Due to the use of signed vs. unsigned chars on our various platforms, one gets
"warning: comparison is always true due to limited range of data type"
from GCC 3.3.
mutually exclusive. The fact that the most recent one specified on the
command line is the one that takes effect is an implementation detail and
users should not rely on this.
The initial stack_block is staticly allocated and will be aligned
according to the alignment requirements of pointers, which does not
necessarily match the alignment enforced by ALIGN. To solve this a
more involved change is required: remove the static initial stack
and deal with an initial condition of not having a stack at all. This
change is therefore more risky than the previous ones, but unavoidable
(other than not using the platform default alignment).
Discussed with: tjr
Approved and reviewed by: tjr
Tested on: alpha, i386, ia64 and sparc64
The problem with the previous attempt, as noticed by Marcel, was that
stacknxt was being aligned to a pointer boundary instead of an
ALIGNBYTES + 1 boundary, which broke sparc64.
using the alignment from sys/param.h (16) instead of the alignment
from machdep.h (8) tickled a nasty bug in the memory allocator that I
haven't been able to track down yet.
are later stripped with rmescapes() in expandarg(). If the filename has
already been unescaped, doing it again in rmescapes() can walk off the
end of the string, leading to memory corruption and eventually SIGSEGV.
Noticed by: kris
commands. Commands like "if then ... fi" and "while do ... done" are no
longer accepted. Bodies of compound commands are still allowed to be
empty, because even though POSIX does not allow them, most shells do.
pointers. This fixes two format warnings on 64 bits
archs which are fatal now that WFORMAT=0 has been removed.
It doesn't fully fix the sh(1) build on 64 bits platforms
though, there is still some quad_t issues that need to be
fixed.
Tested on: i386, sparc64
adding history and vi/emacs-style line editing to the shell itself.
Atty was a user-mode terminal emulator (like screen and window) that did
line editing and history.
forkshell() after it has been freed. This caused mysterious behaviour
when anything but the first command in a pipeline tried to access the
terminal when the `junk' malloc() option was enabled (which is the default).
itself does that if you set EL_SIGNAL. Instead, set a flag and check it
before calling el_gets(). This is safer, but slower to respond to changes.
Pointed out by: mp
shell function and alias lookup. The -p option has been implemented, the
UPE -v and -V options have not. The old `command' command has been renamed
to `builtin'.
The pgrp member of struct job was declared as a short and could not store
every possible process group ID value, the rest of them were benign because
pid_t happens to be an int.
2. Instead, open /dev/tty. This problem stopped commands in subshells from
being executed correctly if standard error was redirected.
PR: 36671
Obtained from: NetBSD (but simplified)
reality (and POSIX): current directory isn't searched unless CDPATH has
a "." element or is unset.
PR: 38442
Submitted by: oleg dashevskii <be9@be9.ru>
MFC after: 1 week
keep a linked list of the jobs, most recently used first. This is required
to support the idea of `previous job', and to allow the jobs fg and bg
default to be correct according to POSIX.
to fail when the logical current directory no longer exists. Allow changes
to absolute paths when logical cwd is invalid, fall back to physical cd
if logical cd fails.
o Old-style K&R declarations have been converted to new C89 style
o register has been removed
o prototype for main() has been removed (gcc3 makes it an error)
o int main(int argc, char *argv[]) is the preferred main definition.
o Attempt to not break style(9) conformance for declarations more than
they already are.
o Change
int
foo() {
...
to
int
foo(void)
{
...
used so often that it's worth keeping it as a builtin.
Now that all the printf invocations from within the system startup
scripts, we can safely remove it.
Urged by: sheldonh :)
No MFC is planned so far because it may break compatibility and
violate POLA.
binary size increase is 3,784 bytes (about 0.6%).
I don't drop the printf builtin while I'm here because some /etc/rc.*
scripts seem to use it before mounting /usr where printf(1) resides.
Reviewed by: arch (sheldonh)
Inspired by: NetBSD, ksh
Clued by: ume (on how the printf builtin is used)
setvar() and passed to setvareq(). When the VTEXTFIXED flag is set,
that copy is never freed, causing a memory leak.
PR: 31533
Submitted by: maxim@macomnet.ru
value CTLARI since this might break expansion of arithmetic expressions.
Don't access memory below start of stackblock.
Problem analyzed by hunt@iprg.nokia.com, slightly different patch applied.
PR: 24443
Submitted by: hunt@iprg.nokia.com
Avoid using parenthesis enclosure macros (.Pq and .Po/.Pc) with plain text.
Not only this slows down the mdoc(7) processing significantly, but it also
has an undesired (in this case) effect of disabling hyphenation within the
entire enclosed block.
errexit (-e) processing. This solves a problem where 'make clean' would
fail with an unspecified error in certain automake-generated makefiles.
Reviewed by: no objections from -hackers...
MFC after: 2 weeks
When a child is receiving SIGSTOP, eval continues with the next
command. While that is correct for the interactive case (Control-Z
and you get the prompt back), it is wrong for a shellscript, which
just continues with the next command, never again waiting for the
stopped child. Noted when childs from cronjobs were stopped, just to
make more processes (by wosch).
The fix is not to return from a job wait when the wait returned for a
stopped child while in non-interactive mode. This bahaviour seems to
be what bash2 and ksh implement. I tested for correct behaviour for
finnaly killing the child with and without forgrounding it first.
When not foregrouding before killing, the shell continues with the
script, which is what the other shells do as well.
Reviewed by: Silence on -current
Serious fix still needed, see discussion on -current
(Subject: /bin/sh dumps core with here-document of 8bit text)
Problem in this code originally spotted by
Jun Kuriyama <kuriyama@FreeBSD.org>
growstackblock() sometimes relocates a stack_block considered empty
without properly relocating stack marks referencing that block.
The first call to popstackmark() with the unrelocated stack mark
as argument then causes sh to abort.
Relocating the relevant stack marks seems to solve this problem.
The patch changes the semantics of popstackmark() somewhat. It can
only be called once after a call to setstackmark(), thus cmdloop() in
main.c needs an extra call to setstackmark().
PR: bin/19983
Submitted by: Tor.Egge@fast.no
Reviewed by: Gerald Pfeifer <pfeifer@dbai.tuwien.ac.at>
case), so that it doesn't clash with the ncurses function of the same
name when linking statically with -ltermcap.
The linker only complains when -static is used, and it is not clear
whether this is a bug.
PR: bin/18104
Submitted by: Anatoly Vorobey <mellon@pobox.com>
spaces reserved by the header files it includes.
mkinit.c still produces C code with redundant declarations, although
they are more harmless since they automatically derived from the right
places.
in that revision as well as things I broke in that revision. A note-
worthy instance of the latter case was the inversion of -E and -V in the
subsection on Commandline Editing.
than two processes (got that? :-), the stdin fd of the middle
processes that has just been set up was accidetially closed. Don't do
this.
PR: bin/14527
commit and those which cause ugly nroff output have been fixed, since
the purpose of the style guideline which they contravene is to reduce
the sizes of deltas.
Reported by: bde
* Consistently misspell built-in as builtin.
* Add a builtin(1) manpage and create builtin(1) MLINKS for all shell
builtin commands for which no standalone utility exists. These MLINKS
replace those that were created for csh(1).
* Add appropriate xrefs for builtin(1) to the csh(1) and sh(1) manpages,
as well as to the manpages of standalone utilities which are supported
as shell builtin commands in at least one of the shells. In such
manpages, explain that similar functionality may be provided as a
shell builtin command.
* Improve sh(1)'s description of the cd builtin command. Csh(1) already
describes it adequately. Replace the cd(1) manpage with a builtin(1)
MLINKS link.
* Clean up some mdoc problems: use Xr instead of literal "foo(n)"; use
Ic instead of Xr for shell builtin commands.
* Undo English contractions.
Reviewed by: mpp, rgrimes
Fix grammar and spelling nits.
Use .Dq and .Qq where appropriate.
Divorce trailing punctuation from quoted elements.
Use .Dq instead of .Xr for builtins.
Remove trailing whitespace and blank lines.
PR: 13340
statement if blocks[*] when the else could be ambiguous, not defaulting
to int type and removal of some unused variables.
[*] This is explicitly allowed by style(9) when the single statement
spans more than one line.
Reviewed by: obrien, chuckr
representation of the expression is quoted. Take care of this when
doing pattern matching in conjunction with trimming.
#!/bin/sh
c=d:e; echo "${c%:e}"
PR: NetBSD PR#7231
Noticed by: Havard Eidnes <Havard.Eidnes@runit.sintef.no>
make /etc/rc interruptible in cases when programs hang with blocked
signals) isn't standard enough.
It is now switched off by default and a new switch -T enables it.
You should update /etc/rc to the version I'm about to commit in a few
minutes to keep it interruptible.
Don't output double-quotes inside variable expansion/arithmetic
expansion region in here-documents. When leaving the arithmetic
expansion syntax mode, adjust the dblquote flag according to
previous syntax, in order to avoid splitting of quoted variables.
foreground child is running. Formerly, traps were exceuted after the
next child exit.
The enables the user to put a breaking wrapper around a blocking
application:
(trap 'echo trap ; exit 1' 2; ./pestyblocker; echo -n)
The "echo -n" after the child call is needed to prevent sh from
optimizing the trap-executing shell away. I'm working on this.
multiple times when performing nested variable expansion, and
preserve some quoting information in order to avoid removing
apparently empty expansion result.
i.e. this makes emacs usable from system(3). Programs called from
shellscripts are now required to exit with proper signal status. That
means, they have to kill themself. Exiting with faked numerical exit
code is not sufficient.
Exit with proper signal status if script exits on signal.
Make the wait builtin interruptable, both with and without traps set.
Use volatile sig_atomic_t where (and only where) appropriate.
(Almost) fix printing of newlines on SIGINT.
Make traps setable from trap handlers. This is needed for shellscripts
that catch SIGINT for cleanup work but intend to exit on it, hance
have to kill themself from a trap handler. I.e. mkdep.
While I'm at it, make it -Wall clean. -Wall is not enabled in
Makefile, since vararg warnx() macro calls in usr.bin/printf/printf.c
are not -Wall-able.
PR: 1206
Obtained from: Basic SIGINT fix from Bruce Evans
Removed explicit dependencies of foo.o on foo.c. These were mainly
placeholders for comments about missing dependencies of tools objects
on headers. This problem needs to be handled more generally.
urgent need is when you run sh around a program that intentionally
uses SIGQUIT/SIGINT for asynchronous events, i.e. $EDITOR started from
system(2), like many mailers do. This fixes PR bin/1206 and possibly
bin/4241.
The solution committed has been tested for a large number of possible
cases (see recent discussion on cvs-committers). I completed a make
world, made sure 'make world' is interruptable and used the changed
/bin/sh as a login shell all day, including job control and using
SIGQUIT-catching programs (to write this message :-).
PR: bin/1206
Reviewed by: discussion on cvs-commiters
cast value that was always ignored. Rev.1.9 of trap.c made this
more bogus by returning a semantically different value after calling
siginterrupt(). Avoid these problems by not returning a value.
trap 'echo xxx' 1 2 3 15
read x
is not interrupted by ^C (due to restartable read syscall) and must be
interrupted per POSIX
Worse case:
read -t 5 x
hangs forever after ^C pressed (supposed to timeout after 5 secs)
Fixed by adding siginterrupt(signo, 1) after catch handler installed
2) Do not reinstall sighandler immediately after it is called,
BSD do it for us
'read' command to return an error if the user fails to supply any
input withink a given time period. The behaviour of this option is
similar to that of the like-named option in ksh93.
Reviewed by: joerg
This will make a number of things easier in the future, as well as (finally!)
avoiding the Id-smashing problem which has plagued developers for so long.
Boy, I'm glad we're not using sup anymore. This update would have been
insane otherwise.
while remaining (becoming :) compatible with other popular shells.
Specifically these changes include:
1) Implement 'trap -l' to get a list of valid signals names. This
is useful if you wanted to do something like reset all signal
handlers to there defaults values, in which case something like
this will do the trick.
trap `trap -l`
2) Reformat the output of 'trap' so it can be saved and later eval'd
to restore the saved settings.
3) Allow the use of signal names as well as signal numbers.
4) Fix trap handling of SIGCHLD so that commands like the following
(albeit, contrived) won't cause sh(1) to recurse ad infinitum.
trap uname 0 20
5) Make variables static that are used only in trap.c.
6) Minor 'style(9) police' mods.
now handles the getpwd() init problem the same way as bash
and ksh do. Also while I was in here, I cleaned up the format
a little, removed some unnnecessary #if SYMLINKS cruft, and
changed the pwd builtin to use getcwd(3) as Joerg suggested.
getopts should now work as expected. This fix was in the NetBSD
code that I was merging from but missed getting into FreeBSD's
version because of 'drain bamage' on my part.
Submitted by: NetBSD, joerg
This patch causes too many side effects, one of which bites hard is
when interrupting a 'make fetch' in the ports tree (PR#1990).
This whole area is a real can of worms....
This most definately should go into 2.2
Reviewed by: steve, bde
so that simple regresssion tests based on `cmp' work. mkdep still
doesn't work right for these tools. They should probably be in
separate directories.
Sorted dependencies.
face of aliases. Note, bash doesn't do aliases while running scripts, but
"real" ksh does..
Also:
Reduce redundant .Nm macros in (unused) bltin/echo.1
nuke error2, it's hardly used.
More -Wall cleanups
dont do certain history operations if NO_HISTORY defined
handle quad_t's from resource limits
Submitted by: Steve Price <sprice@hiwaay.net> (minor tweaks by me)
- don't put \n on error() calls, error adds it already.
- don't prepend "ulimit" on error() calls in miscbltin.c.
- getopt typo on ulimit -p -> -u conversion
- get/setrlimit() calls were not being error checked
ulimit formatting cleanup from me, use same wording as bash on Bruce's
suggestion. Add ulimit arg to output on Joerg's suggestion.
merge of parallel duplicate work by Steve Price and myself. :-]
There are some changes to the build that are my fault... mkinit.c was
trying (poorly) to duplicate some of the work that make(1) is designed to
do. The Makefile hackery is my fault too, the depend list was incomplete
because of some explicit OBJS+= entries, so mkdep wasn't picking up their
source file #includes.
This closes a pile of /bin/sh PR's, but not all of them..
Submitted by: Steve Price <steve@bonsai.hiwaay.net>, peter
Requested by: joerg
(Note, this is mostly going to be conflicts, which is expected. Our entire
sh source has a mainline, so this should not change anything except for
a few new files appearing. I dont think they are a problem)
didn't correctly start background jobs anymore. Strange that nobody
was complaining...
Add a dummy target for `builtins' in the Makefile, to prevent it
from attempting to build this file by compiling builtins.c. :-/
traditional behaviour, and it violates Posix.2.
Fixes PR # bin/880: /bin/sh incorrectly parse...
Fixes also an earlier problem report about the shell not evaluating
loops correctly. (Not files via GNATS.)
Submitted by: nnd@itfs.nsk.su (Nickolay N. Dudorov)
This means that a script containing:
echo 1
set -v
echo 2
will now produce output, like it does on SYSV machines and other 'proper'
/bin/sh implementations..
This is done by a slight restructure of the input processor allowing it to
read chunks from the file at a time, but process the data by line from the
chunk.
Obtained from: Christos Zoulas for NetBSD. <christos@deshaw.com>
o fix brokeness for 1>&5 redirection, where `5' was an invalid file
descriptor, but no error message has been generated
o fix brokeness for redirect to/from myself case
command and badly needed in sh(1) for everybody who wants to modify
the system-wide limits from inside /etc/rc.
The options are similar to other system's implemantations of this
command, with the FreeBSD additions for -m (memoryuse) and -p (max
processes) that are not available on other systems.
sh -c [-aCefinuvx] command_string [ command_name [argument ...] ] 1
4.56.3 Options
-c Read commands from the command_string operand. Set the
value of special parameter 0 (see 3.5.2) from the value of
the command_name operand and the positional parameters
($1, $2, etc.) in sequence from the remaining argument
operands.
Pointed out by: Kaleb Keithly (kaleb@x.org)
When comparing my recent parser change against the ash in 1.1.5.1, i
found that a couple of other problems in the same area has been fixed
there, but not in 2.2. Semicolons and EOF do also delimit words...
The && and || tokens do also terminate a command, not only the
newline.
While i was at it, disabled trace code by default, it served no good
purpose since it required the use of a debugger anyway to be turned
on. Instead, placed a hint in the Makefile on how to turn it on.
This makes the shell ~ 10 % faster and ~ 4 KB smaller. :)
Pointed out by: jan@physik.TU-Berlin.DE (Jan Riedinger)
There is a bug in sh: the built in command "fc -l" generates
a core dump (*NULL in not_fcnumber).
According to the sh manual page (fc -l [-nr] [first [last]]), fc -l
is a correct sequence (in that case, values are defaulted to -16 and -1)
but fails when first is not given.
MKINIT line that doesn't have a comment on it (we have at least two).
This mkinit program was written by someone who obviously doesn't believe
in defensive programming. :-( There's a LOT of work that needs to be done
on this thing. :-( :-( :-(
find it in /bin. This is something of a kludge, I know, but consider
my limited alternatives: I can't make this an execvp() without making
people scream that I introduced a failure point or slowed down pwd,
and I can't make it an optional macro since crunch doesn't let you pass
arbitrary command-line args to the build of one of its crunch-ees.
This is the simplest, if not the nicest looking, solution I could come up
with.
This should fix it (passed my test cases). Originally discovered with
perl's Configure (well, in FreeBSD, I don't know how the NetBSD folks
discovered it).
Reviewed by: sef
Submitted by: jtc@cygnus.com
Obtained from: NetBSD
automagically. -lfoo has to be right to work, but ${LIBFO0} is too
easy to forget or misspell; nothing checks it and it should be
different for shared libraries.
