Summary: | dev-lang/tcl-8.5.7 - application-specific initialization failed: out of stack space (infinite loop?) | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Jeroen Roovers (RETIRED) <jer> |
Component: | Current packages | Assignee: | TCL/TK Project <tcltk> |
Status: | RESOLVED OBSOLETE | ||
Severity: | normal | CC: | gnu_andrew, halcy0n, hppa |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | HPPA | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | 406607, 410511, 410515, 445206 | ||
Bug Blocks: | 280845 | ||
Attachments: | dev-lang:tcl-8.5.7:20090810-000018.log [hppa,fail] |
Description
Jeroen Roovers (RETIRED)
2009-08-10 00:20:40 UTC
Created attachment 200761 [details]
dev-lang:tcl-8.5.7:20090810-000018.log [hppa,fail]
Looks a lot like bug #279528, but perhaps with the stack limit being reached before a segfault occurs. Standard stack limit on Gentoo/HPPA: elmer ~ # ulimit -s 8192 Fails the same way with USE=-threads. After a `ulimit -s unlimited' in the same shell and running emerge again, src_install() succeeds. may be related to: http://sourceforge.net/tracker/?func=detail&atid=110894&aid=1889719&group_id=10894 so you should be able to compile it by adding -DTCL_NO_STACK_CHECK=1 to your CFLAGS. do you confirm? (In reply to comment #5) > so you should be able to compile it by adding -DTCL_NO_STACK_CHECK=1 to your > CFLAGS. do you confirm? Yes, that does seem to work around the problem. Any ideas how to put this in the ebuild? Just a `use hppa && append-cflags -DTCL_NO_STACK_CHECK=1' maybe? > Yes, that does seem to work around the problem. Any ideas how to put this in
> the ebuild? Just a `use hppa && append-cflags -DTCL_NO_STACK_CHECK=1' maybe?
yes I was thinking something like that.
updated tcl-8.5.7.ebuild in CVS ;-) I got the exact same issue while trying to emerge dev-tcltk/expect on ppc. Is it really hardware specific ? Why on Earth did you remove hppa from CC? (In reply to comment #9) > I got the exact same issue while trying to emerge dev-tcltk/expect on ppc. > Is it really hardware specific ? > Same here, I would recommend "append-cflags "-DTCL_NO_STACK_CHECK=1"" for ppc as well. *** Bug 306853 has been marked as a duplicate of this bug. *** (In reply to comment #10) > Why on Earth did you remove hppa from CC? > Sorry, it was not intentional... Anyone feel like committing this? Index: tcl-8.5.7.ebuild =================================================================== RCS file: /var/cvsroot/gentoo-x86/dev-lang/tcl/tcl-8.5.7.ebuild,v retrieving revision 1.13 diff -u -B -r1.13 tcl-8.5.7.ebuild --- tcl-8.5.7.ebuild 15 Jun 2010 23:46:51 -0000 1.13 +++ tcl-8.5.7.ebuild 10 Aug 2010 02:42:11 -0000 @@ -45,7 +45,7 @@ src_compile() { # workaround stack check issues, bug #280934 - if use hppa; then + if use hppa || use ppc; then append-cflags "-DTCL_NO_STACK_CHECK=1" fi Index: tcl-8.5.8-r1.ebuild =================================================================== RCS file: /var/cvsroot/gentoo-x86/dev-lang/tcl/tcl-8.5.8-r1.ebuild,v retrieving revision 1.4 diff -u -B -r1.4 tcl-8.5.8-r1.ebuild --- tcl-8.5.8-r1.ebuild 16 Jun 2010 07:40:18 -0000 1.4 +++ tcl-8.5.8-r1.ebuild 10 Aug 2010 02:42:11 -0000 @@ -42,7 +42,7 @@ src_configure() { # workaround stack check issues, bug #280934 - if use hppa; then + if use hppa || use ppc; then append-cflags "-DTCL_NO_STACK_CHECK=1" fi Index: tcl-8.5.8.ebuild =================================================================== RCS file: /var/cvsroot/gentoo-x86/dev-lang/tcl/tcl-8.5.8.ebuild,v retrieving revision 1.3 diff -u -B -r1.3 tcl-8.5.8.ebuild --- tcl-8.5.8.ebuild 15 Jun 2010 23:46:51 -0000 1.3 +++ tcl-8.5.8.ebuild 10 Aug 2010 02:42:11 -0000 @@ -45,7 +45,7 @@ src_compile() { # workaround stack check issues, bug #280934 - if use hppa; then + if use hppa || use ppc; then append-cflags "-DTCL_NO_STACK_CHECK=1" fi Go ahead. (In reply to comment #15) > Go ahead. It won't do. After reviewing this bug and bug #306853, it appears this isn't affecting ppc (I can't reproduce it like volkmar can), but bug #306853 was filed for a ppc64 system, and someone with an x86 had it as well. We'd have to go for a generic append-flags but I have no idea how that would affect our distribution/what upstreams thinks of this. *** Bug 352187 has been marked as a duplicate of this bug. *** This was discussed here: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=452679 Looks like there is an upstream solution already. Anyone care if I apply the patch from upstream? (In reply to comment #19) > Anyone care if I apply the patch from upstream? Please go ahead. I should have checked this first... tcl-8.5.10 already has this patch applied. Its been in the tree for awhile now, could we just go ahead with that version? 8.5.7 is quite all, so going for .10 makes sense. (In reply to comment #22) > 8.5.7 is quite all, so going for .10 makes sense. * /all/old ;) ppc done We need to wait for that other bug to remove the broken versions. + 10 Feb 2013; Justin Lecher <jlec@gentoo.org> -tcl-8.4.19.ebuild, + -tcl-8.5.11-r1.ebuild, -tcl-8.5.13.ebuild: + Drop old fixes #280934 + |