Subj : Exceptions in js32.dll 1.5 rc5 (with stacktraces)
To   : netscape.public.mozilla.jseng
From : "Rob Swindell" <rob@synchro.net>
Date : Fri Sep 19 2003 01:42 am

As I previously posted, this version (1.5 rc5, thread-safe) seems to run
almost perfectly, with the occasional exception in JS_malloc().

Here are a couple example stack traces:

_heap_alloc_dbg(unsigned int 0x00000010, int 0x00000001, const char *
0x00000000, int 0x00000000) line 327
_nh_malloc_dbg(unsigned int 0x00000010, int 0x00000000, int 0x00000001,
const char * 0x00000000, int 0x00000000) line 242 + 21 bytes
malloc(unsigned int 0x00000010) line 130 + 21 bytes
JS_malloc(JSContext * 0x0a5330a0, unsigned int 0x00000010) line 1424 + 9
bytes
js_NewRegExp(JSContext * 0x0a5330a0, JSTokenStream * 0x00000000, JSString *
0x02f07e98, unsigned int 0x00000000, int 0x00000000) line 1185 + 22 bytes
js_NewRegExpOpt(JSContext * 0x0a5330a0, JSTokenStream * 0x00000000, JSString
* 0x02f07e98, JSString * 0x00000000, int 0x00000000) line 1231 + 25 bytes
match_or_replace(JSContext * 0x0a5330a0, JSObject * 0x02f07e88, unsigned int
0x00000001, long * 0x02eae138, int (JSContext *, long, GlobData *)*
0x002d6530 match_glob(JSContext *, long, GlobData *), GlobData * 0x0266ea9c,
long * 0x0266ec0c) line 1129 + 34 bytes
str_match(JSContext * 0x0a5330a0, JSObject * 0x02f07e88, unsigned int
0x00000001, long * 0x02eae138, long * 0x0266ec0c) line 1251 + 34 bytes
js_Invoke(JSContext * 0x0a5330a0, unsigned int 0x00000001, unsigned int
0x00000000) line 843 + 26 bytes
js_Interpret(JSContext * 0x0a5330a0, long * 0x0266fb0c) line 2811 + 15 bytes
js_Execute(JSContext * 0x0a5330a0, JSObject * 0x02edf340, JSScript *
0x02f24438, JSStackFrame * 0x00000000, unsigned int 0x00000000, long *
0x0266fb0c) line 1024 + 13 bytes
JS_ExecuteScript(JSContext * 0x0a5330a0, JSObject * 0x02edf340, JSScript *
0x02f24438, long * 0x0266fb0c) line 3278 + 25 bytes
js_static_service_thread(void * 0x010488c4) line 1173 + 36 bytes
_threadstart(void * 0x010501a0) line 173 + 13 bytes
KERNEL32! 7c4e987

_heap_alloc_dbg(unsigned int 0x00000008, int 0x00000001, const char *
0x00000000, int 0x00000000) line 327
_nh_malloc_dbg(unsigned int 0x00000008, int 0x00000000, int 0x00000001,
const char * 0x00000000, int 0x00000000) line 242 + 21 bytes
malloc(unsigned int 0x00000008) line 130 + 21 bytes
JS_malloc(JSContext * 0x00c100f0, unsigned int 0x00000008) line 1424 + 9
bytes
js_InflateString(JSContext * 0x00c100f0, const char * 0x0266ec36, unsigned
int 0x00000003) line 2784 + 17 bytes
JS_NewStringCopyZ(JSContext * 0x00c100f0, const char * 0x0266ec36) line 3540
+ 17 bytes
js_NumberToString(JSContext * 0x00c100f0, double 313.00000000000) line 626 +
13 bytes
js_Interpret(JSContext * 0x00c100f0, long * 0x0266fb04) line 1833 + 32 bytes
js_Execute(JSContext * 0x00c100f0, JSObject * 0x00c0b740, JSScript *
0x02cc0240, JSStackFrame * 0x00000000, unsigned int 0x00000000, long *
0x0266fb04) line 1024 + 13 bytes
JS_ExecuteScript(JSContext * 0x00c100f0, JSObject * 0x00c0b740, JSScript *
0x02cc0240, long * 0x0266fb04) line 3278 + 25 bytes
js_static_service_thread(void * 0x010488c4) line 1174 + 36 bytes
_threadstart(void * 0x010501a0) line 173 + 13 bytes
KERNEL32! 7c4e987c()

These occur only intermittently (maybe once every 24-48 hours) when running
the largest long-running application/server script I have (an IRC daemon).

I tried to upgrade to 1.5 rc5a (as I posted previously), but am hitting an
exception in js_MarkScriptFilename (almost immediately) when using that
version.

Thanks for any pointers,

-Rob

.