Commit Graph

43464 Commits

Author SHA1 Message Date
Luke Wagner
42890c79a8 Double browser max heap size. Change shell heap size so it matches. 2010-06-30 20:02:00 -07:00
Luke Wagner
abc0ed6b6d Fix build bustage 2010-06-30 18:23:30 -07:00
Luke Wagner
d1c1fdd62b Fix build bustage on MSVC 2010-06-30 18:05:19 -07:00
Luke Wagner
8c46bea4d0 Forgot to add jsval.h and jsvalue.h 2010-06-30 17:52:02 -07:00
Luke Wagner
68403a872a Get narcissus compiling 2010-06-30 16:32:44 -07:00
Luke Wagner
71da5b9aa9 Split extract*Type functions and add comments explaining usage 2010-06-30 15:45:54 -07:00
Luke Wagner
c2f68a0182 Pull jsval impl details into jsvalimpl.h (included by jspubtd.h) and pull js::Value into jsvalue.h (engine-private) 2010-06-30 15:25:47 -07:00
Luke Wagner
f82e2f8002 Take the "Raw" out of MarkRaw and MarkGCThingRaw. It offends Andreas, and it might be a bad name 2010-06-29 23:02:23 -07:00
Luke Wagner
19f7cefc18 Drop the distinction in js::Value of funobj vs. nonfunobj. Instead, values can just have JSVAL_TYPE_OBJECT. The trace-jit still uses funobj/nonfunobj and translates as necessary. 2010-06-29 22:33:52 -07:00
Luke Wagner
6986d7f53e Change runtime size to 128MB from 64MB. It seems like we are running up to this limit with v8-splay and so we spend a lot more time GC'ing. Need to see if we should do this for the browser as well. 2010-06-29 02:35:39 -07:00
Luke Wagner
9d26f036a3 Remove unnecessary TODO 2010-06-29 02:10:48 -07:00
Luke Wagner
b8a731853a Put JSContext argument back at the end of SameValue (apparently that means something) 2010-06-29 02:08:07 -07:00
Luke Wagner
ae0f7e12b6 Change js::DumpX back to js_DumpX 2010-06-29 01:49:55 -07:00
Luke Wagner
86f4b4504b Use the same tag/payload trick in js_UnboxInt32 as js_UnboxDouble 2010-06-29 01:42:28 -07:00
Luke Wagner
313b355d22 Rearrange JSStackFrame... for 2% on both SS and V8 2010-06-29 01:07:12 -07:00
Luke Wagner
10745fb373 Merge 2010-06-28 18:10:50 -07:00
Luke Wagner
ab3a0d3ffa Add 64-bit tracing support 2010-06-28 18:08:28 -07:00
David Anderson
da423a57a7 Fixed Linux 64-bit opt build. 2010-06-28 15:51:50 -07:00
David Mandelin
452e333382 Bug 573578 followup: restore JS_IdToValue API to original functionality 2010-06-28 11:40:33 -07:00
David Mandelin
9da17d8247 Bug 574881: complete implementation of IdToString as needed with fatvals 2010-06-25 17:49:11 -07:00
David Mandelin
67eb767a51 Bug 574874: relax assertion as needed 2010-06-25 17:47:48 -07:00
David Mandelin
113fdd40ed Bug 574518: Better solution: recur until error so that test autotunes for memory usage 2010-06-25 16:47:30 -07:00
David Mandelin
ffa14c0ca4 Check for void jsid returned from JS_NextProperty in ctypes 2010-06-25 16:42:57 -07:00
David Mandelin
08b4c31ff3 Clean up unneeded initializers from previous patch 2010-06-25 14:58:34 -07:00
David Mandelin
53570dcb31 Bug 574745: fix com ptr usage causing leak on fatval buildbots 2010-06-25 14:22:02 -07:00
Luke Wagner
b6ba48936a Merge 2010-06-24 22:09:56 -07:00
Luke Wagner
e80b60a250 Fix bug where vp_ins was being passed uninitialized to SetPropertyByName 2010-06-24 22:09:03 -07:00
David Mandelin
fb83e7796d Fix jsreftest js1_5/Regress/regress-159334.js 2010-06-24 17:39:01 -07:00
David Mandelin
001939737c Bug 574518: adjust test case to account for greater stack memory usage in Interpret
--HG--
extra : rebase_source : d8cee395b898d809020b503b83a402de8cb88eee
2010-06-24 17:14:06 -07:00
Luke Wagner
2365632b1c Merge 2010-06-24 16:29:36 -07:00
Luke Wagner
d59fc36dbf Add back the js_UnboxDouble check in TraceRecorder::box_value, simplify code that skips the js_UnboxDouble 2010-06-24 16:28:08 -07:00
David Mandelin
c6783f6875 Removing failure note from test that passes 2010-06-24 14:04:26 -07:00
Luke Wagner
558aa8414b Avoid unnnecessary extra LIR in unbox_value 2010-06-24 11:21:20 -07:00
Luke Wagner
561b9386c9 Have box_value callers pass the AccSet to use 2010-06-23 17:50:53 -07:00
Luke Wagner
d474e4bd4a Tweak names/comments to account for JS_NEITHER change 2010-06-23 17:23:57 -07:00
Luke Wagner
42622db476 Merge 2010-06-23 17:13:46 -07:00
David Mandelin
bb7c22a9ba Bug 571623: restore support for traceable natives. With much help from peterv. 2010-06-23 17:11:06 -07:00
Luke Wagner
eba2e1f102 Merge and fix 2010-06-23 17:06:18 -07:00
Luke Wagner
ed28dae403 64-bit builds and passes trace tests on OS X with --disable-tracejit. reftests has failures, but these also show up on tm and should be fixed by a rebase 2010-06-23 16:58:32 -07:00
David Mandelin
5635d6ffb5 Adding a space to push to cedar 2010-06-22 15:46:26 -07:00
David Mandelin
a2628ff847 Windows C build fix 2010-06-22 15:26:36 -07:00
Luke Wagner
50d922350a Fix bustage compiling jsapi.h as Objective-C++ 2010-06-22 11:54:27 -07:00
Luke Wagner
bae4687e65 Reuse existing JS_ValueToId instead of introducing new JSVAL_TO_JSID 2010-06-22 11:51:46 -07:00
Luke Wagner
9cdf8a36e7 Fix bug in TraceRecorder::storeHole introduced earlier 2010-06-22 10:47:54 -07:00
Luke Wagner
c872ea6cb5 Fix bugs introduced by overzealous application of setNumber in cset 5c5bb620d2c8 2010-06-22 10:14:11 -07:00
Luke Wagner
f536e360bf Take advantage of value representation in the tracer when mapping values to and from trace types 2010-06-22 09:43:21 -07:00
Luke Wagner
7db2d34bde Merge 2010-06-21 16:30:17 -07:00
Luke Wagner
ebb2d4890b Be careful when using setDouble that we shouldn't use setNumber instead to fit in an int when possible. Was missing a lot in jsdate.cpp 2010-06-21 16:28:02 -07:00
David Mandelin
4ab1d47307 Bug 573578: intern string before converting to jsid in PropertyOpForwarder 2010-06-21 16:07:31 -07:00
Luke Wagner
90a6ed7cd3 Merge 2010-06-21 12:51:25 -07:00