mozilla
/
pjs
зеркало из https://github.com/mozilla/pjs.git
1
0
Форкнуть 0
Код Задачи Пакеты Проекты Релизы Вики Активность
76 041 коммит 40 Ветки 136 Теги 1,1 GiB
Граф коммитов

3072 Коммитов

Автор SHA1 Сообщение Дата
waldemar%netscape.com 907e88a921 Added 'operator' non-reserved word 2000-12-02 09:09:41 +00:00
jeff.dyer%compilercompany.com e1e34965e1 Fixed a couple typos. 2000-12-02 02:04:51 +00:00
jeff.dyer%compilercompany.com 0305f2d290 Updated to reflect latest changes. 2000-12-02 01:49:18 +00:00
jeff.dyer%compilercompany.com 0f99399a6e Import of Initial Source from MVCC 2000-12-02 00:46:31 +00:00
pschwartau%netscape.com 643e118e14 Initial add; see Bugzilla bug 61266 - 2000-12-01 20:20:43 +00:00
pschwartau%netscape.com e2e1752e2e Initial add; see Bugzilla bug 57631 2000-12-01 20:17:11 +00:00
jband%netscape.com eaf3928939 fixes for bug 60303, bug 60356, bug 57370, and bug 58982 r=mccabe sr=hyatt sr=brendan 2000-11-30 06:58:37 +00:00
mccabe%netscape.com 6df0ad455d Fix to 61183. Detect whether the OS has given us a date string ending in /xx (as in xx/xx/xx) for Date.prototype.toLocale[Date]String, and pad it out to the real number of digits. 
r=rogerl
 2000-11-30 05:52:44 +00:00
mccabe%netscape.com 93184e2bfe Remove unused property definitions and fix some errors that had crept in. 2000-11-29 23:32:53 +00:00
mccabe%netscape.com b93f77c5b2 Twiddle Make usage message. 2000-11-29 22:36:06 +00:00
mccabe%netscape.com d9cd56c60b Fix to 58673. Alias Date.prototype.toGMTString to Date.prototype.toUTCString (rather than defining seperate function objects for each). ECMA conformance. 
r=rogerl
 2000-11-29 03:03:05 +00:00
mccabe%netscape.com 1509017ef1 Fix to 57048. Bring the JS engine's implementation of Object.prototype.propertyIsEnumerable in line with the ECMA spec. 
Thanks to David Flanagan for noticing this problem.

r=rogerl
 2000-11-29 02:05:10 +00:00
mccabe%netscape.com 6105620bbc Restore toplevel Makefile (with an added note in the usage message about it's new limitations wrt debugger and bsf) and update the js library makefile to account for changed files. 2000-11-28 21:24:41 +00:00
nboyd%atg.com 1d790c4961 Fix bug 61267. 2000-11-28 16:23:24 +00:00
mstoltz%netscape.com 668715ee7b bug 59766. Added code to push current JS context onto context stack before all LC calls into JS, becuase the security manager looks on the context stack for permission info. r=jband, mccabe 2000-11-28 02:50:21 +00:00
beard%netscape.com 870fcd146a Keep JDK 1.1 compatible. 2000-11-27 23:40:15 +00:00
nboyd%atg.com ac975755b3 Sense JDK1.3 dependencies. 2000-11-27 22:23:40 +00:00
pschwartau%netscape.com 807f22a27f Removing outdated testcase. See Bugzilla bug 49233. 2000-11-27 19:39:18 +00:00
pschwartau%netscape.com c9886e3a1e Removing outdated testcase. See bug 49233. 2000-11-27 19:32:41 +00:00
nboyd%atg.com a218a1a793 Add hasProperty and use it for the 'in' operator 2000-11-27 17:46:56 +00:00
nboyd%atg.com d05cf1d8e6 Fix 19734, accepting patch 2000-11-27 15:13:35 +00:00
nboyd%atg.com 03243413e6 Switch to ant for building. 2000-11-27 15:04:48 +00:00
nboyd%atg.com e33e06b0e8 Add debugger contribution from Christopher Oliver. 2000-11-27 15:00:45 +00:00
pschwartau%netscape.com daa5773e5c Initital add - 2000-11-27 01:22:47 +00:00
pschwartau%netscape.com fb875c9e90 Adding shell.js for ECMA3 Date tests, and tests for ECMA3 15.9.5.3 - 15.9.5.7 2000-11-25 01:17:24 +00:00
beard%netscape.com 9b8ac41686 importable XML project for CWPro6. 2000-11-23 04:27:33 +00:00
beard%netscape.com 2736e7e776 removed import of non-existent org.mozilla.javascript.tools.debugger.* package. 2000-11-23 04:24:34 +00:00
nboyd%atg.com 0380d9c7d9 * Add ContextListener interface 
* Add ability to access shell functions from compiled scripts, and to
  add shell functions to embedded apps more easily
 2000-11-22 15:00:05 +00:00
waldemar%netscape.com 6cd25e1b96 Added support for white space. Removed initial / production. 2000-11-21 04:07:05 +00:00
waldemar%netscape.com fa7d21a130 Renamed simple to span 2000-11-21 04:06:44 +00:00
brendan%mozilla.org e21159cb73 Store real JSString*, not address of auto JSSubString, in local root (60674, r=beard, sr=jband). 2000-11-21 00:51:43 +00:00
beard%netscape.com 6c99e43e31 updated dependencies 2000-11-21 00:31:10 +00:00
beard%netscape.com e1e12894c0 added explicit getString() method to StringFormatter, and removed operator String() methods. 2000-11-21 00:29:39 +00:00
beard%netscape.com 785eaacfb2 Made getStaticSlot() static, to remove warning. 2000-11-21 00:15:39 +00:00
beard%netscape.com 8329d7a4b6 fixing unsigned/signed conversion warnings. 2000-11-21 00:14:55 +00:00
nboyd%atg.com 7d318fbd85 Merge debugger changes into trunk. 2000-11-20 16:16:32 +00:00
nboyd%atg.com b7d36f7601 Fix for updated filename. 2000-11-20 14:44:56 +00:00
cls%seawood.org 8bd122b3d5 Resurrect REQUIRES so that we have some sort of means to track intermodule dependencies. Bug #59454 r=blizzard@mozilla.org 2000-11-20 07:16:06 +00:00
pavel%gingerall.cz 3f402edbe2 - script compilation support (JS::CompileScript) 
- compiled script may be executed (JS::exec)
- memory leak fixed
  (JSVALtoSV returning objects; magic has to to a mortal value)
 2000-11-19 14:40:33 +00:00
beard%netscape.com 24a6cc787c CW Pro 6 compatibility. 2000-11-18 03:14:07 +00:00
nboyd%atg.com fb39f788bc timestamping file 2000-11-18 02:24:06 +00:00
nboyd%atg.com 387cf7985c Add support in Rhino for BSF without the bsfengines.jar. 2000-11-18 02:14:46 +00:00
danm%netscape.com b9887b3ac3 removing glowcode directives. won't affect you unless you use an obsolete version of glowcode. 2000-11-18 02:13:42 +00:00
nboyd%atg.com 2e9a2bd01f Fix 60184, and add SwingApplication.js example, with change 
to shell to prevent early exit.
 2000-11-17 13:05:46 +00:00
rogerl%netscape.com 0f1b1996aa Added lexutils & exception... to default build. 2000-11-16 23:58:53 +00:00
rogerl%netscape.com 3923459330 Example from Waldemar's spec. 2000-11-16 23:50:37 +00:00
rogerl%netscape.com c65d7b8376 Added BindThis argument, removed 'this' arg from Call. 2000-11-16 23:49:29 +00:00
rogerl%netscape.com 8ccb885553 Added notion of binding a 'this' to a method to construct a mini-closure 
for call semantics. Fixed MSVC warnings from lexutils. Added BindThis
instructionand removed 'this' from Call instruction (is now extracted
from target argument).
 2000-11-16 23:48:42 +00:00
mccabe%netscape.com 92e9039cb0 Not part of the Mozilla build. Update tool paths for server build. 2000-11-16 00:03:20 +00:00
mccabe%netscape.com dea57eb30f Not part of the Mozilla build. Add js and liveconnect config files for OSFV5.0 2000-11-15 23:57:49 +00:00
nboyd%atg.com 09771a8b39 Subject: 
[Rhino] importPackage() when not in Rhino shell?
        Date:
             Tue, 14 Nov 2000 09:37:39 -0000
       From:
             "Benjamin Geer" <geerb@midas-kapiti.com>
 Organization:
             Another Netscape Collabra Server User
 Newsgroups:
             netscape.public.mozilla.jseng




The importPackage() and importClass() functions provided by the Rhino shell
seems as if they would be very generally useful.  Unless I've missed
something, they don't seem to be available to scripts compiled using the
JavaScript compiler, or to scripts that are run using Script.exec().  Is
there any chance these functions could be made available for all scripts to
use?  This would save a lot of typing; one could then always write a = new
Foo() instead of a = new Package.com.baz.bar.foo.Foo().

--Benjamin Geer
 2000-11-14 20:10:42 +00:00
nboyd%atg.com 72a244c2da Update javadoc. 2000-11-14 20:08:53 +00:00
nboyd%atg.com 8db6eec108 Fix bug: 
slobo@espial.com wrote:
>
> Hello Mike
>
> In the following test case, tester returns "undefined cat" in Rhino
> while in NN it returns "meow cat".
>
> Thanks
>
> Steven
>
> /////////////////////////////////////////////////////////////////////
> function tester(nest){
>     var nest = nest+" cat";
>     alert(nest);
>     // nest now contains the value undefined.
> }
>
> tester("meow");
 2000-11-14 13:30:55 +00:00
rginda%netscape.com 806cf8e188 refactored lexer bits from icodeasm into lexutils 
refactored exceptions from icodeasm into exception*
fixed LeadingCap method names to be interCapped
added string8 and string16 typedefs (as opposed to string vs String)
 2000-11-14 03:11:54 +00:00
nboyd%atg.com 58b79a147b Subject: 
More changes to support non NativeJavaObject wrappers
   Date:
        Fri, 03 Nov 2000 17:56:38 +0100
   From:
        Igor Bukanov <igor@icesoft.no>
     To:
        nboyd@atg.com




Hi, Norris,

In post 1.5 rhino one can introduce own wrappers for arbitrary Java
objects. But I think to fully support this
org.mozilla.javascript.ScriptRuntime should be changes as well: its eq
and shallow_eq contain references to NativeJavaObject, this should be
replaced at least to Wrapper (see the atached patch). Even better
solution would be to add to WrapHandler methods to compare wrappers: I
can send a patch for that as well.

There is a small usability problem as well: if
org.mozilla.javascript.JavaMembers would be public I do not need to copy
it to a package with non NativeJavaObject.java wrapper.

Regards, Igor
 2000-11-13 22:10:35 +00:00
jeff.dyer%compilercompany.com 843e30ec1a Fix the first startup test. 2000-11-08 21:05:58 +00:00
mccabe%netscape.com 1691baff22 Fix to 58946. 
Make try { ... } catch(exn) { return exn } work by ensuring that the return value (exn) is maintained on the stack as we pop off scopes to return from the try/catch/finally.  The newly added JSOP_SWAP opcode helps us bubble.

This fixes a regression uncovered by the fix to 56716.

(I've noticed that this causes *depend* builds of the standalone JS shell to crash on this construct, but I've tested in the Mozilla build, and the dependencies seem to solve the problem there.)

r=brendan.mozilla.org
sr=jband@netscape.com
 2000-11-08 20:39:43 +00:00
beard%netscape.com 06b45baadc Replaced occurrences of $$, $1, $2 with result, first, second. CodeWarrior doesn't like $ characters in identifiers. 2000-11-08 19:50:39 +00:00
pschwartau%netscape.com 8171a0962f The const keyword is non-ECMA and not currently implemented in Rhino 2000-11-08 01:10:43 +00:00
pavel%gingerall.cz 7ac6f89e90 - jsperl.c compiled with js programm, not with libjs 
- better JS value tie methods
- memory leaks fixed
- support of digging native js error from exception
 2000-11-07 15:24:49 +00:00
nboyd%atg.com 0f96b20b62 Clean up debug APIs 2000-11-07 03:20:56 +00:00
rogerl%netscape.com 0a6215f169 Fixed warnings & bogus qualifier. 2000-11-07 01:09:45 +00:00
rogerl%netscape.com 5a3c3bbedc Fixed statics and constructors bitrot. 2000-11-07 00:59:27 +00:00
rogerl%netscape.com 9468068269 More tests and changed 'verify' to live separately. 2000-11-07 00:58:42 +00:00
rogerl%netscape.com d56d1d28c0 Testcase for adding setter to non-final field. 2000-11-06 23:00:00 +00:00
rogerl%netscape.com 7f2318fcc5 Fixed compare bugs. Report results from js2 test cases. Rewrote generic 
equal to match ECMA 1.5
 2000-11-06 22:57:10 +00:00
waldemar%netscape.com c4904771e2 Defined UseIncludesExcludes. Removed support for a:🅱️:c. 2000-11-04 10:47:47 +00:00
waldemar%netscape.com a16abdbf7e Added 'include' and 'exclude' non-reserved words 2000-11-04 10:46:24 +00:00
mccabe%netscape.com 8a87f76dc2 Fix to 58008. Add missing ECMA-specified fuctions Date.prototype.toDateString, toTimeString, toLocaleDateString, toLocaleTimeString. Commoned some code to toLocale_helper and date_format. 
Also make several formerly class static variables lazily initialized.
 2000-11-03 21:28:49 +00:00
rogerl%netscape.com 18a5e3a269 More tests 2000-11-03 02:42:23 +00:00
rogerl%netscape.com 7b8898dcf1 Fixed adding setter to virtual inherited field. 2000-11-03 02:42:01 +00:00
pschwartau%netscape.com 5fb5604727 initial add 2000-11-03 02:27:44 +00:00
jeff.dyer%compilercompany.com 94f99233df Don't care if required named come after optional named parameters. 2000-11-03 01:44:30 +00:00
rogerl%netscape.com 1851703e79 Fixed bogus rest reference, added named tests. 2000-11-02 19:40:42 +00:00
mccabe%netscape.com 1cb16802f8 Fix previous fix to conform to brendan-nit. 2000-11-02 03:38:09 +00:00
mccabe%netscape.com 980b771c8d Fix to 58007. Add Date.prototype.toDateString, toTimeString methods, by commoning existing date_format helper function with toString. ECMA-262 ed. 3 conformance. 
r=brendan@mozilla.org
sr=jband@mozilla.org
 2000-11-01 23:57:37 +00:00
nboyd%atg.com 35f28776c5 Upgrade to Ant 1.2 and remove deprecated tasks. 2000-11-01 18:43:55 +00:00
rogerl%netscape.com e74eea4764 Fixed signed/unsigned whinage. 2000-11-01 02:04:04 +00:00
rogerl%netscape.com 56310a15c2 Simple call tests. 2000-11-01 01:55:46 +00:00
rogerl%netscape.com 2edfa7302d Fixed rest parameter index handling. 2000-11-01 01:55:13 +00:00
rogerl%netscape.com 375bd5d07d Fixed signed/unsigned whine. 2000-11-01 01:48:15 +00:00
rogerl%netscape.com 2e5785dfe5 Fixed JSArray for 1.5 semantics stuff. Changed readEvalFile to stop 
doing line at a time nonsense. Fixed ExprNode::Index cases to use
InvokeExprNode instead of Binary.
 2000-11-01 01:39:43 +00:00
mccabe%netscape.com 82a15a1687 Fix sense of don't-fail-on-warning check propagated from js.c. 
Thanks to John Bandhauer for the catch.

Not part of Mozilla.
 2000-10-30 22:42:48 +00:00
nboyd%atg.com c109da414a Fix bug 57060 for/in loop enumerates array elements as numbers, not strings 2000-10-30 19:46:02 +00:00
nboyd%atg.com fa72e89acf Fix bugs: 
56318  function literals with names don't work right
57045  negative integers as object properties: weird behavior
58479  functions defined within conditional phrases are always crea
 2000-10-30 19:34:11 +00:00
shaver%mozilla.org f188006b0a #56716, run finally{} before return from try{}/catch{}, r=mccabe, sr=brendan 2000-10-29 18:54:58 +00:00
brendan%mozilla.org 9d7b76c742 Use 0 - i rather than -i to avoid unsigned warning on Windows. 2000-10-29 18:08:40 +00:00
nboyd%atg.com d223b095ba Fix formatting 2000-10-29 18:07:26 +00:00
brendan%mozilla.org cb814a824f Handle negative-int-fits-in-jsval string literals as property names equivalent to their int conversions (57043, r=shaver, sr=jband). 2000-10-29 01:30:16 +00:00
waldemar%netscape.com 139931c770 Allowed multiple units in a unit expression 2000-10-28 05:55:47 +00:00
waldemar%netscape.com b65d4f2f46 Renamed OrdinaryInitialIdentifierCharacter and OrdinaryContinuingIdentifierCharacter to InitialIdentifierCharacter and ContinuingIdentifierCharacter 2000-10-28 05:55:19 +00:00
waldemar%netscape.com c8e5c8c645 Initial version 2000-10-28 05:54:44 +00:00
waldemar%netscape.com e25ff660ba Added JS20/Units 2000-10-28 05:53:23 +00:00
rogerl%netscape.com c93f2b6295 Fixed gcc whine about char16. 2000-10-28 00:07:36 +00:00
rogerl%netscape.com e89c1b85f2 Removed unused parameter. 2000-10-28 00:01:58 +00:00
rogerl%netscape.com 2c070bbf9e Named parameter work. 2000-10-27 23:55:31 +00:00
nboyd%atg.com 0e8b09b0b6 Add javadoc. 2000-10-27 20:17:56 +00:00
nboyd%atg.com 466695f1aa Fix tree and icode printing. 2000-10-27 20:17:19 +00:00
mccabe%netscape.com 133caad64a Changes to update standalone JavaScript and Liveconnect library builds to the New Era of NSPR 4 and JDK 1.2. 
Fixes for OSF; they they assume the existence of /share/builds/components/[jdk|nspr]/SOME_VERSION/etc/etc.  Sorry, external folks!

Not part of the Mozilla build.
 2000-10-27 01:10:09 +00:00
mccabe%netscape.com a516acc1d9 Changes to update standalone JavaScript and Liveconnect library builds to the New Era of NSPR 4 and JDK 1.2. 
Fixes for Linux and SunOS; they assume the existence of /share/builds/components/[jdk|nspr]/SOME_VERSION/etc/etc.  Sorry, external folks!

Not part of the Mozilla build.
 2000-10-27 00:08:31 +00:00
mccabe%netscape.com b75a03d416 Fix standalone liveconnect build to not complain about missing fdlibm when asked to clean or clobber. 
Not part of the Mozilla build.
 2000-10-26 23:48:16 +00:00
mccabe%netscape.com 0773be3ba0 Make commandline shells for javascript and xpconnect-enabled javascript not return failure error codes when a warning occurs. 
The failure codes were confusing our test drivers.

Not part of the Mozilla product.
 2000-10-26 23:45:43 +00:00
jeff.dyer%compilercompany.com 969051855b Added a couple missing redesignates. 2000-10-26 19:51:31 +00:00
beard%netscape.com a8b33eb1d5 adding nodefactory.h to dependency mix 2000-10-26 03:43:41 +00:00
beard%netscape.com 9adcc15ce8 fixing confusing code for linux, static Arena() -> getArena() 2000-10-26 03:40:51 +00:00
mccabe%netscape.com ea452afb28 Add magic emacs modeline comment to make it display in Makefile mode. 
Not part of the Mozilla build.
 2000-10-26 02:59:03 +00:00
mccabe%netscape.com 8502dbb654 Add magic emacs modeline comment to make it display in Makefile mode. 
Not part of the Mozilla build.
 2000-10-26 02:13:07 +00:00
jeff.dyer%compilercompany.com 5e52ae4dac Added to FunctionDefinition a pointer that marks the position in the 
parameter list that the named parameter marker occurs. The names at or
after that position may or may not have aliases associated with them.
 2000-10-25 12:46:23 +00:00
brendan%mozilla.org 1d811ae9b8 Fix obj.toSource to get attributes using the object in which id was found (56482, r=jband, sr=shaver). 2000-10-25 07:00:31 +00:00
brendan%mozilla.org 30d3951dc9 Home sp before calling js_ValueToStringAtom and js_ValueToFunction (bug 57826, r=mccabe, sr=jband). 2000-10-25 04:04:29 +00:00
jeff.dyer%compilercompany.com 22833e1539 Fix assertions that I missed in partial debug mode. Also fixed problem with 
optional parameter after the named parameter marker.
 2000-10-23 22:23:32 +00:00
jeff.dyer%compilercompany.com 73411c491c Parsing named parameters and arguments according to the Sep-23rd revision 
of the grammar. WARNING: struct VariableBinding has grown. Recompile all
after getting this revision.
 2000-10-23 19:39:55 +00:00
nboyd%atg.com cbdf39c92e Subject: 
[Rhino] Optimization for OptRuntime.thisGet
        Date:
             Mon, 23 Oct 2000 17:50:53 +0200
       From:
             Hannes Wallnoefer <hannes@helma.at>
 Organization:
             Another Netscape Collabra Server User
 Newsgroups:
             netscape.public.mozilla.jseng




I found a little oddity in
org.mozilla.javascript.optimizer.OptRuntime.thisGet().

get() is called twice on thisObj, once right at the beginning, and once
when starting to walk down the prototype chain. Below is what I think
this should look like - the prototype walk now begins with thisObj's
prototype, if it exists.

Also, (thisObj == null) was checked only after thisObj.get() was called,
so I moved that up in front.

Hannes
PS: I just made the changes in the news msg editor, so there may be
stupid mistakes.
 2000-10-23 17:48:27 +00:00
nboyd%atg.com b1266792b7 Subject: 
updated Global, Main and ImporterTopLevel
   Date:
        Mon, 23 Oct 2000 14:37:45 +0100
   From:
        Matthias Radestock <matthias@lshift.net>
     To:
        nboyd@atg.com




Norris,

I've made some more changes to shell.Main and shell.Global in order to
reduce their mutual dependency, enable "quit" and get "load" to operate
in the local scope.

see attachments for updated .diffs.


Matthias.
 2000-10-23 15:55:42 +00:00
beard%netscape.com 3c21b0ca58 Fixed typo in package target. 2000-10-23 15:44:05 +00:00
nboyd%atg.com bcfdcd3e36 Fix 56883. 2000-10-23 14:31:05 +00:00
nboyd%atg.com 210c869cd5 Patch from beard@netscape.com to fix 56879. 2000-10-23 14:30:40 +00:00
brendan%mozilla.org 7527c8a330 Comment change only: document TOK_TRY and TOK_CATCH nodes. 2000-10-22 22:47:40 +00:00
beard%netscape.com 87e4c0401b Fixed access paths, link errors. [Not part of build.] 2000-10-21 07:29:53 +00:00
beard%netscape.com e77b33af7d Added static library targets, JSRef.o and JSRefDebug.o. [Not part of build.] 2000-10-21 06:46:21 +00:00
beard%netscape.com a0fa5086d0 LiveConnect support for Mac. [Not part of standard build.] 2000-10-21 06:44:54 +00:00
beard%netscape.com 5fa536e789 fixed some spacing problems, Makefiles use hard tabs! 2000-10-21 00:57:54 +00:00
beard%netscape.com 16f4fdae81 update depenencies for linux 2000-10-21 00:55:59 +00:00
beard%netscape.com 8f28eb5769 added xmlparser.cpp. 2000-10-21 00:51:32 +00:00
beard%netscape.com ee94aef90f fixing unsigned warnings, a leetle error checking. 2000-10-21 00:50:44 +00:00
beard%netscape.com 7f9bf0e529 fixing unsigned/signed conversion warnings. 2000-10-21 00:43:35 +00:00
rogerl%netscape.com acc5d653f4 Fixed signed/unsigned warning. 2000-10-21 00:38:59 +00:00
rogerl%netscape.com ee98657443 Changed ArgumentList to ArgumentList*. (Still need to handle the 
destructor though).
 2000-10-21 00:31:22 +00:00
rogerl%netscape.com 8e8a28606a Changed ArgumentList to ArgumentList*. Changed ICodeGenerator to take a 
Context instead of world & global pair. Connected to ICodeAssembler from
XMLParser.
 2000-10-21 00:30:23 +00:00
rginda%netscape.com 25bccbbb49 fixups for lame windows parser, and multiple parse per instance issues. 2000-10-20 22:18:52 +00:00
rginda%netscape.com 2457927e68 lalala, things run, lalala. (well, kinda.) 2000-10-20 10:33:57 +00:00
rginda%netscape.com adcd95d4ac more work on producing real live instructionstreams 2000-10-20 03:33:37 +00:00
rginda%netscape.com 143404294d starting to turn the icode StatementNode vector into actual Instructions... 2000-10-20 01:13:03 +00:00
rogerl%netscape.com a9f2c3b79c changes to XMLLexer etc. 2000-10-19 23:07:58 +00:00
rogerl%netscape.com 5b3d409795 Adding xmlparser stuff 2000-10-19 22:34:23 +00:00
brendan%mozilla.org c9423a9980 - Treat 08, 09, 078, 079, etc. as decimal, but warn that they're not legal 
ECMA-262 octal literals.  Old code would split 08 into 0 and 8 if JS1.2 or
  other non-ECMA version, and always split 078 into 07 and 8, resulting in
  missing ; syntax errors.

- Fix CheckFinalReturn to be aware of JS_HAS_EXCEPTIONS, finally (sic).  Lots
  of help from jag (Peter Annema, disttsc@bart.nl), thank him.

Both changes got lumped under bug 49233, and are r=jband, sr=shaver.
 2000-10-19 19:21:53 +00:00
jeff.dyer%compilercompany.com feccfe330f b=53849, r=mstoltz@netscape.com,edburns@acm.org, a=brendan@mozilla.org. 
This fix makes Liveconnect smarter about getting a security principal
when verifying that call from an applet to JS should be allowed.
 2000-10-19 11:51:22 +00:00
beard%netscape.com 34165f9eb4 More const String& happiness. 2000-10-19 03:37:02 +00:00
beard%netscape.com 91533bfcd5 const String& to soothe g++ 2000-10-19 03:33:51 +00:00
beard%netscape.com 592a474b0a Fixing linux breakage, incorrect constructor name. 2000-10-19 03:32:23 +00:00
rginda%netscape.com dbbc7dcf0e made (most) parse*operand functions do the right thing... the rest choke for now. 2000-10-19 00:03:12 +00:00
rogerl%netscape.com 5a461f3c89 More warning whining. 2000-10-19 00:02:37 +00:00
rogerl%netscape.com ab0a5b79b7 Quiet warnings 2000-10-18 23:55:47 +00:00
rogerl%netscape.com 797d199480 Added icodeasm & xmlparser to build 2000-10-18 23:46:39 +00:00
rogerl%netscape.com e30ce8373b Added XMLparsing to load a class. Fixed handling of forward references to 
class methods/fields.
 2000-10-18 23:37:44 +00:00
brendan%mozilla.org 0b83f3aa96 Don't nuke the stackPool in JS_GC if it's in use by cx->stackHeaders or cx->fp (or other future users; 57096, r=shaver, a=jband). 2000-10-18 22:21:46 +00:00
rginda%netscape.com ff3db92711 making argument name optional in argumetlists 2000-10-18 22:01:17 +00:00
rginda%netscape.com 1cb3536755 waugh, "..." isn't supported on windows. Added full icode map generated by the new perl script. Added a comment character to the assembler, and required newlines after instructions. 2000-10-18 20:44:14 +00:00
rginda%netscape.com 0f8daa3979 adding genmap.pl to generate the icode map 2000-10-18 20:41:03 +00:00
brendan%mozilla.org 66e3327295 Fix cx-for-acx typo-blunder (57070, r=mccabe, a=jband). 2000-10-18 16:00:39 +00:00
rginda%netscape.com 7ba55fdebd fixed testcase and icodemap size 2000-10-18 04:33:54 +00:00
rginda%netscape.com df1d06fc40 whee! the parse "tree" minimally "works" 2000-10-18 03:12:42 +00:00
rginda%netscape.com 335a61b469 more icodeasm work... parse "tree" is finally testable... it might even work. 2000-10-18 01:02:13 +00:00
rginda%netscape.com 8685390f44 working in labels 2000-10-17 08:02:16 +00:00
rginda%netscape.com 9b5fdfd8c2 icode assembler changes 2000-10-17 00:41:54 +00:00
jband%netscape.com aee1c25d21 fix crasher bug 56843 in case of if('wrappedJSObject' in foo). Better solution for bigger problem in the works. a=brendan@mozilla.org r=shaver@mozilla.org 2000-10-16 22:34:26 +00:00
jband%netscape.com deded45781 remove JS gc roots on wrappedjs objects during xpconnect shutdown to avoid needless rooting and noise in root leak logs. bug 55426. a=brendan@mozilla.org r=mccabe@netscape.com 2000-10-15 05:42:25 +00:00
jband%netscape.com f3f7e3cf67 only emit leaked root warnings when the JSRuntime is being deleted. This changes around the fix of bug 55117 and cleans up the incorrect leaked root warnings in bug 55426. a=brendan@mozilla.org r=mccabe@netscape.com 2000-10-15 00:20:48 +00:00
rginda%netscape.com 534a67dfea guess what? yep, more icode asm hacking. 2000-10-14 00:06:42 +00:00
rogerl%netscape.com bdb6266783 Bug #46196, a=brendan, r=mccabe. Fixed recursive death in out of memory 
situation.
 2000-10-13 22:11:52 +00:00
brendan%mozilla.org c7d1950d0e Checking in for rogerl (53614, r=mccabe, a=me). 2000-10-13 01:58:40 +00:00
rginda%netscape.com 36bdf14e61 adding some comments 2000-10-12 23:59:25 +00:00
rginda%netscape.com 0d069aae17 icode assembler still in progress 2000-10-12 23:58:51 +00:00
rginda%netscape.com c43c176e8d move # to first column for Very Lame Compilers. 2000-10-12 23:58:30 +00:00
nboyd%atg.com 1d9b17f17b Fix bug 56185. 2000-10-12 16:10:55 +00:00
nboyd%atg.com 908350121b Fix bug 56158 2000-10-12 14:59:19 +00:00
rginda%netscape.com 7d14c260a9 more icodeasm hacks 2000-10-12 07:55:13 +00:00
beard%netscape.com ee2e996d8a Outer class loader support, using ClassManager. 2000-10-12 06:30:29 +00:00
rginda%netscape.com b0f3d95321 more hacking (flailing, maybe) on the icode assembler 2000-10-12 01:32:01 +00:00
rginda%netscape.com f38bdbdbf3 initial checkin of simple query script for the icode hash 2000-10-12 01:31:33 +00:00
rogerl%netscape.com 22b5bf23a9 Bug #55114, a=brendan, r=mccabe. Fixed failure in enumerating Java string 
objects.
 2000-10-12 01:23:13 +00:00
rogerl%netscape.com d533f7e957 Bug #53614, a=brendan,r=mccabe. Fixed crash for calling RegExp literal. 2000-10-12 00:18:40 +00:00
rogerl%netscape.com 3379e625da Bug #52479 , a=brendan,r=mccabe. Fixed crash on bad \uXXXX in input. 2000-10-11 22:34:40 +00:00
rginda%netscape.com 1863037f91 initial checkin of icode assembler 2000-10-11 02:44:14 +00:00
rginda%netscape.com e1e32378c7 seperate icode hash into module 2000-10-11 02:42:39 +00:00
waldemar%netscape.com 7442e0026e Added s flag 2000-10-11 02:28:13 +00:00
beard%netscape.com 5a1f437267 Map all exceptions to just return null. 2000-10-10 18:14:20 +00:00
nboyd%atg.com 2a806a1b13 More javadoc. 2000-10-10 14:58:47 +00:00
waldemar%netscape.com 8808f7286f Changed classof x to x.class. Removed eval. Made include be a non-reserved word. Changed syntax of import to allow an automatic use of selected namespaces. 2000-10-10 02:49:34 +00:00
waldemar%netscape.com 6f4582329f Removed "eval" and "include" keywords 2000-10-10 02:48:17 +00:00
rogerl%netscape.com 8141a721b4 Mods for getter/setter methods. 2000-10-09 22:21:26 +00:00
rogerl%netscape.com 12bba253e7 Bug #54408, r=mccabe, a=brendan. LiveConnect top-level property names 
predefined readonly.
 2000-10-09 21:39:22 +00:00
beard%netscape.com 848d0b5817 Added ClassManager.java. 2000-10-09 18:16:06 +00:00
beard%netscape.com cdae200b86 Manages interactions with outer class loaders. 2000-10-09 17:43:51 +00:00
brendan%mozilla.org 3525efa958 Remove not-found property caching (55624, r=shaver a=jband). 2000-10-09 14:38:21 +00:00
nboyd%atg.com 297522e053 Add support for new Ant-based buildfile. 2000-10-06 16:26:17 +00:00
nboyd%atg.com 1623d01e8f Remove reference that javadoc couldn't find 2000-10-06 16:25:36 +00:00
nboyd%atg.com 4e15c42226 Reformat code. 2000-10-06 16:25:14 +00:00
mccabe%netscape.com 2f0d7b87e1 Fix to 55117. Warn on DEBUG of any GC roots remaining at engine shutdown. 
r=rogerl@netscape.com
a=brendan@mozilla.org
 2000-10-05 21:55:38 +00:00
jband%netscape.com cecbb83988 fix bug 55139. fix broken JSObject parenting of wrapped natives for JS components. r=shaver@mozilla.org a=brendan@mozilla.org 2000-10-05 04:15:43 +00:00
jband%netscape.com 46893817df part of bug 55139. Add debug only diagnostic code for dump JSObjects. r=shaver@mozilla.org a=brendan@mozilla.org 2000-10-05 04:13:53 +00:00
jband%netscape.com f803edb01b fix bug 55151. This fixes a trunk only regression caused by me in a previous checkin. I missed that an 'if' block contained both a report of an error (that we might want to skip) and also a few lines of necessary error handling code. r=mccabe@netscape.com a=brendan@mozilla.org 2000-10-05 04:13:48 +00:00
brendan%mozilla.org 00c4402750 Fix old bug that stored iter_state even after OBJ_ENUMERATE error (55201, r=rogerl, a=jband). 2000-10-05 00:43:17 +00:00
nboyd%atg.com 1114de0a19 Allow superclass to appear anywhere in the list of classes in the JavaAdapter argument list. 2000-10-03 18:34:02 +00:00
beard%netscape.com 646b2ea55c avoid an infinite loop on EOF. 2000-10-03 06:23:11 +00:00
nboyd%atg.com 2307ef527d Avoid creating multiple wrappers. 2000-10-02 20:46:08 +00:00
nboyd%atg.com 10bd56f484 Do cheap test first. 2000-10-02 17:10:00 +00:00
jband%netscape.com e80b9b250d Mostly submitted patches... 
- Fix bug 54264 from Jon Smirl <jonsmirl@mediaone.net>
  Do cleanup of thread local storage on main thread.

- Fix bug 54275 from Jon Smirl <jonsmirl@mediaone.net>
  Release components in shell before shutting down xpcom

- Fix bug 54310 from Jon Smirl <jonsmirl@mediaone.net>
  Call JS_DestroyScript in xpcshell and js.c

- Fix bug 54352 from Jon Smirl <jonsmirl@mediaone.net>
  Cleanup what static data we can in xpclog.

- Initial fix of bug 54473
  Don't report warnings as errors in wrapped JS calls.

- Fix bug 54462 from Mark Hammond <MarkH@ActiveState.com>
  Fix jband's stupid use of uint8 for method indexes.

- Use environment rather than prefs for #ifdef'd debug options

- Don't report NS_ERROR_FACTORY_REGISTER_AGAIN as an error.

r=mccabe@netscape.com a=jband@netscape.com
 2000-09-30 08:16:09 +00:00
mccabe%netscape.com 5911e488e0 Propagate fix to 41872 from the standalone JS shell to the standalone XPConnect-enabled JS shell. Make warnings on by default in the standalone JS shell, and introduce the -W option to suppress warnings if needed. Retained the -w option (enable warnings) in case somebody was using it. 
This file doesn't go into the Mozilla distribution.

r=brendan,r=jband
 2000-09-30 01:42:29 +00:00
mccabe%netscape.com 40e861a13e Fix to 41872. Make warnings on by default in the standalone JS shell, and introduce the -W option to suppress warnings if needed. Retained the -w option (enable warnings) in case somebody was using it. 
This file is not part of the Mozilla build.

r=brendan
 2000-09-30 00:24:22 +00:00
nboyd%atg.com 42b2d478cd Allow null returns from SecuritySupport methods if someone only 
wants to implement LiveConnect filtering
 2000-09-29 13:49:47 +00:00
brendan%mozilla.org 4ed9d79e0a Fix bungled GC_KEEP_ATOMS logic (51954, r=mccabe, a=jband). 2000-09-29 00:18:15 +00:00
brendan%mozilla.org 8d730c70da Don't leave fp->sp above fp->spbase when returning, js_Invoke may GC on the way out (53123, r=mccabe, a=jband). 2000-09-29 00:11:49 +00:00
cls%seawood.org 799d3ef591 Do not use c++ comments in c file. Breaks WS5.0 compiler. 2000-09-28 08:08:02 +00:00
mccabe%netscape.com ba4d6a252b Fix to 54307. Patch courtesy Jon Smirl <jon@mediaone.net>. 
Clean up several locks that jsdtoa uses, when JS_THREADSAFE.

a=brendan
r=mccabe
 2000-09-28 05:14:32 +00:00
mccabe%netscape.com 897e928322 Remove unused function prototype. 
r=jband
 2000-09-28 02:47:40 +00:00
mccabe%netscape.com 3b57292fa9 Fix to 54305, free array holding fat locks. Patch courtesy Jon Smirl <jonsmirl@mediaone.net>. 
a=brendan
r=mccabe
 2000-09-27 21:52:27 +00:00
jband%netscape.com b416a15d43 NOT PART OF BUILD. fixing tools bustage 2000-09-26 10:23:23 +00:00
pschwartau%netscape.com 0aefd9f03b Adding new test case for scope check and creating js1_5 test suite. 2000-09-25 21:24:09 +00:00
nboyd%atg.com 7f2239a2c4 Miscellaneous formatting changes. 2000-09-25 14:07:45 +00:00
nboyd%atg.com 142af94b8f Drop unneeded references. 2000-09-25 14:06:47 +00:00
nboyd%atg.com adc731c1e6 Fix enum.js regression. 2000-09-25 14:04:20 +00:00
waldemar%netscape.com eded78869f Added void and classof operators 2000-09-24 06:08:53 +00:00
waldemar%netscape.com 0e94e874c9 Made implements clauses js2-only 2000-09-23 08:15:06 +00:00
brendan%mozilla.org 5c322995b7 Fix two distinct bugs with similar symptoms (53123, r=law, a=jband). 2000-09-23 01:23:53 +00:00
jeff.dyer%compilercompany.com d5d760afb0 #45692:r=edburns,a=brendan. Followup fixes for oji liveconnect security. 2000-09-22 13:48:20 +00:00
waldemar%netscape.com e9fac36551 Major revisions for Sep 18 and 21 language changes 2000-09-22 08:16:22 +00:00
waldemar%netscape.com baf53149b3 Removed "constructor" keyword 2000-09-22 08:14:56 +00:00
jband%netscape.com 418b39b154 - Use the thread context service rather than our own JSContext. This makes 
us less unthreadsafe. Use THREADSAFE nsISupports impl macro. bug 52936
- Add JS_{Begin,End}Request. bug 39373
- Call xpc->InitClass on each global - not just the superglobal. bug 52591
- Remove some gotos using auto classes for cleanup.
- Converted WITH_SERVICE calls to do_GetService.
- Consistent placement of contractID strings.
a=shaver@mozilla.org
 2000-09-21 04:30:32 +00:00
jband%netscape.com 3dcad1b0a6 rest of the fix for bug 53268. a=brendan@mozilla.org 2000-09-21 03:56:42 +00:00
brendan%mozilla.org ccf7a9fae9 Fix 53268, r=jband. 2000-09-21 01:37:02 +00:00
rogerl%netscape.com 668c8ce522 Cleanup warning fix. 2000-09-19 20:56:49 +00:00
rogerl%netscape.com 9a22c0ee19 Fixed warnings for gcc build. 2000-09-19 20:35:05 +00:00
beard%netscape.com a160285334 Fixed startup/terminate routines. 2000-09-19 05:38:23 +00:00
beard%netscape.com 49e2480b9c [Not part of build] Really increase default heap size, to 8192. 2000-09-19 02:50:00 +00:00
beard%netscape.com 89a295bdc4 Bumped up memory/stack. 2000-09-19 02:34:41 +00:00
brendan%mozilla.org 2861dff1f3 Add strict warning for undefined property reference (foo.bar where bar names no object in foo or its prototypes; r=jband). 2000-09-19 02:24:11 +00:00
brendan%mozilla.org 490bf8236e Fix JS_BeginRequest to avoid self-deadlock if called from a finalizer (52808, r=jband). 2000-09-19 02:18:17 +00:00
jband%netscape.com ef7bb615cc fix shutdown crasher bug 52940. We can't leave the thread context stack service thinking that the context we are about to delete is still valid. a,r=brendan@mozilla.org 2000-09-19 01:09:48 +00:00
beard%netscape.com b35d6552f0 Create SIOUX console with menus, etc. r=jband, a=brendan 2000-09-18 23:49:46 +00:00
brendan%mozilla.org de5d6d74b9 Censor all Call objects, per ECMA-262 (r=rogerl, 53037). 2000-09-18 23:35:22 +00:00
brendan%mozilla.org 1f03cb71d4 - Fix encodeURI, decodeURI, etc. common subroutines to avoid realloc'ing char 
by char, by using a larger chunk size (64 chars) for linear growth.  Also got
  rid of ASCII-oriented add_bytes subroutine and related sprintf usage.
- Avoid reloading loop invariant str->chars all the time in encode and decode.
- Avoid creating garbage strings for unescaped and reserved character sets, by
  using statically initialized jschar array constants.
- Expand tabs, clean up 80th column violations, use prevailing style, etc.
 2000-09-18 22:17:30 +00:00
nboyd%atg.com b8f186828a Expand javadoc. 2000-09-18 18:38:02 +00:00
nboyd%atg.com 7457687f89 Remove old-style debug codegen. 2000-09-18 17:57:02 +00:00
nboyd%atg.com a6150b9fff Better error message for common error. 2000-09-18 17:28:59 +00:00
beard%netscape.com dee42157a2 Make XPC shell compile on XP_MAC. r=jband, a=brendan 2000-09-17 15:34:39 +00:00
jband%netscape.com 51e5128d17 NOT PART OF BUILD. Fix xpconenct test uses of progids missed in the big progid to contractid renaming 2000-09-16 22:17:23 +00:00
brendan%mozilla.org 3d0b204150 - Back out part of patch from tlundeen@webcrossing.com, my fault for taking it 
in part (the entire patch made JSContexts ref-counted, but that is not an API
  compatible change, and it doesn't help clean up at JS_Finish time if the API
  user leaks JSContext refs anyway).  52835, r=jband.
- First part of 64-bit portability fix for 52792, r=jnance.  More work needed.
- Fix bogus assert and minimization in js_AllocStack, too.
 2000-09-16 22:17:22 +00:00
jband%netscape.com d50e6f9d65 NOT PART OF BUILD. Fix xpconenct test uses of progids missed in the big progid to contractid renaming 2000-09-16 21:47:32 +00:00
beard%netscape.com d8c9e5842f [Not part of build] Project to build XPConnect shell testbed. 2000-09-16 02:43:29 +00:00
jband%netscape.com 6eabb6acbd fix bug 39373. make xpconnect calls into JS within a request for safety. a,r=brendan@mozilla.org 2000-09-15 06:59:38 +00:00
jband%netscape.com feed45599c fix bug 52579. Avoid infinite recursion dialog when JS is run by xpconnect on DOM context. a,r=brendan@mozilla.org 2000-09-15 06:17:08 +00:00
brendan%mozilla.org 84aaf06fe8 Better patch for bug and bogus assertion found by rogerc@geocast.com. 2000-09-15 06:05:28 +00:00
rayw%netscape.com d39d9e6d59 Bug fix for 52648 -- remaining stash of progids which I missed on 
the mega-change yesterday.  This will fix a few more issues.

r=warren
 2000-09-14 23:20:49 +00:00
rogerl%netscape.com b0bb9eeddc Bug #46429. r=rginda. Fixed bug in nested quantifiers. 2000-09-14 22:39:21 +00:00
brendan%mozilla.org ed8364faab Fix bogus new assertion in js_AllocGCThing (r=rogerc@geocast.com). 2000-09-14 19:58:11 +00:00
brendan%mozilla.org da262c552a Fixes to make JS GC truly exact: 
- All jsvals for which JSVAL_IS_GCTHING evaluates to true must contain tagged
pointers into the GC heap -- therefore jsapi.c's JS_DefineConstDoubles cannot
"cheat" by tagging addresses of static jsdoubles to avoid js_NewNumberValue.

- Finalization is now interleaved with the Sweep phase, to avoid allocating
memory for finalization records while sweeping.  Instead, the JSRuntime holds a
preallocated JSGCThing vector (gcFinalVec) that the Sweep phase fills and
flushes via gc_finalize_phase, repeatedly.

This means that finalizers cannot allocate a new GC thing, an incompatible but
plausible change.  js_AllocGCThing asserts and then checks whether it is called
while rt->gcLevel is non-zero, and fails the allocation attempt if so.  But this
fixes bug 38942, where the old sweep-then-finalize with a sweep => malloc
dependency could lead to memory exhaustion.

- Instead of scanning whole stackPool arenas, which led to UMRs (bug 27924) and
sometimes to gross over-scanning that depended on the GC bounds-checking all
thing pointers against its heap, we scan exactly those stack slots in use:
  - arguments reachable from fp->argv;
  - variables reachable from fp->vars;
  - operands now reachable from fp->spbase, bounded above by the lesser of
    fp->sp or fp->spbase + fp->script->depth for an interpreted frame; if the
    latter, fp->sp has advanced logically above the operand budget, in order to
    call a native method, and all unused slots from fp->sp up to depth slots
    above fp->spbase must be set to JSVAL_VOID;
  - stack segments pushed when calling native methods, prefixed by JSStackHeader
    structs and linked from cx->stackSegments through each header.
The stack segment headers help the GC avoid scanning unused portions of the
stack: the generating pc slots running depth slots below fp->spbase, and slots
at the end of an arena that aren't sufficient to satisfy a contiguous allocation
for more args, vars, or operands.

- Exact GC means the stack pointer must remain above live operands until the
interpreter is done with them, so jsinterp.c got heavily whacked.  Instead of
POPs of various kinds followed by a PUSH for binary operators (e.g.), we use
FETCH and STORE macros that index by -1 and -2 from sp, and minimize adjustments
to sp.  When sp is homed to fp->sp, this allows js_DecompileValueGenerator to
find the value reliably, and if possible its generating pc.

- Finally, the O(n**2) growth rate of gc_find_flags has been fixed, using the
scheme sketched in bug 49816 and documented in a new major comment in jsgc.c.
Briefly, by allocating flags and things from one arena, we can align things on
1024-byte "thing page" boundaries, and use JSGCPageInfo headers in each page to
find a given thing's flags in O(1) time.

/be
 2000-09-14 06:14:45 +00:00
rayw%netscape.com 52476868db fix to mal-formed contract id introduced in mega-check-in:x 2000-09-14 04:18:07 +00:00
rayw%netscape.com 6cc70ebd6c Bug 37275, Changing value of all progids, and changing everywhere a progid 
is mentioned to mention a contractid, including in identifiers.

r=warren
 2000-09-13 23:57:52 +00:00
brendan%mozilla.org d3ef2ed32f Fix uninitialized var bug found by one warning, silence another that's a can't-happen (r=jwbaker@acm.org). 2000-09-12 19:42:01 +00:00