Update of bug #18796 (project freeciv):
Planned Release: 2.3.1,2.4.0 = 2.3.1, 2.4.0, 2.5.0
___
Reply to this item at:
http://gna.org/bugs/?18796
___
Message sent
Update of bug #18796 (project freeciv):
Status: Ready For Test = Fixed
Open/Closed:Open = Closed
___
Reply to this item at:
Follow-up Comment #10, bug #18796 (project freeciv):
Since I've ended up with this ticket, I'll apply Matthias' patch to S2_3,
S2_4, and trunk soon.
___
Reply to this item at:
http://gna.org/bugs/?18796
Follow-up Comment #8, bug #18796 (project freeciv):
File #14361 does make the segfaults go away and introduces no assertion
failures. Thanks.
(Nit: there's now an orphaned comment /* human players also need this for
building advice */.)
It looks like the same issue ought to apply to trunk,
Update of bug #18796 (project freeciv):
Planned Release: 2.3.1 = 2.3.1,2.4.0
___
Follow-up Comment #9:
(Nit: there's now an orphaned comment /* human players also need
this for building advice
Follow-up Comment #6, bug #18796 (project freeciv):
Could you test the attached patch (on top of S2_3; svn20364)? I can't
reproduce the bug but it should fix it. The patch moves the start of the phase
_before_ the call chain which resulted in the segfault (the bugtrace in
comment #1 was
Follow-up Comment #7, bug #18796 (project freeciv):
if needed the corresponding patch for trunk ...
(file #14362)
___
Additional Item Attachment:
File name: 0019-trunk-fix-crash-due-to-not-started-advisor-phase.patch Size:1
KB
Follow-up Comment #4, bug #18796 (project freeciv):
For trunk, the reason for this has to be fixed.
We don't have evidence that the problem exists on trunk. Could it have been
fixed in the many rearrangements that have happened since S2_3 branched?
I can't reproduce the issue from the savegame
Follow-up Comment #5, bug #18796 (project freeciv):
Should the KLUDGE for stable branch added in bug #18095 have
caught this?
Answering myself: it doesn't, because in that case ai_data_phase_done() is
called before ai_data_get() returns; this leaves static members of ai_data as
they are
Update of bug #18796 (project freeciv):
Status: Confirmed = Ready For Test
Assigned to:None = jtn
Summary: Server segfault on turn change = Server segfault
on turn change (auto-explore)
URL:
http://gna.org/bugs/?18796
Summary: Server segfault on turn change
Project: Freeciv
Submitted by: jtn
Submitted on: Sat Oct 8 12:01:57 2011
Category: general
Severity: 3 - Normal
Follow-up Comment #1, bug #18796 (project freeciv):
In ai_channel(), ai-channels is NULL (hence the segfault, presumably), and
ai-phase_is_initialized is FALSE.
I'm guessing this is one of the sort of bugs that would have been signalled
by the assertion failure which was disabled for S2_3 in
Follow-up Comment #2, bug #18796 (project freeciv):
In fact, the backtrace looks a lot like the one in bug #18095 comment 3
http://gna.org/bugs/?18095#comment3.
I'm not that familiar with this area; perhaps the band-aid applied in that
bug to minimise bad effects isn't sufficiently effective?
Follow-up Comment #3, bug #18796 (project freeciv):
Could you test to add an assert for ai-channels == NULL in ai_channels(). It
could help as a fast bug fix for 2.3 (can go on without crash). For trunk, the
reason for this has to be fixed. It seems to be that not all calls to AI
functions are
14 matches
Mail list logo
