Magic-Mirror/gecko-dev
1
0
Fork 0
mirror of https://github.com/mozilla/gecko-dev.git synced 2024-12-29 03:44:37 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
64,882 Commits 616 Branches 98 Tags 5.9 GiB
da45d9b0eb
Commit Graph

978 Commits

Author SHA1 Message Date
drapeau%eng.sun.com
7f384b86a6 Two main things: 
1) Fix for 23775 (three files changed for this)

2) Many OJI unit tests added.
 2000-06-23 02:28:31 +00:00
pavel%gingerall.cz
f8ce45f144 - workaround of -rdynamic (JS_PERLCONNECT only) 2000-06-16 08:56:03 +00:00
brendan%mozilla.org
5c53e520b1 sfraser@netscape.com's fine GC_MARK_DEBUG enhancement for XPConnect. 2000-06-16 04:34:00 +00:00
pavel%gingerall.cz
fff3bf7638 - fixes in original version of perlconnect (JSVALToSV SVToJSVAL etc.) 
- object delegation (like JSCreateObject) Perl->JS
- ParlValue handles PerlObject correctly
- undef values handled correctly (in both directions)
- JS arrays may be tied to perl arrays
- error handlers supported on Perl side
- no globals
- several minor fixes
 2000-06-14 07:23:58 +00:00
pavel%gingerall.cz
0c07208dc5 - minor change of jsperl.h inclusion, matters for perlconnect build only 2000-06-14 07:18:41 +00:00
rginda%netscape.com
2ebc0a5b20 -- NOT PART OF THE BUILD -- 
Conditionally build lcshell w/ jdk 1.1.8 or 1.2.2 on windows
 2000-06-14 00:14:56 +00:00
mkaply%us.ibm.com
737d9fb1b4 # 37239 
r = mccabe, a = brendan
OS/2 bring-up - PR_CALLBACK for VisualAge
 2000-06-14 00:07:08 +00:00
mccabe%netscape.com
3a709facd8 First part of fix for 38495, support for exposing plugin methods to JavaScript. 
This patch teaches XPConnected objects to look in their JavaScript __proto__ chain for any names they can't resolve themselves.  The rest of the fix to this bug sets the original DOM node object as the prototype of a new xpconnect-exposed plugin object, so javascript accesses will see names from both objects.

r=jst,brendan
a=beard
 2000-06-13 23:18:21 +00:00
scc%mozilla.org
507357239c fix type equivalence between |PRUnichar| and |jschar| now that |PRUnchar| may be |wchar_t| on select platforms 2000-06-12 23:52:31 +00:00
brendan%mozilla.org
d8f3cd1921 Comply with weird ECMA nit: call (o.f)() (note parens around the function expression) must bind 'this' to the global object, not to o\! (41864, r=shaver). 2000-06-08 06:46:18 +00:00
brendan%mozilla.org
73d4167370 Use localizable error message for out of memory. 2000-06-06 04:54:04 +00:00
brendan%mozilla.org
28d3dcb5fc Better fix, really (r=shaver for sure). 2000-06-06 04:41:05 +00:00
brendan%mozilla.org
4e92401f64 Better control flow for catch clause code generation (r=shaver). 2000-06-06 04:27:37 +00:00
mccabe%netscape.com
3b337ab6af Fix to potential leak introduced with fix to 40406. 
Be conservative in handling the lifetime of the safe context created by XPConnect to execute JS Components, and save it off to be destroyed at cleanup time, even if some other safe context is registered with XPConnect via SetSafeJSContext.

r=vishy, a=brendan
 2000-06-06 00:01:25 +00:00
brendan%mozilla.org
d2c7c21d1b Avoid zero-length malloc (and assertbotch) in array_sort, just return true early\! 2000-06-03 19:00:28 +00:00
warren%netscape.com
512c8bf433 Renaming nsIAllocator to nsIMemory (and nsAllocator to nsMemory). API cleanup/freeze. Bug #18433 2000-06-03 09:46:12 +00:00
brendan%mozilla.org
6ca20f928f Fix ECMA DontDelete compliance problems, which create getter/setter security holes (40760, r=shaver). 2000-06-02 00:02:46 +00:00
mkaply%us.ibm.com
716fff7b7c # 40177 
r = leaf, a = brendan
Fix tab in makefile
 2000-06-01 14:15:39 +00:00
brendan%mozilla.org
aca040859b Use JS_ValueToId to go from user to internal property id, for integer-id optimality (40731, r=shaver). 2000-05-31 22:10:53 +00:00
brendan%mozilla.org
d84057e951 Make JS_ExecuteScriptPart call the debugger hooks (41066, r=MyNGs@HotMail.com). 2000-05-31 21:57:46 +00:00
brendan%mozilla.org
6a220f2f41 Fix unreviewed changes made to fix 'Uninitialized variable compiler warnings'. 2000-05-28 00:02:26 +00:00
jst%netscape.com
60b53dbf6e Fixing bustage on solaris native builds. reported and reviewed by Tomi.Leppikangas@oulu.fi 2000-05-27 13:14:31 +00:00
edburns%acm.org
67cfd5f36b r=brendan 
a=brendan
bug: 27362

This fix makes it so nsCLiveconnect.cpp doesn't #include
files within an extern "C" {} block.  To make this work, I
simply moved the extern "C" {} to the minimum necessary
range.  This required placing an "ifdef __cplusplus extern "C""
block in jsj_private.h, since nsCLiveconnect.cpp is the only
c++ file that includes jsj_private.h.
 2000-05-27 01:12:40 +00:00
shaver%mozilla.org
634ad644e4 use NS_METHOD for register/unregister callback, for Win32 2000-05-26 05:55:00 +00:00
shaver%mozilla.org
12fe3f7799 fix build bustage -- I'm an ass 2000-05-26 05:18:31 +00:00
shaver%mozilla.org
ee8d2a82fe remove useless nsIModule paste (#22922), add debug() for JS components, use categories for component loaders, XPCONNECT_STANDALONE (#39875), r=dp,waterson a=brendan 2000-05-26 04:56:23 +00:00
mkaply%us.ibm.com
367ea5b7df # 34106 
r = leaf, a = brendan
Mistake was made with an earlier makefile checkin - removing double definition of EXTRA_DSO_OPTS
 2000-05-25 14:25:04 +00:00
mccabe%netscape.com
b05da32465 Fix the fix to 40406, fixing bustage. 
Don't destroy the per-thread safe JS context on cleanup if that context was provided through the setter.

r=waterson
 2000-05-25 08:33:53 +00:00
mccabe%netscape.com
ba80f85fb1 Fix for 40406. Add setter to XPConnect to set default JS context used to execute JS components when no appropriate context can be found. 
The setter changes per-thread data - for the DOM/UI thread, this means that we can set the JS Context to one with DOM magic.  This magic allows some DOM JS <-> XPConnect JS conversions to succeed, and in particular allows creation of a DOM window from within a JS component.

Unblocks nsbeta2 work by Vishy and Pavlov.

r=mccabe,vishy.  a=brendan,sleestack.
 2000-05-24 22:54:57 +00:00
cls%seawood.org
cb31f6e62d More OS/2 changes. Batches 19 & 20 of Bug #34106 r=pavlov 2000-05-21 13:32:11 +00:00
brendan%mozilla.org
addb7242fb Fix remove-while-enumerating, beef up metering and add JS_DHashTableDumpMeter (these files are not part of the client build). 2000-05-20 18:42:22 +00:00
brendan%mozilla.org
0858bc61ea Fix old stddev calc bug (#ifdef'd code, not part of build). 2000-05-20 18:41:13 +00:00
cls%seawood.org
df5725e323 Finishing off the standalone modules framework. 
List js/src/liveconnect & js/src/xpconnect in toplevel Makefile.in to allow js to be built using the standalone framework.
Use the <module>_STANDALONE defines anytime BUILD_MODULES != all.
r=slamm
 2000-05-20 02:45:59 +00:00
brendan%mozilla.org
e1b56a4ce1 Use int for shift type consistently; optimize primary hash match (NOT PART OF BUILD). 2000-05-19 18:26:37 +00:00
brendan%mozilla.org
725ff208a6 Don't clear memory gratuitously (after create-error; after destroy). 2000-05-19 05:10:58 +00:00
brendan%mozilla.org
83d65f9508 Don't destroy what hasn't been created (NOT PART OF BUILD). 2000-05-19 03:49:32 +00:00
brendan%mozilla.org
91e4edcef8 Experimental double-hashing code, hope to donate to nsprpub/lib/ds (NOT PART OF BUILD). 2000-05-19 02:51:20 +00:00
mccabe%netscape.com
65950e22ec Patch xpcshell.cpp to not try to call JS_malloc with 0 bytes. 
Patch courtesy Mark G. Adams <madams@livepage.com>

r=mccabe

This code is part of xpcshell, and doesn't go into the mozilla product.  Prior to this fix, xpcshell crashed on startup.
 2000-05-19 00:34:46 +00:00
cls%seawood.org
a8000b8bb5 Mass replace of -lmozjs with $(MOZ_JS_LIBS) needed for OS/2 and consistency. 2000-05-17 06:45:45 +00:00
brendan%mozilla.org
7bc9169a36 Include jsstddef.h first in all js*.c files (r=tlundeen@webcrossing.com). 2000-05-17 06:24:45 +00:00
brendan%mozilla.org
7dd0a4f6c4 JS_DestroyRuntime nukes all contexts; add JS_G/SetRuntimePrivate (r=tlundeen@webcrossing.com). 2000-05-17 06:23:03 +00:00
pavlov%netscape.com
803190ad4b changes all usages of NS_NewTimer to CreateInstance 2000-05-17 02:49:35 +00:00
sfraser%netscape.com
7c1335fa2a Fix to only broadcast to the autoreg observer if we are really registering. 2000-05-17 01:54:18 +00:00
mkaply%us.ibm.com
68897b7076 # 37239 
r= rogerl
OS/2 bring up continues - fix for floating point trap
 2000-05-16 22:51:12 +00:00
rayw%netscape.com
18fabd3cae Added the fixes for 35866 and one older bug to the js component loader. 
This means it is more careful about how it writes binary filenames in
the registry and it calls an observer to report autoregistrations.

This should have gone out with equivalent Native Component Loader fixes,
but it didn't.
 2000-05-16 05:09:31 +00:00
cls%seawood.org
8a94c9df50 Latest OS/2 makefile changes. Bug 34106. Batch 14, 15, & 16. 2000-05-16 04:41:54 +00:00
brendan%mozilla.org
98af3bcf58 Fix GC request debiting to cope with multiple JSContexts on the same thread (39321, r=shaver). 2000-05-16 02:44:27 +00:00
brendan%mozilla.org
26edd3b893 Common a string literal (js_call_str; r=shaver). 2000-05-15 20:21:05 +00:00
brendan%mozilla.org
453918e0e2 Remove deadwood brought to my attention by tlundeen@webcrossing.com (r=shaver). 2000-05-15 20:04:41 +00:00
brendan%mozilla.org
d69376f26f Fix stupid 0-for-enumerator-equal-to-0 warning (r=shaver). 2000-05-15 20:03:13 +00:00