Subj : Re: Access violation in JavaScript-C 1.5 pre-release 5 2003-01-10 for Win32 To : netscape.public.mozilla.jseng From : "Rob Swindell" Date : Tue May 06 2003 12:28 am "Brendan Eich" wrote in message news:3EB74A13.1030005@meer.net... > Rob Swindell wrote: > > >jshash.c line 180: access violation > >stack trace: > >JS_HashTableRawLookup(JSHashTable * 0x03186c00, unsigned long 6223614, const > >void * 0x02f7b7f0) line 180 + 3 bytes > >js_PurgeDeflatedStringCache(JSString * 0x02f7b7f0) line 2597 + 19 bytes > >js_FinalizeStringRT(JSRuntime * 0x023a7428, JSString * 0x02f7b7f0) line 2633 > >+ 9 bytes > >js_FinalizeString(JSContext * 0x0224b390, JSString * 0x02f7b7f0) line 2612 + > >16 bytes > >js_GC(JSContext * 0x0224b390, unsigned int 2) line 1311 + 11 bytes > >js_ForceGC(JSContext * 0x0224b390, unsigned int 2) line 993 + 13 bytes > >js_DestroyContext(JSContext * 0x0224b390, int 2) line 225 + 11 bytes > >JS_DestroyContext(JSContext * 0x0224b390) line 895 + 11 bytes > >js_service_thread(void * 0x01021d60) line 865 + 15 bytes > >_threadstart(void * 0x01021cc0) line 173 + 13 bytes > >KER > > > >Sorry, I don't know the specific steps to reproduce the problem as this was > >running an NNTP service (written in JS) and some combination of client > >commands caused this transient error. > > > >Is this a known problem? > > > >If not, would someone enter it into bugzilla for me (I hate that thing)? > > > >I also get occasional exceptions in JS_HashTableRawRemove (jshash.c line > >306). I wonder if that JS_ASSERT line is correct? > > > > Yes, that assertion is correct. > > If it's botching, it means you've mutated a deflated string key so that > it matches another entry in the table. Hmm.... in English? :-) > Is it possible you're calling free on the result of a JS_GetStringBytes call? I searched through my code and could find no such cases. And generally, I already know that would be a "Bad Thing {tm}", so I would've been very surprised to find such a bug. Does this assertion appear related to the access violation in JS_HashTableRawLookup or are they two unrelated anomolies? Sorry I'm not familiar with the inner-workings of the Spidermonkey GC engine. :-( Thanks again, -Rob .