* Re: Bug#101523: zsh segmentation fault [not found] <20010619225633.A7077@Amber.lab.icm.edu.pl> @ 2001-06-19 22:33 ` Clint Adams 2001-06-20 4:01 ` Bart Schaefer 2001-06-20 4:57 ` Bart Schaefer 0 siblings, 2 replies; 10+ messages in thread From: Clint Adams @ 2001-06-19 22:33 UTC (permalink / raw) To: zsh-workers; +Cc: Michal Politowski, 101523-forwarded [-- Attachment #1: Type: text/plain, Size: 323 bytes --] On Tue, Jun 19, 2001 at 10:56:33PM +0200, Michal Politowski wrote: > zsh segfaults when trying to source /usr/share/shoop/shoop.sh from the shoop package. > Unfortunately this file is rather complicated and I can't provide a smaller > example. I can reproduce this with zsh --no-equals -fx shoop.sh shoop.sh is attached [-- Attachment #2: shoop.sh --] [-- Type: text/x-sh, Size: 3320 bytes --] #!/bin/sh -e # This script has been preprocessed prior to installation # It has had comments, blank lines, and leading spaces # removed, and \-style lines combined. This was done so # that it could run quicker under some shells. _shoop () { local TRUEOBJ=$1 TRYOBJ=$2 METH=$3 TRUEMETH=$1_$3 TRYMETH=$2_$3 LASTMETH=$METH shift 3 case "$1" in =|=q|=p|.=|.=q|.=p|.=qp|:|:p) local varmeth=$1 append="" quiet="" private=""; shift if [ "${varmeth%p}" != $varmeth ]; then private=1; varmeth=${varmeth%p}; fi if [ "$_shoop_introspect" ] && eval [ -z \"\$_shooptype_$TRYMETH\$private\" ]; then eval "_shoopdefines_$TRUEOBJ=\"\$_shoopdefines_$TRUEOBJ $METH\"" fi if [ -z "$_shoopnocache_" ]; then eval $_shoopcacheclear_ fi if [ "${varmeth#.}" != $varmeth ]; then append=1 varmeth=${varmeth#.}; fi if [ "${varmeth%q}" != $varmeth ]; then quiet=1 varmeth=${varmeth%q}; fi if [ "$varmeth" = = ]; then if [ "$append" ];then set -- "$(eval eval "\$_shoop_$TRUEMETH") $@"; fi if [ ! "$quiet" ]; then echo -n $@; fi eval "_shoop_$TRUEMETH='echo -n $@' _shooptype_$TRUEMETH=variable" else if [ "$quiet" ]; then echo "Invalid modifier(q) on assignment!($TRUEOBJ.$METH)" >&2; fi if [ "$append" ];then eval eval "_shoop_$TRUEMETH=\'\$_shoop_$TRUEMETH;\$@\' _shooptype_$TRUEMETH=method" else eval "_shoop_$TRUEMETH='$@' _shooptype_$TRUEMETH=method" fi fi return ;;esac if eval [ \"\$_shooptype_$TRYMETH\" ]; then local THIS=$TRUEOBJ eval eval "\$_shoop_$TRYMETH" return else eval local P PARENTS=\"$(eval eval "\$_shoop_${TRYOBJ}_parent")\" THIS=$TRUEOBJ GETMETH="" NEWPARENTS="" if [ -z "$_shoopnocache_" ]; then eval local CACHE=\"\$_shoopcache_link_$TRUEMETH\" if [ "$CACHE" ]; then eval eval \$$CACHE return fi fi for P in $PARENTS; do eval GETMETH=\"\$_shoop_${P}_$METH\" if [ "$GETMETH" ]; then eval "$GETMETH" return fi NEWPARENTS="$NEWPARENTS $(eval eval "\$_shoop_${P}_parent")" done local orgargs="$@" set -- $NEWPARENTS while [ $# -gt 0 ];do P=$1 eval GETMETH="\$_shoop_${P}_$METH" if [ "$GETMETH" ]; then set -- $orgargs if [ -z "$_shoopnocache_" ]; then eval _shoopcache_link_${THIS}_$METH=_shoop_${P}_$METH _shoopcache_=\"\$_shoopcache_ _shoopcache_method_$METH _shoopcache_link_${THIS}_$METH \" _shoopcache_method_$METH=\"\$_shoopcache_method_$METH _shoopcache_link_${THIS}_$METH\" _shoopcache_linkmethod_${P}_$METH=\"\$_shoopcache_linkmethod_${P}_$METH _shoopcache_link_${THIS}_$METH\" fi eval "$GETMETH" return fi shift set -- $(eval eval "\$_shoop_${P}_parent") "$@" done echo "\"$METH\" is undefined for $TRYOBJ." >&2 return 1 fi } IFS=" " _shoopcacheclear_=" if eval [ \\\"\\\$_shoopcache_method_\$METH\\\" ]; then if eval [ -z \\\"\\\$_shoopcache_linkmethod_\$TRUEMETH\\\" ]; then eval unset _shoopcache_method_\$METH \\\$_shoopcache_method_\$METH _shoopcache_linkmethod_\$TRUEMETH \\\$_shoopcache_linkmethod_\$TRUEMETH fi fi " _shoop_introspect=1 IFS=" " _shoop OBJECT OBJECT new :p ' local OBJNAME=$1 eval "$OBJNAME () { shift; _shoop $OBJNAME $OBJNAME \"\$@\"; };" if [ $THIS != $OBJNAME ]; then _shoop $OBJNAME $OBJNAME parent = $THIS >/dev/null fi eval unset _shoopcache_ \$_shoopcache_ || true ' _shoop OBJECT OBJECT new OBJECT OBJECT . parent = "" OBJECT . super :p '_shoop $THIS $($THIS . parent) "$LASTMETH" "$@"; return' unset _shoop_introspect ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: Bug#101523: zsh segmentation fault 2001-06-19 22:33 ` Bug#101523: zsh segmentation fault Clint Adams @ 2001-06-20 4:01 ` Bart Schaefer 2001-06-20 6:51 ` Andrej Borsenkow 2001-06-20 10:01 ` Peter Stephenson 2001-06-20 4:57 ` Bart Schaefer 1 sibling, 2 replies; 10+ messages in thread From: Bart Schaefer @ 2001-06-20 4:01 UTC (permalink / raw) To: Clint Adams, zsh-workers; +Cc: Michal Politowski, 101523-forwarded On Jun 19, 6:33pm, Clint Adams wrote: } } I can reproduce this with zsh --no-equals -fx shoop.sh It's happening when the `=' parameter is (not) restored upon exit from the function. I'm not sure that's really the root cause, though. Trace below. I will point out that there's a rather nasty bug in shoop.sh on line 61: local orgargs="$@" This will almost certainly not do what is wanted; it'll set orgargs to be $1 and will make local variables whose names are the rest of the arguments. Then later, set -- $orgargs will also not produce the desired effect. ------------ BUG: in restoring special parameters zsh: segmentation fault (core dumped) zagzig<11> gdb Src/zsh core GNU gdb 4.17.0.4 with Linux/x86 hardware watchpoint and FPU support Copyright 1998 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. Type "show copying" to see the conditions. There is absolutely no warranty for GDB. Type "show warranty" for details. This GDB was configured as "i386-redhat-linux"... Core was generated by `Src/zsh -o noequals -fx /home/schaefer/Mail/detach.dir/shoop.sh'. Program terminated with signal 11, Segmentation fault. Reading symbols from /lib/libnsl.so.1...done. Reading symbols from /usr/lib/libncurses.so.4...done. Reading symbols from /lib/libm.so.6...done. Reading symbols from /lib/libc.so.6...done. Reading symbols from /lib/ld-linux.so.2...done. Reading symbols from /lib/libnss_files.so.1...done. #0 0x805f3c2 in restore_params (restorelist=0x40112580, removelist=0x40112588) at ../../zsh-4.0/Src/exec.c:2439 2439 tpm->flags = pm->flags; (gdb) where #0 0x805f3c2 in restore_params (restorelist=0x40112580, removelist=0x40112588) at ../../zsh-4.0/Src/exec.c:2439 #1 0x805eecc in execcmd (state=0xbffffaa8, input=0, output=0, how=18, last1=2) at ../../zsh-4.0/Src/exec.c:2315 #2 0x805bacb in execpline2 (state=0xbffffaa8, pcode=5699, how=18, input=0, output=0, last1=0) at ../../zsh-4.0/Src/exec.c:1192 #3 0x805b043 in execpline (state=0xbffffaa8, slcode=11266, how=18, last1=0) at ../../zsh-4.0/Src/exec.c:982 #4 0x805a9dd in execlist (state=0xbffffaa8, dont_change_job=0, exiting=0) at ../../zsh-4.0/Src/exec.c:826 #5 0x805a6f7 in execode (p=0x401122c8, dont_change_job=0, exiting=0) at ../../zsh-4.0/Src/exec.c:729 #6 0x806fcd5 in loop (toplevel=1, justonce=0) at ../../zsh-4.0/Src/init.c:160 #7 0x80724b4 in zsh_main (argc=5, argv=0xbffffb38) at ../../zsh-4.0/Src/init.c:1209 #8 0x804a8b4 in main (argc=5, argv=0xbffffb38) at ../../zsh-4.0/Src/main.c:37 (gdb) p pm $1 = (struct param *) 0x40112590 (gdb) p *pm $2 = {next = 0x40112568, nam = 0x8124bbc "=", flags = 786432, u = { data = 0x8137640, arr = 0x8137640, str = 0x8137640 " ", val = 7309465757271946816, dval = 4.2633628668410461e+180, hash = 0x8137640}, sets = {cfn = 0x998b646e, ifn = 0x998b646e, ffn = 0x998b646e, afn = 0x998b646e, hfn = 0x998b646e}, gets = {cfn = 0x99, ifn = 0x99, ffn = 0x99, afn = 0x99, hfn = 0x99}, unsetfn = 0, ct = 0, env = 0x0, ename = 0x0, old = 0x0, level = 1074866392} (gdb) p tpm $3 = (struct param *) 0x0 (gdb) l 2434 Param tpm = (Param) paramtab->getnode(paramtab, pm->nam); 2435 2436 DPUTS(!tpm || PM_TYPE(pm->flags) != PM_TYPE(tpm->flags) || 2437 !(pm->flags & PM_SPECIAL), 2438 "BUG: in restoring special parameters"); 2439 tpm->flags = pm->flags; 2440 switch (PM_TYPE(pm->flags)) { 2441 case PM_SCALAR: 2442 tpm->sets.cfn(tpm, pm->u.str); 2443 break; (gdb) p pm->nam $4 = 0x8124bbc "=" -- Bart Schaefer Brass Lantern Enterprises http://www.well.com/user/barts http://www.brasslantern.com Zsh: http://www.zsh.org | PHPerl Project: http://phperl.sourceforge.net ^ permalink raw reply [flat|nested] 10+ messages in thread
* RE: Bug#101523: zsh segmentation fault 2001-06-20 4:01 ` Bart Schaefer @ 2001-06-20 6:51 ` Andrej Borsenkow 2001-06-20 7:59 ` Andrej Borsenkow 2001-06-20 16:04 ` Bug#101523: zsh segmentation fault Bart Schaefer 2001-06-20 10:01 ` Peter Stephenson 1 sibling, 2 replies; 10+ messages in thread From: Andrej Borsenkow @ 2001-06-20 6:51 UTC (permalink / raw) To: Bart Schaefer, Clint Adams, zsh-workers Cc: Michal Politowski, 101523-forwarded > > I will point out that there's a rather nasty bug in shoop.sh on line 61: > > local orgargs="$@" > > This will almost certainly not do what is wanted; it'll set orgargs to be > $1 and will make local variables whose names are the rest of the > arguments. > Then later, > > set -- $orgargs > > will also not produce the desired effect. > This is related to export foo=$(cmd) problem we have discussed. Obviously, other shells treat {export|local|typeset} foo=bar w.r.t. word splitting exactly the same as foo=bar, i.e. they do not split in assignments here. Did not we have some patch for it, at least in HEAD? Looks, like it should be in 4.0. Have ^ permalink raw reply [flat|nested] 10+ messages in thread
* RE: Bug#101523: zsh segmentation fault 2001-06-20 6:51 ` Andrej Borsenkow @ 2001-06-20 7:59 ` Andrej Borsenkow 2001-06-20 16:47 ` KSH_TYPESET and 4.0.2 Bart Schaefer 2001-06-20 16:04 ` Bug#101523: zsh segmentation fault Bart Schaefer 1 sibling, 1 reply; 10+ messages in thread From: Andrej Borsenkow @ 2001-06-20 7:59 UTC (permalink / raw) To: zsh-workers > > > > I will point out that there's a rather nasty bug in shoop.sh on line 61: > > > > local orgargs="$@" > > > > This will almost certainly not do what is wanted; it'll set > orgargs to be > > $1 and will make local variables whose names are the rest of the > > arguments. > > Then later, > > > > set -- $orgargs > > > > will also not produce the desired effect. > > > > This is related to export foo=$(cmd) problem we have discussed. Obviously, > other shells treat {export|local|typeset} foo=bar w.r.t. word splitting > exactly the same as foo=bar, i.e. they do not split in assignments here. > > Did not we have some patch for it, at least in HEAD? Looks, like it should > be in 4.0. 14858 -andrej ^ permalink raw reply [flat|nested] 10+ messages in thread
* KSH_TYPESET and 4.0.2 2001-06-20 7:59 ` Andrej Borsenkow @ 2001-06-20 16:47 ` Bart Schaefer 2001-06-20 18:33 ` Peter Stephenson 0 siblings, 1 reply; 10+ messages in thread From: Bart Schaefer @ 2001-06-20 16:47 UTC (permalink / raw) To: zsh-workers On Jun 20, 11:59am, Andrej Borsenkow wrote: } Subject: RE: Bug#101523: zsh segmentation fault } } > > local orgargs="$@" } > } > This is related to export foo=$(cmd) problem we have discussed. } > } > Did not we have some patch for it, at least in HEAD? Looks, like it should } > be in 4.0. } } 14858 Can we get a few more votes on whether KSH_TYPESET should go in 4.0.2? I note that we still haven't resolved the bash-compatibility issue. Other than that, though, I see no reason not to include it. -- Bart Schaefer Brass Lantern Enterprises http://www.well.com/user/barts http://www.brasslantern.com Zsh: http://www.zsh.org | PHPerl Project: http://phperl.sourceforge.net ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: KSH_TYPESET and 4.0.2 2001-06-20 16:47 ` KSH_TYPESET and 4.0.2 Bart Schaefer @ 2001-06-20 18:33 ` Peter Stephenson 0 siblings, 0 replies; 10+ messages in thread From: Peter Stephenson @ 2001-06-20 18:33 UTC (permalink / raw) To: Zsh hackers list "Bart Schaefer" wrote: > Can we get a few more votes on whether KSH_TYPESET should go in 4.0.2? > > I note that we still haven't resolved the bash-compatibility issue. Other > than that, though, I see no reason not to include it. It's not exactly a bug fix, but after fixing the setting of individual positional params it doesn't seem to be problematic. If it looks like people are running across this difference I don't mind it going in. -- Peter Stephenson <pws@csr.com> Software Engineer CSR Ltd., Unit 300, Science Park, Milton Road, Cambridge, CB4 0XL, UK Tel: +44 (0)1223 392070 ********************************************************************** The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. ********************************************************************** ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: Bug#101523: zsh segmentation fault 2001-06-20 6:51 ` Andrej Borsenkow 2001-06-20 7:59 ` Andrej Borsenkow @ 2001-06-20 16:04 ` Bart Schaefer 1 sibling, 0 replies; 10+ messages in thread From: Bart Schaefer @ 2001-06-20 16:04 UTC (permalink / raw) To: Andrej Borsenkow, Clint Adams, zsh-workers Cc: Michal Politowski, 101523-forwarded On Jun 20, 10:51am, Andrej Borsenkow wrote: } Subject: RE: Bug#101523: zsh segmentation fault } } > local orgargs="$@" } > } > This will almost certainly not do what is wanted; it'll set orgargs to be } > $1 and will make local variables whose names are the rest of the } > arguments. } } This is related to export foo=$(cmd) problem we have discussed. Obviously, } other shells treat {export|local|typeset} foo=bar w.r.t. word splitting } exactly the same as foo=bar, i.e. they do not split in assignments here. I don't doubt it, but as you pointed out yourself the POSIX-defined behavior is to split the arguments of those commands. If the authors of shoop, whoever they are, want it to be portable, they should be avoiding this conflict. It's easy enough to use local orgargs="$*" instead. -- Bart Schaefer Brass Lantern Enterprises http://www.well.com/user/barts http://www.brasslantern.com Zsh: http://www.zsh.org | PHPerl Project: http://phperl.sourceforge.net ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: Bug#101523: zsh segmentation fault 2001-06-20 4:01 ` Bart Schaefer 2001-06-20 6:51 ` Andrej Borsenkow @ 2001-06-20 10:01 ` Peter Stephenson 2001-06-20 10:10 ` Peter Stephenson 1 sibling, 1 reply; 10+ messages in thread From: Peter Stephenson @ 2001-06-20 10:01 UTC (permalink / raw) To: Zsh hackers list "Bart Schaefer" wrote: > It's happening when the `=' parameter is (not) restored upon exit from the > function. I'm not sure that's really the root cause, though. No, I think it's when it (or something completely different --- there's no `=' special parameter) is being saved. The following looks a safe fix. Unless anyone knows any better, this will go on both branches. (Saved parameters go on the heap since the list they're saved on is removed when the scope exits.) Index: Src/exec.c =================================================================== RCS file: /cvsroot/zsh/zsh/Src/exec.c,v retrieving revision 1.28 diff -u -r1.28 exec.c --- Src/exec.c 2001/05/28 09:00:22 1.28 +++ Src/exec.c 2001/06/20 09:58:32 @@ -2395,7 +2395,7 @@ } else if (!(pm->flags & PM_READONLY) && (unset(RESTRICTED) || !(pm->flags & PM_RESTRICTED))) { Param tpm = (Param) zhalloc(sizeof *tpm); - tpm->nam = s; + tpm->nam = dupstring(s); copyparam(tpm, pm, 1); pm = tpm; } -- Peter Stephenson <pws@csr.com> Software Engineer CSR Ltd., Unit 300, Science Park, Milton Road, Cambridge, CB4 0XL, UK Tel: +44 (0)1223 392070 ********************************************************************** The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. ********************************************************************** ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: Bug#101523: zsh segmentation fault 2001-06-20 10:01 ` Peter Stephenson @ 2001-06-20 10:10 ` Peter Stephenson 0 siblings, 0 replies; 10+ messages in thread From: Peter Stephenson @ 2001-06-20 10:10 UTC (permalink / raw) To: Zsh hackers list Peter Stephenson wrote: > - tpm->nam = s; > + tpm->nam = dupstring(s); > copyparam(tpm, pm, 1); That could simply have been pm->nam on the right hand side, saving the string duplication, but I don't have the energy to change it, test it, and check it in again. If anyone else wants to... -- Peter Stephenson <pws@csr.com> Software Engineer CSR Ltd., Unit 300, Science Park, Milton Road, Cambridge, CB4 0XL, UK Tel: +44 (0)1223 392070 ********************************************************************** The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. ********************************************************************** ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: Bug#101523: zsh segmentation fault 2001-06-19 22:33 ` Bug#101523: zsh segmentation fault Clint Adams 2001-06-20 4:01 ` Bart Schaefer @ 2001-06-20 4:57 ` Bart Schaefer 1 sibling, 0 replies; 10+ messages in thread From: Bart Schaefer @ 2001-06-20 4:57 UTC (permalink / raw) To: Clint Adams, zsh-workers; +Cc: Michal Politowski, 101523-forwarded On Jun 19, 6:33pm, Clint Adams wrote: } } On Tue, Jun 19, 2001 at 10:56:33PM +0200, Michal Politowski wrote: } } > zsh segfaults when trying to source /usr/share/shoop/shoop.sh Incidentally, it executes without crashing if I `emulate sh' first. -- Bart Schaefer Brass Lantern Enterprises http://www.well.com/user/barts http://www.brasslantern.com Zsh: http://www.zsh.org | PHPerl Project: http://phperl.sourceforge.net ^ permalink raw reply [flat|nested] 10+ messages in thread
end of thread, other threads:[~2001-06-20 18:34 UTC | newest] Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- [not found] <20010619225633.A7077@Amber.lab.icm.edu.pl> 2001-06-19 22:33 ` Bug#101523: zsh segmentation fault Clint Adams 2001-06-20 4:01 ` Bart Schaefer 2001-06-20 6:51 ` Andrej Borsenkow 2001-06-20 7:59 ` Andrej Borsenkow 2001-06-20 16:47 ` KSH_TYPESET and 4.0.2 Bart Schaefer 2001-06-20 18:33 ` Peter Stephenson 2001-06-20 16:04 ` Bug#101523: zsh segmentation fault Bart Schaefer 2001-06-20 10:01 ` Peter Stephenson 2001-06-20 10:10 ` Peter Stephenson 2001-06-20 4:57 ` Bart Schaefer
Code repositories for project(s) associated with this public inbox https://git.vuxu.org/mirror/zsh/ This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).