Re: svn commit: r32558 - trunk/subversion/libsvn_fs_fs

Okay, yes, I think BDB was fine. The logic there is a little different from
that of FSFS, and that was causing me some doubt. Thanks (seriously, not
sarcastically) for causing me to double-check that logic Just In Case.

C. Michael Pilato wrote:
> Hrm. Let me double-check the surrounding code, though. Something doesn't
> feel right...
>
>
> C. Michael Pilato wrote:
>> No, the BDB logic jumps out of the loop if the pred_id is NULL before trying
>> to dup it.
>>
>>
>> Kamesh Jayachandran wrote:
>>> Is it not applicable for bdb?
>>>
>>> With regards
>>> Kamesh Jayachandran
>>>
>>> cmpilato_at_tigris.org wrote:
>>>> Author: cmpilato
>>>> Date: Tue Aug 19 10:23:16 2008
>>>> New Revision: 32558
>>>> Log:
>>>> Fix SEGFAULT introduced in r32545.
>>>> * subversion/libsvn_fs_fs/tree.c
>>>> (fs_node_origin_rev): Avoid dup'ing a NULL pointer.
>
>

-- 
C. Michael Pilato <cmpilato_at_collab.net>
CollabNet   <>   www.collab.net   <>   Distributed Development On Demand