFORMAT=elf. Removed src/lkm/Makefile.inc since it became
empty and is worse than useless.
the way, most of the things below are "Reviewed by: hoek" as well.)
The changes are (roughly in order of appearance):
(1) Revamp comments at beginning of file. Major rewrites. Reorder
them into more relevant sections. Make clear which ones are user
variables and not to be set in ports' Makefiles.
Reviewed by: hoek (well, he's the only one who sent any comments)
(2) Include ${.CURDIR}/Makefile.local if it exists. This is a local
configuration file (ala rc.conf.local and make.conf.local) so
please do not commit a file with this name to the repository.
Suggested by: dillon
(3) MANCOMPRESSED now takes three possible values: "yes", "no" and
"maybe". (It used to be a binary variable -- the old behavior is
now accomplished by "MANCOMPRESSED=yes". Ports that defined this
variable to other values have been corrected.)
"yes" means the manpages are installed compressed, "no" means they
are not, and "maybe" means the port already respects the value of
NOMANCOMPRESS. The default is "yes" for USE_IMAKE ports without
NO_INSTALL_MANPAGES, and "no" otherwise.
Add "compress-man" target and move manpage {,de}compression there.
Reviewed by: hoek etc. and a full build of the ports tree
(4) Add LIBDIR="${LIBDIR}" to MAKE_ENV. Makes easier to "fix" ports
for /usr/lib/aout mess.
Submitted by: ohashi@mickey.ai.kyutech.ac.jp (Takeshi Ohashi)
(5) Change ${TMPPLIST} from ${WRKDIR}/PLIST.mktmp to
${WRKDIR}/.PLIST.mktmp.
Suggested by: hoek
Strongly seconded by: steve
(6) Change a couple more relative pathnames to absolute ones.
Submitted by: hoek
(7) Move checksum into real-extract.
(8) Change way rules are chained. Instead of:
build: configure ${BUILD_COOKIE}
${BUILD_COOKIE}:
@cd ${.CURDIR} && ${MAKE} ${.MAKEFLAGS} real-build
we now have
build: ${BUILD_COOKIE}
${BUILD_COOKIE}:
@cd ${.CURDIR} && ${MAKE} ${.MAKEFLAGS} configure
@cd ${.CURDIR} && ${MAKE} ${.MAKEFLAGS} real-build
Other than being more PC (pmake clean), this really speeds up
skipping ports already built. For instance, "make package" on a
fully packaged games subtree used to take 269 seconds on average,
now it's 45 seconds on average.
The flip side of this is that it will create more processes when
the targets actually chain, but when you're actually compiling
things, your make is creating half a bazzilion processes anyway so
I don't think it matters.
(9) ${TMPPLIST} is now a real dependency. Create "generate-plist"
target to generate ${TMPPLIST}. Make sure it's called when needed
(usually between do-install and post-install), and that the
required files exist.
(10) Change some messages so we can tell where "make index" failed.
(11) Check if LIB_DEPENDS really generated the required shared lib or
not.
Seconded by: "Brent J. Nordquist" <bjn@visi.com>
when certain .mk files include other .mk files. This will remove the
need for multiple include protection in some other makefiles around the
tree (and helps some elf conditionals).
means (change PREFIX to ${X11BASE}) and add a new variable USE_X11
which means "this port requires the X window system (actually the
library)". USE_X_PREFIX implies USE_X11. USE_X11 adds a
LIB_DEPENDS to libX11 with the x11/XFree86 port.
Reviewed by: the ports list, hoek in particular
(2) Remove NO_CONFIGURE and NO_PATCH, which never meant anything.
Yell if they are defined.
Reviewed by: the ports list
(3) Add new variable OSREL, which is automatically set to the
numeric OS version (e.g., 2.2.1, 3.0). It can be used to refer to
files in gcc installations, for instance.
(4) Finally remove EXEC_DEPENDS hack after all these years.
Submitted by: hoek
(5) Put quotes around some echo ${*_DEPENDS} statements so they won't
blow up when the variables include regular expressions like
"qt\\.1\\.\\\(33\\\|40\\\):${PORTSDIR}/x11/qt140".
recently added definitions from sys.mk to bsd.own.mk. Include the
src-relative bsd.own.mk in src/Makefile to pick up all new definitions.
Don't check that MACHINE_ARCH is defined in src/Makefile, since it is
(and should have been) guaranteed to be defined.
(mostly for includes) separate from direct dependencies (so that ${.ALLSRC}
can be used to find full paths to the sources for the direct dependencies
only). The `::' hack just forgot the indirect dependendencies. This
broke building doc/usd/13.viref with `make -jN' - the index got corrupted
by being built twice concurrently.
Cleaned up the ${DFILE} rule. There was a .else clause with dead code in
it following a .else clause (make accepts this bad syntax). ${.ALLSRC}
now works in the USE_SOELIMPP case. Some client Makefiles no longer need
the SRCDIR=${.OBJDIR} hack.
`make world' to avoid problems with picking up (new) target or (stale)
host shared libraries.
Don't honor -static in LDFLAGS for linking LKMs. LDFLAGS is not
actually for ld, but we use it anyway, and must prevent -static being
misinterpreted as -s.
Don't hide any of the link steps.
Move a.out libraries to /usr/lib/aout to make space for ELF libs.
Make rtld usr /usr/lib/aout as default library path.
Make ldconfig reject /usr/lib as an a.out library path.
Fix various Makefiles for LIBDIR!=/usr/lib breakage.
This will after a make world & reboot give a system that no
longer uses /usr/lib/*, infact one could remove all the old
libraries there, they are not used anymore.
We are getting close to an ELF make world, but I'll let this
all settle for a week or two...
the right solution or not, bsd.port.mk is broken unless bsd.locale.mk
is installed.
Note that if LOCALE is not defined, port-building explodes:
"/usr/share/mk/bsd.locale.mk", line 135: if-less elif
"/usr/share/mk/bsd.locale.mk", line 135: Need an operator
(For each .if testing LOCALE)
Notes:
- We no longer use -fgnu-runtime in bsd.lib.mk, since it is the default
and bsd.lib.mk is the wrong place to override it.
- Gnu C doesn't have a special compiler driver for Objective C like it
does for C++. The defaults are suitable for Gnu C. Use `OBJCLIBS='
in /etc/make.conf for POC.
is used in the dependency list for ${DEPENDFILE}. `make depend' was
broken for a few days. `make world' only uses `make depend' when
NOCLEAN is defined, so only a few people noticed the bug.
Submitted by: mostly by jmg
bsd.obj.mk instead of bsd.dep.mk for defining the _SUBDIR target
and a default tags target. Abuse bsd.obj.mk for defining default
cleandepend and depend targets.
that want a y.tab.h file. This want must be specified by putting y.tab.h
in SRCS (and defaulting to or putting -d in YFLAGS). This only works if
there is only one yacc parser, of course. One improvement: copy y.tab.c
to foo.c instead of renaming it, so that `#line...y.tab.c' statements in
it refer to an existing file.
Regress to not generating explicit rules for .l and .y sources containing
slashes. This case is unusual and hard to handle properly.
Don't generate an unused dependency when -d is not in YFLAGS.
sources. This will be used to fix `make -jN' races in many
Makefiles in /usr/src, and to simplify these and other Makefiles.
To get the fixes and simplifications, application Makefiles should
put the raw lex and yacc source names (foo.l and bar.y) in SRCS and
not put the names of any generated files in SRCS or CLEANFILES. A
few Makefiles already do this, although it didn't actually work before
(mkdep couldn't create complete dependencies because there is no
intermediate .c file for the .y.o and .l.o implicit rules).
Complications: if bar.y is in SRCS and -d is in YFLAGS, then bar.h
will be generated whether or not it is used, even if this clobbers a
real source file. This is so that bar.c can be generated using the
-j-safe and debugger-friendly -o option to YACC. There are smaller
warts for handling y.tab.h. y.tab.c and lex.yy.c are not supported.
`BINFORMAT=foo make checkdpadd' in /usr/src now reports only 2 false
negatives (in libss and init). (BINFORMAT=foo is to turn off better
handling of the a.out case.)
- the two `_EXTRADEPEND::' targets potentially clobbered each other for
`make -jN'. In practice, the output for the second target sometimes
disappeared.
- bogus dependencies were generated for static libraries.
headers in ${SRCS}, as in bsd.lmod.mk and bsd.prog.mk. This helps
`make [-j]' work when .depend doesn't exist. Even plain `make'
sometimes only worked because of magic ordering in ${SRCS}.
use the default without losing any (currently unused) features.
(CLEANDIRS is only used by for libgmp and libmp via bsd.lib.mk, and
only documented everywhere it is supported except of course where it
is actually used.)
I don't know what the hell I was thinking in: rev. 1.268
(2) Create ${PREFIX} before calling mtree if it doesn't exist. This
may not be the best solution, but pre-install is called after
mtree so there really isn't any way to fix this from the port
Makefiles and thus has to be done here.
Reviewed by: the lists
(2) Change MASTER_SITE_CTAN to reflect current reality.
Submitted by: fenner
(3) Add new port variable NO_LATEST_LINK. When this is set, the
"Latest" package symlink is not created. Use this for ports that
are betas when there is also a port for an older, more stable,
version.
(4) Don't be too stupid about "make deinstall".
Submitted by: fenner
in kernel Makefiles. Nothing in /usr/include is used (provided
relative paths for sys/* and <machine> can be found), so there is
no need for the -I/usr/include kludge as in kernel Makefiles.
them in the include path. This fixes recent breakage of the syscons
LKMs and general brokenness of the include paths (headers under
/usr/include were used in many cases).
it is built with this defined (which it isn't by default). This change
to sys.mk treats the absence of MACHINE_ARCH as i386 on the assumption
that it will be appropriately defined (as something else) on any other
architecture. When building FreeBSD's make with NetBSD tools, both
MACHINE and MACHINE_ARCH are correctly set (e.g. when bootstrapping
FreeBSD's make on NetBSD/mvme68k, MACHINE=mvme68k and
MACHINE_ARCH=m68k). This isn't really needed for the alpha which
has both defined as 'alpha', but I thought it was worth getting the
distinction between a MACHINE and a MACHINE_ARCH correct now.
Now, shouldn't PC98 have MACHINE=pc98 and MACHINE_ARCH=i386 ??!!
(1) Allow multiple checksums of same file.
Submitted by: hoek
(2) Add "deinstall" target as an alias to "pkg_delete $(make package-name)"
(well, something like that, see diff for details).
(3) Add new port variable USE_AUTOCONF. It appends BUILD_DEPENDS to
devel/autoconf and runs autoconf before configure.
Submitted by: ohashi@mickey.ai.kyutech.ac.jp (Takeshi Ohashi)
(4) Clarify USE_X11 and USE_IMAKE usage.
(5) Add new user-overridable variable MASTER_SITE_KDE.
Submitted by: vanilla
(6) Add support for "Latest" package links.
Idea by: Terry Lambert
(7) Try to catch obsolete tcl/tk installations that could cause problems.
Annoyed by various bogus commits by: you-know-who
don't remove a.out explicitly. a.out should only be generated for
libraries and is removed in the non-default rule in bsd.lib.mk.
Removed undocumented cleanfiles target. It was the same as the
default clean target except it didn't descend into subdirs. It was
different from special clean targets in other ways. This feature
hasn't been missed for more important targets.
Removed unused default cleandepend target. bsd.dep.mk has a better
version which is always used.
Use a better rule for checkdpadd in the BINFORMAT=aout case. This
mainly checks that ld -f is working correctly. The old rule is
still available via `make BINFORMAT=foo checkdpadd' and should be
used to check for regressions under 2.2 where ld -f is not available.
(as in bsd.prog.mk). Include it if `checkdpadd' is being made, so that
it can be checked until it goes away.
Don't clean files that we don't create.
Fixed style of empty test.
when there is no /usr/src/share/info if it were actually used.
Added comments to explain duplicated tex commands.
Use substitution in IFILENS to simplify some things.
Removed /g from many substitutions. It is bogus for anchored matches.
Don't echo nothing.
Don't add things that wouldn't be built with the current options to
CLEANFILES (except for some cases involving tex).
Reviewed by: wosch
we won't have double-slashes.
Add support for new port variable MANUAL_PACKAGE_BUILD. If this is
defined as well as the user variable PACKAGE_BUILDING, the port will
be ignored. This is used to mark ports that can be built normally
except on a machine that has a lot of conflicting ports (i.e., our
package building machine).
DESCR files when building README.html .
Don't use control characters in sed statement.
Problems reported by "Chris G. Demetriou" <cgd@pa.dec.com> in NetBSD PR
pkg/4341.
variables were lost when we removed -W, and 23 new ones including at
least one serious one have crept in for LINT.
Restored -Winline to CFLAGS. This gives only 3 old warnings and 1 new
for LINT.
Run tex twice due cross references.
Cleanup many tex generated files in `make clean'
Format latin1 output in 80 characters column (was ~110 characters).
.if in Makefiles. bsd.prog.mk and bsd.lib.mk do not depend on it however.
Allow overriding of the -soname arg when building the lib*crypt.so* libs
since libdescrypt.so and libscrupt.so both need a -soname of libcrypt.so
so that the symlink is obeyed at runtime rather than at compile time.
since 2.1.x make(1) apparently does not have the -m switch to set both
the the bsd.*.mk and sys.mk location, and this breaks 'make world' from a
2.1.x system.
note, using "-Wl,-f" to generate a library objects list doesn't work
anymore since the hack to ld hasn't been incorporated into binutils-2.8.
(and the -f switch is used for something else already)
This is disabled by default, don't panic! :-)
you can get one from.)
(2) Use "reinstall" as DEPENDS_TARGET if target is "reinstall". In
particular, this will make it possible to do a "make reinstall" on
several NFS clients and have them install all dependencies
correctly.
Tested by: davidn
a couple *.mk files to enable -current world building on really old
machines (e.g., 2.1.5).
Reviewed by: too many many people to list here, special thanks to bde
in a few places (in bsd.lib.mk and bsd.prog.mk); this merely fixed part of
the brokenness by not setting it here.
This fixes building of secure telnetd when DESTDIR is defined.
(Otherwise, it will try to link libtelnet from ${DESTDIR}/usr/lib.)
Reviewed by: bde, jkh
