Commit Graph

3435 Commits

Author SHA1 Message Date
antonm@chromium.org
69bb110075 Fix the case of no words to copy.
CopyWords cannot actually copy zero words---it'd clobber destiantion with
the first word of source.

Add an ASSERT to check this condition plus update array builtins to verify
for amount of copied data when necessary.

TBR=vitalyr@chromium.org

Review URL: http://codereview.chromium.org/1559004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4313 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-29 22:07:52 +00:00
lrn@chromium.org
17eece5748 Fix bug in string replace regexp with function when returning non-string.
Review URL: http://codereview.chromium.org/1528005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4312 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-29 21:12:54 +00:00
serya@chromium.org
4f01977e33 Fixing Windows build.
Review URL: http://codereview.chromium.org/1563002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4311 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-29 18:32:25 +00:00
antonm@chromium.org
cd826b5c41 Refactor word copying logic.
Review URL: http://codereview.chromium.org/1530005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4310 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-29 18:02:06 +00:00
serya@chromium.org
9e76d99fb9 Percise rounding parsing octal and hexadecimal strings.
Rounding happens when the number exceeds 53 bits of floating point mantissa. Current implemetation ignores digits after some limits. 0x1000000000000081 was rounded to 0x1000000000000100 while 0x100000000000008000001 was rounded to 0x100000000000000000000.

Review URL: http://codereview.chromium.org/1374005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4309 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-29 15:46:58 +00:00
kmillikin@chromium.org
08ed4504f6 Fix presubmit failure.
TBR=fschneider@chromium.org

Review URL: http://codereview.chromium.org/1563001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4308 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-29 14:43:41 +00:00
kmillikin@chromium.org
81779f6324 Rework flow graph construction.
The flow graph has been simplified to remove the special branch, join,
and exit nodes.  All nodes are now basic blocks (possibly empty to
preserve edge-split form) with a distinguished entry and exit block.

Most trivial expressions are not added to the flow graph as
instructions.  The assigned variable analyzer has been changed to
sometimes work right-to-left so that right subexpressions can be
marked as trivial.

The reaching definitions analysis has been temporarily removed, and
the analyses that depended on it (primitivity analysis, dead code
marking) as well.

Review URL: http://codereview.chromium.org/1530003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4307 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-29 14:23:55 +00:00
vitalyr@chromium.org
30efbb0622 Call binary op stub instead of runtime in count operations.
Review URL: http://codereview.chromium.org/1560001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4306 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-29 14:10:49 +00:00
vitalyr@chromium.org
3f2083286e Start using String type info:
* Improved string concatenation.

 * Fixed type inference in prefix/postfix count operations.

Review URL: http://codereview.chromium.org/1520001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4305 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-29 11:48:57 +00:00
serya@chromium.org
0b42d25c72 Optimization for parsing integers of limited length (in digits).
Review URL: http://codereview.chromium.org/1367004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4304 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-29 11:17:34 +00:00
ricow@chromium.org
3cf90789f9 Prepare push of version 2.2.0 to trunk. Now working on version 2.2.1.
Review URL: http://codereview.chromium.org/1527002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4301 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-29 07:25:54 +00:00
vitalyr@chromium.org
9117dfcff9 Some string optimizations:
* Faster hashing for sequential strings.

 * When adding short external two-byte strings try to convert them
   back to ascii. Chances are high the embedder uses two-byte
   representation even for ascii strings. This optimization saves
   memory and makes hashing faster.

Review URL: http://codereview.chromium.org/1444001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4300 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-26 23:33:37 +00:00
antonm@chromium.org
a25312a53e Omit write barrier when setting trimmed elements.
Review URL: http://codereview.chromium.org/1310002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4299 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-26 18:47:14 +00:00
lrn@chromium.org
845fd15dd6 Pre-create properties on JSRegExp objects
Initialize properties in single runtime call.

Review URL: http://codereview.chromium.org/1350003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4298 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-26 14:19:47 +00:00
whesse@chromium.org
75003d7e76 Change a TypeInfo function from inline to non-inline, reducing stack usage during code compilation.
Review URL: http://codereview.chromium.org/1369003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4297 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-26 11:34:00 +00:00
ager@chromium.org
adeb2f43f2 Prepare push of version 2.1.10 to trunk. Now working on version 2.2.0.
Review URL: http://codereview.chromium.org/1397003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4293 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-26 09:14:50 +00:00
ager@chromium.org
2bcc49740c Port number dictionary probing in generated code to ARM.
Fix bug in ARM pixel array load code and a typo in the x64 number
dictionary load code.

Fix bug in string dictionary probing where we did not bail out if the
object has an interceptor.

BUG=640

Review URL: http://codereview.chromium.org/1332003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4292 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-26 09:05:22 +00:00
sgjesse@chromium.org
2d0c823305 Proper constraints for inline implementation in assembly.
Review URL: http://codereview.chromium.org/1359002


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4291 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-26 07:58:24 +00:00
ricow@chromium.org
65e099e277 Land http://codereview.chromium.org/1311003/diff/8001/9001 to allows us to push to trunk. Corrected the ASSERT from the review.
Review URL: http://codereview.chromium.org/1404001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4290 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-26 07:55:38 +00:00
kasperl@chromium.org
33548b59e2 Avoid having a LIBS setting when compiling the V8 library
because this interferes with the shared library build.
Only prepend v8 to the LIBS list when building samples
or cctest.
Review URL: http://codereview.chromium.org/1387004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4289 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-26 07:47:41 +00:00
vitalyr@chromium.org
6fbad6e230 Simplify passing of AccessorInfo to interceptors:
* Use slots on the native stack when possible instead of Relocatable.
 * Got rid of a gap in AccessorInfo fields.
 * Added test for non-cacheable post-interceptor lookup.

Review URL: http://codereview.chromium.org/1327002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4288 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 17:08:22 +00:00
kmillikin@chromium.org
70bbac9c56 Move flow graph and helper classes to their own file.
The FlowGraph, FlowGraphBuilder, and flow graph node classes are moved
to src/flow-graph.cc.

Review URL: http://codereview.chromium.org/1253009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4287 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 16:22:48 +00:00
antonm@chromium.org
94a2a0956b Add a test to verify correctness of remembered set update for Heap::CopyJSObject.
Review URL: http://codereview.chromium.org/1256002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4286 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 15:32:58 +00:00
mikhail.naganov@gmail.com
27663efb85 Align OProfile agent invocations style with logging.
This allows to unclutter logging-related code.
I also fixed compilation issues with 'profilingsupport=off'.

Review URL: http://codereview.chromium.org/1317003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4285 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 14:55:53 +00:00
serya@chromium.org
203ba181ad StringToDouble rewritten not using String::Get and memory allocations.
It converts the number to "canonical" form removing insignificant digits,
leading zerroes and spaces what guarantees to fit a fixed size buffer and
does not changes result of strtod.
Review URL: http://codereview.chromium.org/1216005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4282 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 14:39:39 +00:00
antonm@chromium.org
5e24ffaa22 Slightly faster variant of Array.push.
Review URL: http://codereview.chromium.org/1265003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4279 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 14:20:56 +00:00
antonm@chromium.org
f45c35ebd3 Allow empty handles to escape handle scopes.
BUG=39170

Review URL: http://codereview.chromium.org/1235002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4278 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 14:07:17 +00:00
erik.corry@gmail.com
6a8398cffa Fix my fast-codegen.cc gyp file stupidity.
Review URL: http://codereview.chromium.org/1207007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4277 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 13:56:29 +00:00
sgjesse@chromium.org
e2392bd586 Prepare push to trunk, we are now working on 2.1.10
Review URL: http://codereview.chromium.org/1242007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4276 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 13:55:52 +00:00
fschneider@chromium.org
394a519343 Fix build again after renaming.
Review URL: http://codereview.chromium.org/1335001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4275 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 13:35:05 +00:00
erik.corry@gmail.com
d20232096f Gyp causing me gyp.
Review URL: http://codereview.chromium.org/1323003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4274 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 13:31:19 +00:00
whesse@chromium.org
a255a71181 Speed up conversion of untagged int32 result if we know it must be a smi.
Review URL: http://codereview.chromium.org/1233003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4273 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 13:22:37 +00:00
erik.corry@gmail.com
f5523ecc6a Split the virtual frame into heavy and light versions.
The heavy version is for x86 and x64.  The light version
is for ARM and MIPS.  Remove the elements_ array from the
virtual frame in the light version.  More simplifications
to come, followed by light register allocation.
Review URL: http://codereview.chromium.org/1164002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4272 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 13:18:00 +00:00
fschneider@chromium.org
5b50fd70ef Fix build.
Some old names reappeared due to some merge problem.

Review URL: http://codereview.chromium.org/1334001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4271 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 13:14:25 +00:00
antonm@chromium.org
9477ea8735 Fix LargeObjectSpace::Contains to check if addr is in new space.
Otherwise page header check is not quite robust: if there is a smi
at the same offset as Page::is_normal_page field, wrong result would
be returned.

That shouldn't be the problem for paged spaces as objects in those
pages do not span page boundaries and thus cannot mess with ::is_normal_page field.

Review URL: http://codereview.chromium.org/1175001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4270 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 13:10:50 +00:00
lrn@chromium.org
498b074bd0 Run string replace regexp with function in C++ code loop.
Reuses the result array to save on allocation.
Matches Safari's behavior.

Review URL: http://codereview.chromium.org/1109010

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4269 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 12:57:58 +00:00
fschneider@chromium.org
0737ec8dcb Rename NumberInfo to TypeInfo.
Since we add more type (StringType, PrimitiveType) the name
NumberInfo does not make sense anymore.

Review URL: http://codereview.chromium.org/1207006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4268 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 12:44:15 +00:00
lrn@chromium.org
4cd39f12b4 Fix bug in RegExp first-character-lookahead.
Review URL: http://codereview.chromium.org/1253008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4267 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 12:41:26 +00:00
pfeldman@chromium.org
41830911f3 Make following ES5 functions work with undetectable parameters/target (document.all):
ObjectIsPrototypeOf
ObjectDefineGetter
ObjectLookupGetter
ObjectDefineSetter
ObjectLookupSetter
ObjectKeys
ObjectGetPrototypeOf
ObjectGetOwnPropertyDescriptor
ObjectGetOwnPropertyNames
ObjectDefineProperty
ObjectDefineProperties

I did not implement tests covering calls with undetectable parameters since I would need to make these tests native. Just thought it was not worth the effort. If you think tests would make sense, I can add them. We might want to allow other functions to receive undetectable parameters, but I am not seeing any good candidates.

Review URL: http://codereview.chromium.org/1297003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4264 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 12:28:24 +00:00
pfeldman@chromium.org
65025e4659 Fail with out-of-bounds exception upon illegal access to the ExecutionState's frame.
Review URL: http://codereview.chromium.org/1284001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4263 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 12:20:59 +00:00
sgjesse@chromium.org
6a63910650 Re-apply "Inline floating point compare"
This re-applies r4220 and r4233, which was reverted in r4254 due to a bug. This bug has now been fixed, with the only change being line 2884 changed from

  __ SmiTag(left_side->reg());

to

  __ SmiTag(operand->reg());

Added a regression test.

BUG=http://crbug.com/39160
TEST=test/mjsunit/regress/regress-crbug-39160.js

Review URL: http://codereview.chromium.org/1251009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4261 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 12:04:34 +00:00
whesse@chromium.org
244513555c Avoid calling ToRegister(register) when result is in register already, and register is shared.
Review URL: http://codereview.chromium.org/1325004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4260 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 11:52:42 +00:00
lrn@chromium.org
5f764c82a6 Change StringSearch to not call exec and build unnecessary intermediate array.
Review URL: http://codereview.chromium.org/1223006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4259 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-25 09:09:42 +00:00
fschneider@chromium.org
64bc213c2a Check for negative 0 when deciding whether a constant has type Integer32.
Review URL: http://codereview.chromium.org/1225005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4258 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-24 16:37:27 +00:00
fschneider@chromium.org
764d8a9530 Fix Win32 build.
Use correct constants for min/max-int.

Review URL: http://codereview.chromium.org/1239004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4257 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-24 15:55:15 +00:00
fschneider@chromium.org
51495beb78 Store type information with constants.
Instead of testing the value of a constant frame element to determine
the type we compute its type information at construction time.

This speeds up querying the type information during code generation.

This change also adds support for Integer32 constants and sets 
the type information accordingly.

Review URL: http://codereview.chromium.org/1277001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4256 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-24 15:29:41 +00:00
kmillikin@chromium.org
c0c1ebcaa4 Revert "Inline floating point compare"
This reverts svn revisons r4220 and r4233.

Review URL: http://codereview.chromium.org/1279001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4254 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-24 14:26:14 +00:00
ager@chromium.org
652335b995 Implement support for reattaching a global object to a context. This
can be used to use the same global object for different contexts at
different points in time.

Review URL: http://codereview.chromium.org/1207003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4253 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-24 13:24:46 +00:00
erik.corry@gmail.com
82ae9cbe96 Fix Android scons build by putting -lv8 before -lgcc.
Review URL: http://codereview.chromium.org/1275001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4252 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-24 13:10:18 +00:00
pfeldman@chromium.org
54f0a5b221 Expose debug context into the native code. This change is doing several things:
1) Exposes Local<Context> v8::Debug::GetDebugContext().
Rationale: We can already get debugger context instance using various workarounds,
so exposing it explicitly in the API only makes things more clear.

2) Removes debugger.HasJavaScriptFrames() requirement for entering debugger context.
Rationale: Sometimes we'd like to call into debugger from the external native code.

3) Makes Debugger v8::Debug::Call execute on debugger context's global object.
Rationale: This is somewhat arguable, but temporary measure. We've agreed that we
should introduce a DebugAPI object that would expose necessary API in the debug-delay.
The problem is that it would take some time to define this API and it is not really
convenient to do that on the v8 land given the difference in v8 / host lifecycle.
The plan is to compose this API as a Debug.* wrapper outside v8 by means of exposing
debugger context's global object here, and once API settles down, upstream it to v8 and
start making Debug::Calls on this new DebugAPI object instead.


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4251 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-03-24 13:09:02 +00:00