Thue Janus Kristensen <thuejk@gmail.com> writes:
> I finally succeeded in creating a test case, after much experimentation.
> Running the attached sql crashes my postgresql server 100% of the time.
Cute problem. The attached fix cures it for me; would you see if it
solves your original case too?
regards, tom lane
Index: src/backend/commands/trigger.c
===================================================================
RCS file: /cvsroot/pgsql/src/backend/commands/trigger.c,v
retrieving revision 1.248.2.2
diff -c -r1.248.2.2 trigger.c
*** src/backend/commands/trigger.c 24 Jan 2010 21:49:31 -0000 1.248.2.2
--- src/backend/commands/trigger.c 19 Aug 2010 15:14:06 -0000
***************
*** 2534,2539 ****
--- 2534,2540 ----
else
events->tail->next = chunk;
events->tail = chunk;
+ /* events->tailfree is now out of sync, but we'll fix it below */
}
/*
***************
*** 2935,2940 ****
--- 2936,2950 ----
{
chunk->freeptr = CHUNK_DATA_START(chunk);
chunk->endfree = chunk->endptr;
+
+ /*
+ * If it's last chunk, must sync event list's tailfree too. Note
+ * that delete_ok must NOT be passed as true if there could be
+ * stacked AfterTriggerEventList values pointing at this event
+ * list, since we'd fail to fix their copies of tailfree.
+ */
+ if (chunk == events->tail)
+ events->tailfree = chunk->freeptr;
}
}