AUDIT: r15548 - gnucash/trunk/src/import-export/hbci - Really clean up enter/leave pairs. Avoid duplicated sanity checks.

[Christian Stimming]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Derek Atkins schrieb:
>> The bugfix is what has been added in r15531 - a strcmp() on potentially NULL 
>> that shows up in the crash in bug#405851. In r15531, this is avoided by 
>> checking and returning prematurely if there is NULL instead of an object, and 
>> additionally by using safe_strcmp.
> 
> Well, backporting 15539 is proving to be really painful.  I've just
> spent the better part of the last hour attempting to backport and
> there are LOTS of conflicts.  As a result of this difficulty I'm going
> to politely refuse to backport these three changesets (15339, 15542,
> 15548) as-is.

Oh, if it's really that involved, then I think the already back-ported
r15531 is just enough and the three others (15539, 15542, 15548) should
not get back-ported. I was just puzzled because a "g_return_val_if_fail
undeclared" error showed up on my development system, which means the
respective file might be missing an include for some glib versions. But
if you don't see this, all is fine obviously.

> If you want to get this merged into 2.0 I recommend you create a
> "bugfix branch" off of the 2.0 branch where you can backport the
> changes without fear of destroying 2.0, and then we can audit and
> merge the 'pre-merged' changeset.  I'm just afraid that merging
> directly into 2.0 from here requires too much hand-patching and
> something might go wrong.

Thanks for the explanation. But I was probably wrong with the AUDIT
request of the three changesets anyway.

Christian
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.1 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iQCVAwUBRdq212XAi+BfhivFAQLqNAP/X1xDPYbFRSTjquhNHBmH2ldmd4JlwvoS
WXYOFFgiGznKLMNmVXh/WieYaDqlBu+1Llw68neCkMVAMajmipc126oMUgos9+mQ
6e8c+kgmNSNTqkeGE/WpqRSoQhaDb6J8ldFPG+A+VkMv34JV5oX0jB3e9hUGMlds
wzqXFyaPQF8=
=tBu3
-----END PGP SIGNATURE-----