This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH v1.1] __gconv_translit_find: Actually append ".so" to module name [BZ #17187]
- From: "Carlos O'Donell" <carlos at redhat dot com>
- To: Florian Weimer <fweimer at redhat dot com>, Andreas Schwab <schwab at suse dot de>
- Cc: GNU C Library <libc-alpha at sourceware dot org>
- Date: Tue, 26 Aug 2014 10:36:43 -0400
- Subject: Re: [PATCH v1.1] __gconv_translit_find: Actually append ".so" to module name [BZ #17187]
- Authentication-results: sourceware.org; auth=none
- References: <53CD0F15 dot 3030806 at redhat dot com> <53F62D4F dot 3040702 at redhat dot com> <mvmvbpfqwox dot fsf at hawking dot suse dot de> <53FC7CA5 dot 2000100 at redhat dot com>
On 08/26/2014 08:25 AM, Florian Weimer wrote:
> On 08/26/2014 02:09 PM, Andreas Schwab wrote:
>> Florian Weimer <fweimer@redhat.com> writes:
>>
>>> + /* This function always fails. Transliteration module loading is
>>> + not implemented. */
>>> + return 1;
>>
>> Since it always fails you can just remove the function completely.
>
> If taken to the logical conclusion, this has a ripple effect and is not suitable for backporting:
>
> <https://sourceware.org/ml/libc-alpha/2014-08/msg00119.html>
>
> So I had to stop somewhere, and I think the patch from last week is a reasonable compromise for backports and the 2.20 release.
Yes, and no. The ripple effect is still present in that all callers
must actually handle the function returning one instead of zero, thus
we still have to audit the callers. I tend to agree with Andreas
here, it's basically the same amount of audit work, *but* your smaller
patch will apply more easily to older branches, and reduces the
chance we have compiler/assembler/linker problems due to refactored
code.
Cheers,
Carlos.