From: <no...@so...> - 2002-12-08 07:50:05
|
Bugs item #650353, was opened at 2002-12-07 22:58 You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=110894&aid=650353&group_id=10894 Category: None Group: 8.4.1 Status: Open Resolution: None >Priority: 8 Submitted By: Nobody/Anonymous (nobody) Assigned to: David Gravereaux (davygrvy) Summary: crash in SocketThreadExitHandler Initial Comment: This is a bug that started appearing very recently. During Tcl_Finalize, the socket exit handler is trying to fetch data from a thread local storage index (line 562). The actual fetch succeeds. However, the return value is NULL, the SocketThreadExitHandler immediately tries to dereference this pointer (line 564) without checking for NULL. Checking these pointers for NULL would be nice... However, I believe this is just a symptom of a bigger problem which needs to be uncovered. Based on the diagnostic messages I added to the thread local storage functions, I believe that the data for this particular thread local storage index is NEVER set to a non-NULL value. SocketThreadExitHandler(void * 0x00000000) line 562 + 10 bytes SocketExitHandler(void * 0x00000000) line 528 + 9 bytes Tcl_Finalize() line 840 + 12 bytes I'm wondering if this code has been recently changed. -- JJM ---------------------------------------------------------------------- Comment By: Nobody/Anonymous (nobody) Date: 2002-12-07 23:10 Message: Logged In: NO A search for "SocketThreadExitHandler(clientData);" in the various CVS revisions reveals that this is/was broken as of CVS revision 1.28 (Wed Nov 27 00:58:01 2002 UTC (11 days, 6 hours ago)). JJM ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=110894&aid=650353&group_id=10894 |