Summary: | [patch] fix dev-lang/perl ebuild for uclibc -fstack-protector causes segfault | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | dougztr <doug> |
Component: | [OLD] Development | Assignee: | Gentoo Perl team <perl> |
Status: | RESOLVED FIXED | ||
Severity: | normal | ||
Priority: | High | ||
Version: | unspecified | ||
Hardware: | AMD64 | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: |
Comments out parts of perl's Configure script which cause the segfaults
[ebuild R ] dev-lang/perl-5.10.1-r12 USE="-berkdb -debug -doc -gdbm -ithreads" This is the complete build log which shows what happens without the patch |
Description
dougztr
2009-10-19 08:51:22 UTC
Created attachment 207539 [details, diff]
Comments out parts of perl's Configure script which cause the segfaults
Can you please paste the part where it fails? Created attachment 207666 [details]
[ebuild R ] dev-lang/perl-5.10.1-r12 USE="-berkdb -debug -doc -gdbm -ithreads"
Created attachment 207667 [details] This is the complete build log which shows what happens without the patch I was told by a developer on #gentoo-embedded that this particular issue is related to https://bugs.gentoo.org/149292 and might better be corrected with the associated 50_all_gcc44_no_ssp_tls_uclibc.patch (In reply to comment #4) > I was told by a developer on #gentoo-embedded that this particular issue is > related to https://bugs.gentoo.org/149292 and might better be corrected with > the associated 50_all_gcc44_no_ssp_tls_uclibc.patch Do I get it right: this means we don't have to change the perl ebuild? I tested the 50_all_gcc44_no_ssp_tls_uclibc.patch and it did indeed resolve the issue, so I would suggest rather to patch gcc if "use elibc_uclibc" is detected because it would in theory fix any other programs which also use the -fstack-protector switch instead of just fixing issues with the Perl package. |