Subj : Re: JS_InitClass segfaults in my JS/Python bridge
To   : netscape.public.mozilla.jseng
From : jjl@pobox.com (John J. Lee)
Date : Sat Aug 02 2003 01:55 pm

Brendan Eich <brendan@meer.net> writes:

> John J. Lee wrote:
> > jjl@pobox.com (John J. Lee) writes:
> > Not sure this is much help, but here's a traceback (not sure what that
> > ?? is, either):
> > (gdb) bt
> > #0  0x081f7598 in ?? ()
> > #1  0x4040da5a in JS_InitClass (cx=0x81d6248, obj=0x81f6b40, parent_proto=0x81f6b40, clasp=0x8202500,
> >     constructor=0x403dc0f8 <__pyx_f_12spidermonkey_constructor_stub>, nargs=0, ps=0x0, fs=0x0, static_ps=0x0, static_fs=0x0) at jsapi.c:1828
> 
> Stacks are good, but in this case some more info is needed: what is
> *clasp?  Can you run under gdb and stop on line 1828 in jsapi.c?  If
[...]

D'Oh!

(gdb) p *clasp
$1 = {name = 0x82041b8 "Node", flags = 0, addProperty = 0x4040d9a4 <JS_PropertyStub>, delProperty = 0x4040d9a4 <JS_PropertyStub>,
  getProperty = 0x4040d9a4 <JS_PropertyStub>, setProperty = 0x4040d9a4 <JS_PropertyStub>, enumerate = 0x4040d9b4 <JS_EnumerateStub>,
  resolve = 0x4040d9c4 <JS_ResolveStub>, convert = 0x4040d9d4 <JS_ConvertStub>, finalize = 0x4040d9fc <JS_FinalizeStub>,
  getObjectOps = 0x81f7598, checkAccess = 0x81f75a8, call = 0x81f75b8, construct = 0x81f75c8, xdrObject = 0x81f75d8,
  hasInstance = 0x81f75e8, mark = 0x81f75f8, spare = 136279560}
(gdb)

Pyrex doesn't have an aggregate initialisation syntax, and I forgot to
explicitly set those pointers to NULL.

Sorry, I'm a moron.

Thanks!


> How did you build RC5?

Hmm, re-reading the build instructions, I'm not sure what this means:

| NOTE: Do not attempt to use Makefile to build the standalone
|  JavaScript engine.  This file is used only for building the JS-engine
|  in the Mozilla browser.

Does that mean I was wrong to build it this way?

make -f Makefile.ref

(with GNU make)

It seemed to work OK...


John

.