[google] Avoid reading struct member from structure generated by different gcc version

This patch works around a segfault that can happen if we read in profiledata that is stored with an incompatible gcov_info structure.

To be applied only to google/main branch.

Tested by buildit --run-tests.
Noticed unrelated error in libmudflap.

2011-04-17  Martin Thuresson  <martint@google.com>

        * gcc/libgcov.c (gcov_version): Do not access filename in struct
        generated from different version.

[davidxl, 2011-04-18 19:59:13]

LGTM

http://codereview.appspot.com/4446047/diff/1/gcc/libgcov.c
File gcc/libgcov.c (right):

http://codereview.appspot.com/4446047/diff/1/gcc/libgcov.c#newcode155
gcc/libgcov.c:155: filename? filename : "", e, v);
Better split it into two different printfs as the format is different --
otherwise double : will be printed.

[richard.guenther_gmail.com, 2011-04-18 21:24:52]

On Mon, Apr 18, 2011 at 9:59 PM,  <davidxl@google.com> wrote:
> LGTM
>
>
> http://codereview.appspot.com/4446047/diff/1/gcc/libgcov.c
> File gcc/libgcov.c (right):
>
> http://codereview.appspot.com/4446047/diff/1/gcc/libgcov.c#newcode155
> gcc/libgcov.c:155: filename? filename : "", e, v);
> Better split it into two different printfs as the format is different --
> otherwise double : will be printed.
>
> http://codereview.appspot.com/4446047/

I'm confused.  Can we please avoid this kind of mails?

Thanks,
Richard.

[Diego Novillo, 2011-04-18 23:43:39]

On Mon, Apr 18, 2011 at 17:24, Richard Guenther
<richard.guenther@gmail.com> wrote:
> On Mon, Apr 18, 2011 at 9:59 PM,  <davidxl@google.com> wrote:
>> LGTM
>>
>>
>> http://codereview.appspot.com/4446047/diff/1/gcc/libgcov.c
>> File gcc/libgcov.c (right):
>>
>> http://codereview.appspot.com/4446047/diff/1/gcc/libgcov.c#newcode155
>> gcc/libgcov.c:155: filename? filename : "", e, v);
>> Better split it into two different printfs as the format is different --
>> otherwise double : will be printed.
>>
>> http://codereview.appspot.com/4446047/
>
> I'm confused.  Can we please avoid this kind of mails?

It's not altogether easy.  I filed an enhancement request for
Rietveld, but I'm not sure how eager they will be to implement this.
http://code.google.com/p/rietveld/issues/detail?id=291&sort=-id

Two workarounds:

- Reply to the patch using the Reply button in the web interface.
- Reply using to the initial e-mail generated by Rietveld.
- Cut and paste context from the window when replying via the web interface.

Medium/long term, we could probably install our own Rietveld instance
in gcc.gnu.org and modify it to our taste.  Either that or look for
another patch review tool.

If anyone has any other suggestions, please let me know.



Diego.