"Chris Hoover" <[EMAIL PROTECTED]> writes: > #3 0x082c8fa6 in ExceptionalCondition ( > conditionName=0x83bc514 "!(((header->context) != ((void *)0) && > (((((Node*)((header->context)))->type) == T_AllocSetContext))))", > errorType=0x82f89d2 "BadArgument", fileName=0xd3f "", lineNumber=0) at > assert.c:57 > #4 0x082e58ce in pfree (pointer=0x89cde14) at mcxt.c:589 > #5 0x081f9620 in SysLogger_Start () at syslogger.c:1172
Ah-hah: copy-and-paste-o. This chunk of code is assigning the wrong thing to last_csvfile_name. /* instead of pfree'ing filename, remember it for next time */ if (last_csvfile_name != NULL) pfree(last_csvfile_name); last_csvfile_name = filename; } This is a bit distressing in terms of the apparent lack of developer testing on the CSV code; this should have been caught by anyone who'd used CSV logging with --enable-cassert for any length of time. regards, tom lane ---------------------------(end of broadcast)--------------------------- TIP 3: Have you checked our extensive FAQ? http://www.postgresql.org/docs/faq