d5c6199643
Restore dynamic indirect eval code. bug 383682, r=brendan/igor a=brendan
2007-10-09 22:42:42 +00:00
ea5dc9a9b6
Bug 395836 - backed out "patch with nits fixed" due to perf regressions
2007-10-09 22:34:04 +00:00
8611afa425
Oops, removing windows line endings I accidentally checked in when fixing bug 399094.
2007-10-09 20:18:44 +00:00
2c1abc9d43
Implement growable array-based storage for native Array objects with dense allocation
...
patterns. Added specialized implementations of some of the Array built-in
functions. One function's performance improved by 755X as a result.
Also streamlined code for creating array literals.
2007-10-09 15:14:05 +00:00
21d2752b00
Bug 399094 - "Wrappers don't deal with non-native objects". Patch by Blake Kaplan <mrbkap@gmail.com>, r+sr+a=jst.
2007-10-09 06:16:55 +00:00
864e1e094e
Wrap the strftime() call with an empty "invalid parameter handler" on Windows.
...
b=395836, patch by Mats Palmgren <mats.palmgren@bredband.net>, r=crowder a=brendan
2007-10-09 04:24:57 +00:00
93c4a36b6a
When cleaning, nuke the full build directory.
2007-10-07 17:21:51 +00:00
361b03f27a
I just noticed that running tests spins up GUI. Shouldn't happen, and this should take care of it.
2007-10-07 17:21:29 +00:00
86d46a513d
Adding a bit of additional diagnostics -- printing to stdout the name of file being run, and the threads carry the file name in their name too
2007-10-06 17:44:54 +00:00
c945e75657
Don't call setters if there is no setter to call. bug 398040, r=brendan sr=jst a=blocking1.9+
2007-10-06 17:14:34 +00:00
161a963b77
JavaScript Test - ignore XPCCrossOriginWrapper on global object, bug 390946, not part of the build
2007-10-05 20:44:42 +00:00
1a6f7c8a44
Fix bug 397680: Need to support Context.observeInstructionCount for compiled
...
mode. Compile in callbacks to a method that accumulates instruction counts
and calls the observer when the threshold is reached.
2007-10-05 19:39:30 +00:00
8062e6607c
JavaScript Test - regression test for bug 394673, by Igor Bukanov, not part of the build
2007-10-05 16:41:54 +00:00
7f9ef24d51
Bug 397773 - mozilla-central has pedantic errors because of extra commas, r=mrbkap+brendan a=brendan
2007-10-05 13:58:51 +00:00
3e07cea0a9
JavaScript Tests - update Sisyphus related shell scripts to use /bin/bash, bug 397921, not part of the build
2007-10-04 16:06:34 +00:00
bf494d32c0
New skips for new tests.
2007-10-04 13:21:22 +00:00
db8d7087d2
JavaScript Tests - update regression test to support Rhino, bug 398450, by Norris Boyd, not part of the build
2007-10-03 20:56:55 +00:00
1acc041933
fix bug 396969: bitwise and shift ops evaluate operands in reverse order, contrary to ECMA-262
...
also avoid generating debugger callbacks for line source for scripts consisting only of comments
2007-10-03 20:32:11 +00:00
006c8a1ed6
Update JS_GetImplementationVersion to 1.8.0 on trunk, bug 380901, ra=brendan
2007-10-03 13:48:01 +00:00
b81e724d63
Bug 397210: JSTokenStream is stored in JSParseContext. r=brendan
2007-10-02 23:10:09 +00:00
8abbdb13fe
Backing out the patch 397210.
2007-10-01 20:46:10 +00:00
9a750cb448
Bug 397210: JSTokenStream is stored in JSParseContext. r=brendan
2007-10-01 20:25:26 +00:00
9c28e24d10
Bug 396758: the system flag is moved from GC flags to JSObject itself. r=brendan
2007-10-01 19:11:41 +00:00
06a54736a1
JavaScript Tests - sync Sisyphus support, bug 398005, not part of the build
2007-10-01 19:10:41 +00:00
d0e5224d53
Propagate exceptions from the evalcx context to the outer context so they can be caught. bug 353737, r+a=brendan
2007-10-01 00:43:30 +00:00
cb5a2ed1cb
Implement an iterator hook for cross origin wrappers to avoid wrongly walking up the prototype chain during enumeration. bug 390947, r+a=brendan sr=jst
2007-10-01 00:31:22 +00:00
af2668558e
evalcx uses JS_BeginRequest/JS_EndRequest. bug 398045, r=mrbkap, not a part of the build.
2007-09-30 16:12:15 +00:00
138e8b44ab
JavaScript Tests - update spidermonkey-extensions-n.tests, bug 398000, not part of the build
2007-09-29 13:55:22 +00:00
e78090b0e8
Patch from Bob Jervis (bjervis@google.com):
...
In reading the LiveConnect spec, it looks like Mark's problem with overloading
methods is due to a bug in Rhino with respect to how LiveConnect is
implemented.
In NativeJavaObject.java, the getConversionWeight method has the case for
JSTYPE_JAVA_ARRAY and JSTYPE_JAVA_OBJECT shares common code. In the
LiveConnect 3 spec, section 3.3.6.1 describes wrapped Java objects, and 3.3.6.2
describes wrapped Java arrays. Mark has a wrapped Java array of byte.
He is trying to call a method that has an overload of byte[] and for byte as
well. It is reporting an ambiguous choice, which is why we needed the syntax
for selecting the overload. BUt it looks like this should not be necessary.
According to the spec, you can pass a wrapped Java object to a primitive (non
boolean) type, but you cnnot do so for a wrapped Java array. Around line 366
in my version of the file, there is a test to account for the JSTYPE_JAVA_ARRAY
for passing to a primitive type. It looks to me like it should return
CONVERSION_NONE for a wrapped array, but it is returning CONVERSION_NONTRIVIAL.
The relevant lines of code are:
case JSTYPE_JAVA_OBJECT:
case JSTYPE_JAVA_ARRAY:
Object javaObj = fromObj;
if (javaObj instanceof Wrapper) {
javaObj = ((Wrapper)javaObj).unwrap();
}
if (to.isInstance(javaObj)) {
return CONVERSION_NONTRIVIAL;
}
if (to == ScriptRuntime.StringClass) {
return 2;
}
else if (to.isPrimitive() && to != Boolean.TYPE) {
return (fromCode == JSTYPE_JAVA_ARRAY)
? CONVERSION_NONTRIVIAL : 2 + getSizeRank(to);
}
break;
2007-09-28 17:43:38 +00:00
aa4d24bdda
JavaScript Tests - fix emca 262 section, bug 396969, not part of the build
2007-09-28 16:09:35 +00:00
b6599472d8
Fix bug in destructuring assignment for let expressions: e.g.
...
"let ([x, y] = [3, 4]) {}" in a functional context, and let
expressions for both function and script/activation contexts.
2007-09-28 15:37:04 +00:00
d4b0375405
Fix bug 397036: NPE with FEATURE_ENHANCED_JAVA_ACCESS
...
What was going on was a java.lang.SecurityException: "Can not make a
java.lang.Class constructor accessible" was getting thrown. So I've
restructured the new code for dealing with getting access to privates to handle
SecurityExceptions and just use accessible members.
I ran with Context.FEATURE_ENHANCED_JAVA_ACCESS set to true over the regression
suite and it turned up a set of cases where a NPE was thrown because a bad
scope object was passed down from ScriptRuntime.newCatchScope. I fixed that and
tightened up the case where a ClassCache isn't found (which was what was
happening with the invalid scope that later resulted in the NPE.
2007-09-27 19:02:37 +00:00
e25268a218
Patch from Steve Yegge (stevey@google.com) providing access to shell for
...
easier subclassing.
2007-09-27 17:59:00 +00:00
063916524f
Bug 397536 - "performance improvements for JSON.jsm" (optimize string serialization) [p=zeniko@gmail.com (Simon B��nzli) r=sspitzer sr=brendan a1.9=mconnor]
2007-09-27 16:50:29 +00:00
b6b1bd2a6a
JavaScript Tests - fix false negatives due to embedded "false!" in output, bug 397545, r=jorendorff, not part of the build
2007-09-26 14:24:01 +00:00
062c837ad7
Bug 396758: taking the patch out as it broke the test cases.
2007-09-25 17:30:55 +00:00
02994dd5c2
Bug 396758: the system flag is moved from GC flags to JSObject itself. r=brendan
2007-09-25 17:08:41 +00:00
8146cddd11
JavaScript Test - regression test for bug 396684, by Igor Bukanov, not part of the build
2007-09-24 19:20:54 +00:00
fc070995e8
JavaScript Tests - regression tests for bug 396969, by Brendan Eich, not part of the build
2007-09-24 18:41:14 +00:00
7a3c7d9cb1
Bug 397311 - Need a JSAutoRequest in xpcJSWeakReference::Init. r=brendan, sr+a=jst.
2007-09-24 06:09:56 +00:00
c35d1605c7
Fix obsolete test and add a new test to fix orange.
2007-09-22 21:42:36 +00:00
f297b7ebb8
Only allow XOW wrapped prototypes to go to null, not other objects. bug 394815, r+sr+a=brendan
2007-09-22 21:12:32 +00:00
17638547a5
Fix Convert to work for JSTYPE_VOID and make sure it reports an error. bug 394436, r=jst sr=brendan
2007-09-22 21:09:51 +00:00
47ab314acc
Fix old bug where we wouldn't close ts after a compilation error. bug 397136, r+a=brendan
2007-09-22 20:13:35 +00:00
04e097e5bf
JavaScript Test - regression test for bug 396900, by Norris Boyd, not part of the build
2007-09-21 17:11:48 +00:00
17feb968fe
JavaScript Test - regression test for bug 380933, by Jesse Ruderman, not part of the build
2007-09-21 14:55:45 +00:00
cdc5b5e039
JavaScript Test - regression test for bug 352616, not part of the build
2007-09-21 14:52:04 +00:00
ec9633c684
JavaScript Test - regression test for bug 381374, by Jesse Ruderman, not part of the build
2007-09-21 14:48:55 +00:00
c4dae4c238
JavaScript Test - regression test for bug 394967, by Jesse Ruderman, not part of the build
2007-09-21 14:45:48 +00:00
a0d0759501
JavaScript Tests - regression tests for bug 358594, by Jesse Ruderman, not part of the build
2007-09-21 14:39:09 +00:00
mrbkap%gmail.com
crowder%fiverocks.com
bent.mozilla%gmail.com
nboyd%atg.com
bzbarsky%mit.edu
szegedia%freemail.hu
bclary%bclary.com
benjamin%smedbergs.us
igor%mir2.org
reed%reedloden.com