Subj : Re: Exception in js32.dll 1.5 rc5a (with stacktrace) To : netscape.public.mozilla.jseng From : "Rob Swindell" Date : Fri Sep 19 2003 08:16 pm "Brendan Eich" wrote in message news:3F6B702C.3000806@meer.net... > Rob Swindell wrote: > > > "Brendan Eich" wrote in message > > news:3F6B40BD.1030203@meer.net... > > > >>Rob Swindell wrote: > >> > >> > >>>I've been having intermittent exceptions in js32.dll 1.5 rc5 > > > > (thread-safe) - > > > >>>I'll post about those next. > >>> > >>>Anyway, so I've been trying to upgrade to 1.5 rc5a with no success. I, > >>>almost immediately, get this assertion/exception when running one of my > >>>bigger application scripts (an IRC server): > >>> > >>>js_MarkScriptFilename(const char * 0x02a1282d) line 960 + 49 bytes > >> > >> > >>What CVS revision of jsscript.c do you have? > > > > > > 3.46 > > > That's the latest. I trust jsemit.[ch] are up to date also. Yes. > Are you using jsxdrapi.h or any XDR stuff? No (not yet). > > I am using CVS, and I am updated to the "latest". The version string is > > currently "JavaScript-C 1.5 pre-release 5a 2003-05-29", but I guess that's > > only accurate for the tarballs? In that case, I'll run some tests with a DLL built from the 5a tarball too. For some reason, I thought that cvs-mirror.mozilla.org was used for "official" releases only. Perhaps the version string should be changed between release to note that it is a "unstable" build? > Right. > > > > Should I be testing using an ftp tarball instead of from CVS? > > Testing the trunk is fine, we need more of that. Thanks for doing so. > If you need something stabler, use the last tarball. On a different computer, I just checked-out the latest from the CVS trunk (into a fresh directory tree) and I get the same exception in js32.dll: NTDLL! 77fa144b() js_MarkScriptFilename(const char * 0x0229630d) line 960 + 49 bytes js_MarkScript(JSContext * 0x02199198, JSScript * 0x02296220, void * 0x00000000) line 1125 + 12 bytes fun_mark(JSContext * 0x02199198, JSObject * 0x02277e88, void * 0x00000000) line 1344 + 20 bytes js_Mark(JSContext * 0x02199198, JSObject * 0x02277e88, void * 0x00000000) line 3714 + 18 bytes js_MarkGCThing(JSContext * 0x02199198, void * 0x02277e88, void * 0x00000000) line 860 + 37 bytes js_MarkGCThing(JSContext * 0x02199198, void * 0x0219ab40, void * 0x00000000) line 914 + 17 bytes js_MarkGCThing(JSContext * 0x02199198, void * 0x0219ab90, void * 0x00000000) line 914 + 17 bytes gc_root_marker(JSDHashTable * 0x02196120, JSDHashEntryHdr * 0x023f1a64, unsigned long 0, void * 0x02199198) line 970 + 17 bytes JS_DHashTableEnumerate(JSDHashTable * 0x02196120, int (JSDHashTable *, JSDHashEntryHdr *, unsigned long, void *)* 0x00275fb0 gc_root_marker(JSDHashTable *, JSDHashEntryHdr *, unsigned long, void *), void * 0x02199198) line 618 + 34 bytes js_GC(JSContext * 0x02199198, unsigned int 0) line 1183 + 21 bytes js_ForceGC(JSContext * 0x02199198, unsigned int 0) line 995 + 13 bytes JS_GC(JSContext * 0x02199198) line 1675 + 11 bytes JS_MaybeGC(JSContext * 0x02199198) line 1694 + 9 bytes js_BranchCallback(JSContext * 0x02199198, JSScript * 0x023f0950) line 893 + 12 bytes js_Interpret(JSContext * 0x02199198, long * 0x02b7fb04) line 1567 + 110 bytes js_Execute(JSContext * 0x02199198, JSObject * 0x0219ab40, JSScript * 0x023f0950, JSStackFrame * 0x00000000, unsigned int 0, long * 0x02b7fb04) line 1055 + 13 bytes JS_ExecuteScript(JSContext * 0x02199198, JSObject * 0x0219ab40, JSScript * 0x023f0950, long * 0x02b7fb04) line 3384 + 25 bytes js_static_service_thread(void * 0x00bf55a8) line 1179 + 36 bytes _threadstart(void * 0x00bf2048) line 173 + 13 bytes KERNEL32! 7c4e987c() I'll see about building from the 1.5 rc5a tarball now... -Rob .