Re: [elinks-users] elinks-current + spidermonkey - bus error MacOS

2007-09-04 Thread Breen Mullins 
* Kalle Olavi Niemitalo [EMAIL PROTECTED] [2007-09-04 07:55 +0300]:

Breen Mullins [EMAIL PROTECTED] writes:


If JS_InitStandardClasses is somehow causing SpiderMonkey to call
window_set_property, then the private pointer has not yet been
set, and vs becomes NULL and causes a bus error.  Please try
moving the JS_SetPrivate call above the JS_InitStandardClasses
call and see if it helps.

That builds, and I get no bus error. Unfortunately, JS also appears not
to work. Is there a test case you could point me to that _should_ work?


It would also be interesting to know the names and line numbers
of the SpiderMonkey functions in the backtrace if you can get
debug symbols for those.  

I'll try to get that.

Breen
-- 
Breen Mullins
Menlo Park, California
___
elinks-users mailing list
elinks-users@linuxfromscratch.org
http://linuxfromscratch.org/mailman/listinfo/elinks-users

Re: [elinks-users] elinks-current + spidermonkey - bus error MacOS

2007-09-03 Thread Breen Mullins 
* Thomas Adam [EMAIL PROTECTED] [2007-09-02 16:50 +0100]:

On 02/09/07, Breen Mullins [EMAIL PROTECTED] wrote:
 I've been trying to get ECMA/javascript working with elinks.

 I'm running Mac OS 10.3.9. I installed spidermonkey from fink:


ispidermonkey1.60-2  Mozilla JavaScript-C Engine

Doesn't Fink provide a newer version?  Last time I checked, the
libraries were up to 1.8.1.4

The versions of spidermonkey and nspr that I got from fink yesterday 
appear to be up to date:

nspr-4.6.6-1: Netscape Portable Runtime
spidermonkey-1.60-2: Mozilla JavaScript-C Engine

which both match the mozilla website. (Actually moz says that nspr
is at 4.6.5, but the ftp site is up to 4.6.7)

I hacked the suggested linker flag bind_at_load into the Makefile, which
suppressed the error message, but I still get the bus error running the 
resulting binary. I'm out of ideas - can anybody enlighten me?

Breen
-- 
Breen Mullins
Menlo Park, California
___
elinks-users mailing list
elinks-users@linuxfromscratch.org
http://linuxfromscratch.org/mailman/listinfo/elinks-users

Re: [elinks-users] elinks-current + spidermonkey - bus error MacOS

2007-09-03 Thread Breen Mullins 
* Breen Mullins [EMAIL PROTECTED] [2007-09-03 10:56 -0700]:

 * Kalle Olavi Niemitalo [EMAIL PROTECTED] [2007-09-03 20:01 +0300]:

 Breen Mullins [EMAIL PROTECTED] writes:

 I hacked the suggested linker flag bind_at_load into the Makefile, which
 suppressed the error message, but I still get the bus error running the 
 resulting binary.

 Can you post a backtrace?

 kdump output is attached. Hope it's useful; I'm not too familiar with
 ktrace.

And here finally is a backtrace. For some reason I didn't have gdb on
this box, and the failure doesn't always generate a core file.

Here's what it does give me:


[ [EMAIL PROTECTED] src %] gdb ./elinks /cores/core.15024 
GNU gdb 5.3-20030128 (Apple version gdb-330.1) (Fri Jul 16 21:42:28 GMT 2004)
Copyright 2003 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you are
welcome to change it and/or distribute copies of it under certain conditions.
Type show copying to see the conditions.
There is absolutely no warranty for GDB.  Type show warranty for details.
This GDB was configured as powerpc-apple-darwin.
Reading symbols for shared libraries ... done
Core was generated by `./elinks'.
#0  AES_ctr128_inc () at window.c:142
142 struct document_view *doc_view = vs-doc_view;
(gdb) bt
#0  AES_ctr128_inc () at window.c:142
#1  0x0104dffc in ?? ()
#2  0x0104dffc in ?? ()
#3  0x0100e090 in ?? ()
#4  0x01088f38 in ?? ()
#5  0x01089120 in ?? ()
#6  0x0108c47c in ?? ()
#7  0x0108c600 in ?? ()
#8  0x0108d414 in ?? ()
#9  0x0100b314 in ?? ()
#10 0x00046f80 in aep_ctrl () at spidermonkey.c:184
#11 0x00046090 in ecmascript_get_interpreter (vs=0x285bbc8) at ecmascript.c:133
#12 0x0004652c in ecmascript_reset_state (vs=0x285b8dc) at ecmascript.c:233
#13 0x00028d80 in render_document (vs=0x0, doc_view=0x285bbc8, 
options=0x131fea0) at renderer.c:387
#14 0x00029064 in render_document_frames (ses=0x0, no_cache=1344) at 
renderer.c:465
#15 0x0008dd54 in draw_formatted (ses=0x130cc30, rerender=0) at draw.c:351
#16 0x00076060 in doc_loading_callback (download=0x1320aa0, ses=0xddc7c) at 
session.c:578
#17 0x00058200 in RAND_file_name () at connection.c:445
#18 0x000582a4 in done_connection (conn=0x1312040) at connection.c:462
#19 0x00058804 in WriteCompactIndex () at connection.c:620
#20 0x0005452c in XMLNameExpr () at select.c:289
#21 0x000537ec in main (argc=0, argv=0x285b8dc) at main.c:360

Breen
-- 
Breen Mullins
Menlo Park, California
___
elinks-users mailing list
elinks-users@linuxfromscratch.org
http://linuxfromscratch.org/mailman/listinfo/elinks-users

Re: [elinks-users] elinks-current + spidermonkey - bus error MacOS

2007-09-03 Thread Kalle Olavi Niemitalo 
Breen Mullins [EMAIL PROTECTED] writes:

 #0  AES_ctr128_inc () at window.c:142
 142 struct document_view *doc_view = vs-doc_view;

This is apparently not AES_ctr128_inc() but rather
window_get_property() in src/ecmascript/spidermonkey/window.c.
The vs pointer has been read with JS_GetPrivate from window_obj.

 #1  0x0104dffc in ?? ()
 #2  0x0104dffc in ?? ()
 #3  0x0100e090 in ?? ()
 #4  0x01088f38 in ?? ()
 #5  0x01089120 in ?? ()
 #6  0x0108c47c in ?? ()
 #7  0x0108c600 in ?? ()
 #8  0x0108d414 in ?? ()
 #9  0x0100b314 in ?? ()

Some SpiderMonkey functions, I suppose.

 #10 0x00046f80 in aep_ctrl () at spidermonkey.c:184

This would be spidermonkey_get_interpreter() in
src/ecmascript/spidermonkey.c.  I have the following code here
(in ELinks 0.13.GIT commit 1fd2a77416 from 2007-08-12):

179 window_obj = JS_NewObject(ctx, (JSClass *) window_class, NULL, NULL);
180 if (!window_obj) {
181 spidermonkey_put_interpreter(interpreter);
182 return NULL;
183 }
184 JS_InitStandardClasses(ctx, window_obj);
185 JS_DefineProperties(ctx, window_obj, (JSPropertySpec *) window_props);
186 JS_DefineFunctions(ctx, window_obj, (JSFunctionSpec *) window_funcs);
187 JS_SetPrivate(ctx, window_obj, interpreter-vs); /* to @window_class */

If JS_InitStandardClasses is somehow causing SpiderMonkey to call
window_set_property, then the private pointer has not yet been
set, and vs becomes NULL and causes a bus error.  Please try
moving the JS_SetPrivate call above the JS_InitStandardClasses
call and see if it helps.

It would also be interesting to know the names and line numbers
of the SpiderMonkey functions in the backtrace if you can get
debug symbols for those.  We could then check those functions in
the SpiderMonkey source, find the version of SpiderMonkey where
this first started occurring, and log that in elinks/NEWS.
___
elinks-users mailing list
elinks-users@linuxfromscratch.org
http://linuxfromscratch.org/mailman/listinfo/elinks-users

[elinks-users] elinks-current + spidermonkey - bus error MacOS

2007-09-02 Thread Breen Mullins 
I've been trying to get ECMA/javascript working with elinks.

I'm running Mac OS 10.3.9. I installed spidermonkey from fink:


   ispidermonkey1.60-2  Mozilla JavaScript-C Engine
   ispidermonkey-dev1.60-2  Mozilla JavaScript-C Engine
   ispidermonkey-shlibs 1.60-2  Mozilla JavaScript-C Engine

(This needs a hack to configure, which doesn't know about the fink
default install location in /sw:


--- configure.orig  Sun Sep  2 06:52:19 2007
+++ configure   Sun Sep  2 06:52:51 2007
@@ -17932,7 +17932,7 @@
if test ! -d $withval; then
withval=;
fi
-   for spidermonkeydir in $withval  /usr /usr/local /opt/spidermonkey 
/opt/js; do
+   for spidermonkeydir in $withval  /usr /usr/local /opt/spidermonkey 
/opt/js /sw; do
for spidermonkeyinclude in /include /include/js 
/include/smjs /include/mozjs; do
for spidermonkeylib in js smjs mozjs; do
if test $cf_result = no 
)

I get a big warning at the end of the make:


gcc: unrecognized option `-rdynamic'
ld: warning suggest use of -bind_at_load, as lazy binding may result in errors 
or different symbols being used
symbol _poll used from dynamic library /sw/lib/libnspr4.1.dylib(unix.o) not 
from earlier dynamic library /usr/lib/libSystem.B.dylib(poll.o)

and running the resulting binary gives a bus error. 
Looks like a conflict, but I don't know what to do from here. Advice is 
welcome. Thanks!

Breen

-- 
Breen Mullins
Menlo Park, California
___
elinks-users mailing list
elinks-users@linuxfromscratch.org
http://linuxfromscratch.org/mailman/listinfo/elinks-users

Re: [elinks-users] elinks-current + spidermonkey - bus error MacOS

2007-09-02 Thread Thomas Adam 
On 02/09/07, Breen Mullins [EMAIL PROTECTED] wrote:
 I've been trying to get ECMA/javascript working with elinks.

 I'm running Mac OS 10.3.9. I installed spidermonkey from fink:


ispidermonkey1.60-2  Mozilla JavaScript-C Engine

Doesn't Fink provide a newer version?  Last time I checked, the
libraries were up to 1.8.1.4

(Small world, eh, Breen?)

-- Thomas Adam
___
elinks-users mailing list
elinks-users@linuxfromscratch.org
http://linuxfromscratch.org/mailman/listinfo/elinks-users

Re: [elinks-users] elinks-current + spidermonkey - bus error MacOS

2007-09-02 Thread Breen Mullins 
* Thomas Adam [EMAIL PROTECTED] [2007-09-02 16:50 +0100]:

On 02/09/07, Breen Mullins [EMAIL PROTECTED] wrote:
Doesn't Fink provide a newer version?  Last time I checked, the
libraries were up to 1.8.1.4

Hmm. I ran a self-update before installing spidermonkey. Have to 
look into that.

(Small world, eh, Breen?)

Indeed! Thanks for the hint.

Breen
-- 
Breen Mullins
Menlo Park, California
___
elinks-users mailing list
elinks-users@linuxfromscratch.org
http://linuxfromscratch.org/mailman/listinfo/elinks-users