[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: (ITS#4773) syncrepl segfault with incorrect config

Adrian Bridgett wrote:
> On Sat, Dec  9, 2006 at 22:29:50 +0100 (+0100), Pierangelo Masarati wrote:
>> It is already fixed in HEAD; fixing it in re23 might be too 
>> difficult to consider, so, since the correct behavior is clearly 
>> documented, it might be not worth the effort.
> Fair enough,  I'm sure I searched through the bugs for syncrepl items
> and didn't see anything, but I guess it was fixed outside of a bug
> report which would explain it.

Well, AFAIR it was not fixed as a consequence of a bug report; rather, 
HEAD code provides support to tag overlays as specific for the global or 
per-database level, if their use is not appropriate at all levels.  
slapo-syncprov in HEAD checks if it's being instantiated at the right 
level, otherwise it bails out.  What actually happens if it's 
instantiated at the wrong level, namely what causes the core dump you 
reported, is not ficed at all, as it makes no longer sense.  That's what 
I meant when I said that fixing this issue in re23 might not be worth: 
support for tagging overlays and to check if they're appropriately 
instantiated is interspersed in many code changes which may not be 
trivial to isolate.