-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On 10/05/2013 05:54 PM, Bob Proulx wrote:
> Hmm... Regarding valgrind... The coreutils gets run through
> valgrind routinely. There are always lots of false positives.
Well... in this case the 8-byte unfreed region is not. The bug
apparently lies in eglibc, though.
About the uninitialized bytes, I don't know yet. It *looks* like a
false positive, but can't yet assert it.
>> But anyway, that was not even the point. The point was that
>> debugging and bug reporting is difficult because there are no
>> symbols to provide complete backtraces to developers.
>
> Is there anyone that doesn't have the source code available? By
> the time I am geared to up to debug something in coreutils I always
> have the source code available and will use it for the debugging.
> I am not sure there will be much return on investment for the work
> to create and maintain a dbg package.
I really don't know how much work is needed to create a -dbg package,
and I don't have the time to do it either. I know it is hard work and
I am certainly not demanding it. But just let me explain why it is
more important than just blowing it off with "meh... I don't think
it's not worth it".
Source code is good enough to do development, but not to encourage it.
Sure, I can check the source code... but once it starts to get harder
than my current skills I will just give up on it. At this moment, my
current skills include investigating and providing a good bug report,
even possibly to upstream. My skills don't include, yet, understanding
the exact implications of the INTERNAL_SYSCALL macro, for example.
Reporting the bug has turned pointless now; you know what happens:
incomplete debugging information makes it more difficult for upstream.
That's why if I report a bug without a full backtrace, chances of the
bug being disregarded as NEEDINFO raise significantly. And upstream
will just blame the reporter.
So "return of investment" is not delimited to you; it is for the whole
community. You help me, I help upstream, easier to have the bug fixed
for everybody ("everybody" >= Debian).
I'm also wondering if it is at all feasable that debuild automatically
creates it.
> And by the way... It would be nice if the original subject were
> maintained. With just a ping about a bug number it means everyone
> needs to go look up the bug number. It would be nicer if it the
> subject had told us coreutils-dbg as in the original report.
Will surely keep that in mind and suggest an easy fix to the Debian
BTS: append ?subject=SUBJECT to the "Reply" link. Thanks for pointing
it out.
Thanks.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.14 (GNU/Linux)
Comment: Using GnuPG with Icedove - http://www.enigmail.net/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=7HJa
-----END PGP SIGNATURE-----
--
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]
