6d24485f818545901b0e0225a39bb2d9709c35d5: Actually fixed ifeq/ifne fusion now.
David Anderson <danderson@mozilla.com> - Thu, 10 Jul 2008 17:54:19 -0700 - rev 17595
Push 1452 by shaver@mozilla.com at Fri, 22 Aug 2008 00:08:22 +0000
Actually fixed ifeq/ifne fusion now.
b9585554a05fd81e656fd6abaf704689bf2291bf: Merge.
Andreas Gal <gal@mozilla.com> - Thu, 10 Jul 2008 17:42:25 -0700 - rev 17594
Push 1452 by shaver@mozilla.com at Fri, 22 Aug 2008 00:08:22 +0000
Merge.
5c3e6ebd8e7214d0b939fe4b8c16b34a373e2a1c: Merge.
Andreas Gal <gal@mozilla.com> - Thu, 10 Jul 2008 17:41:45 -0700 - rev 17593
Push 1452 by shaver@mozilla.com at Fri, 22 Aug 2008 00:08:22 +0000
Merge.
de3ae43c32ef21abde052836f609c1e72673d252: nanojit doesn't support loads with non-constant offsets so don't do that
Andreas Gal <gal@mozilla.com> - Thu, 10 Jul 2008 17:41:24 -0700 - rev 17592
Push 1452 by shaver@mozilla.com at Fri, 22 Aug 2008 00:08:22 +0000
nanojit doesn't support loads with non-constant offsets so don't do that
cbc0682e1efb92e930beada508ee4f15660d73a4: fix opcode math with the mighty hammer of casting
shaver@mozilla.org - Thu, 10 Jul 2008 20:41:55 -0400 - rev 17591
Push 1452 by shaver@mozilla.com at Fri, 22 Aug 2008 00:08:22 +0000
fix opcode math with the mighty hammer of casting
bfabace39fc3d4577e3c5d91ff078674f9ff2027: the rest of the stdint defs, no idea why I didn't do them before
shaver@mozilla.org - Thu, 10 Jul 2008 20:25:57 -0400 - rev 17590
Push 1452 by shaver@mozilla.com at Fri, 22 Aug 2008 00:08:22 +0000
the rest of the stdint defs, no idea why I didn't do them before
c01216792073c56cb3d056210d43724a9045182d: Merge.
Andreas Gal <gal@mozilla.com> - Thu, 10 Jul 2008 17:05:13 -0700 - rev 17589
Push 1452 by shaver@mozilla.com at Fri, 22 Aug 2008 00:08:22 +0000
Merge.
ee609b5493d2c898afb29969891d685f7768db4f: Steal the reference to cx from the BoxDouble call instead of observing the load.
Andreas Gal <gal@mozilla.com> - Thu, 10 Jul 2008 17:03:59 -0700 - rev 17588
Push 1452 by shaver@mozilla.com at Fri, 22 Aug 2008 00:08:22 +0000
Steal the reference to cx from the BoxDouble call instead of observing the load.
aa3c905521b57a1bf389a877a30b2b1cf32994fd: Fixed the setelem a[i] bug.
Andreas Gal <gal@mozilla.com> - Thu, 10 Jul 2008 16:55:37 -0700 - rev 17587
Push 1452 by shaver@mozilla.com at Fri, 22 Aug 2008 00:08:22 +0000
Fixed the setelem a[i] bug.
5c0c0d1659b9792a658a022c91b8159df73bfdd7: Nanojit needs a LINUX define (which may be my fault), fixing it here for now
David Anderson <danderson@mozilla.com> - Thu, 10 Jul 2008 16:26:28 -0700 - rev 17586
Push 1452 by shaver@mozilla.com at Fri, 22 Aug 2008 00:08:22 +0000
Nanojit needs a LINUX define (which may be my fault), fixing it here for now
0a5b824dcc4eec7fb1018b82fbd37a7874db7e23: use VirtualAlloc for Windows, since it lacks valloc
shaver@mozilla.org - Thu, 10 Jul 2008 18:39:51 -0400 - rev 17585
Push 1452 by shaver@mozilla.com at Fri, 22 Aug 2008 00:08:22 +0000
use VirtualAlloc for Windows, since it lacks valloc
abf6ecb322d18005e3bce34c6e302919a3241784: interp needs tracer.h
shaver@mozilla.org - Thu, 10 Jul 2008 17:31:38 -0400 - rev 17584
Push 1452 by shaver@mozilla.com at Fri, 22 Aug 2008 00:08:22 +0000
interp needs tracer.h
584d2a3dbb9c0691471eb9cf1d2f8e7e3a66bbf9: Decouple jscntxt.h from jstracer.h so that xpconnect doesn't try to include all of
shaver@mozilla.org - Thu, 10 Jul 2008 17:29:16 -0400 - rev 17583
Push 1452 by shaver@mozilla.com at Fri, 22 Aug 2008 00:08:22 +0000
Decouple jscntxt.h from jstracer.h so that xpconnect doesn't try to include all of nanojit.
7f92daca9d4b07f5b93c35541ceb2e52ad0670c2: FASTCALL for Windows
shaver@mozilla.org - Thu, 10 Jul 2008 17:16:31 -0400 - rev 17582
Push 1452 by shaver@mozilla.com at Fri, 22 Aug 2008 00:08:22 +0000
FASTCALL for Windows
9a43f36e5cb1ec415c4ecaa9727b552dc5e9a9a9: int32_t, now available on Windows
shaver@mozilla.org - Thu, 10 Jul 2008 16:58:08 -0400 - rev 17581
Push 1452 by shaver@mozilla.com at Fri, 22 Aug 2008 00:08:22 +0000
int32_t, now available on Windows
c7f76b35889707556b803672374057b11a8fa481: Unlike software developers, g++ doesn't like variable-sized arrays. Have some alloca!
shaver@mozilla.org - Thu, 10 Jul 2008 16:48:42 -0400 - rev 17580
Push 1452 by shaver@mozilla.com at Fri, 22 Aug 2008 00:08:22 +0000
Unlike software developers, g++ doesn't like variable-sized arrays. Have some alloca!
74f3632c456dbf30f71afac1f9ef6fe6a99d72ed: merge
shaver@mozilla.org - Thu, 10 Jul 2008 16:47:12 -0400 - rev 17579
Push 1452 by shaver@mozilla.com at Fri, 22 Aug 2008 00:08:22 +0000
merge
7211696c8180e83aa73c7b18def35163e6c3bd05: update trace-test.js, now crashes calling lsh() the second time
shaver@mozilla.org - Thu, 10 Jul 2008 10:45:11 -0400 - rev 17578
Push 1452 by shaver@mozilla.com at Fri, 22 Aug 2008 00:08:22 +0000
update trace-test.js, now crashes calling lsh() the second time
6c9e4594f64806cfaf3945b9141e0c244344b150: Fixed guarding of eq+ifeq/ifne fusions and enabled tracing JSOP_GOTO (no-op)
David Anderson <danderson@mozilla.com> - Thu, 10 Jul 2008 13:35:17 -0700 - rev 17577
Push 1452 by shaver@mozilla.com at Fri, 22 Aug 2008 00:08:22 +0000
Fixed guarding of eq+ifeq/ifne fusions and enabled tracing JSOP_GOTO (no-op)
ae78505d5548754ebd79733f9557314d592548be: Root all strings and objects first when unboxing. Then box values that might trigger the GC (doubles/ints). This probably needs some performance tuning over time.
Andreas Gal <gal@mozilla.com> - Thu, 10 Jul 2008 13:24:49 -0700 - rev 17576
Push 1452 by shaver@mozilla.com at Fri, 22 Aug 2008 00:08:22 +0000
Root all strings and objects first when unboxing. Then box values that might trigger the GC (doubles/ints). This probably needs some performance tuning over time.
(0) -10000 -3000 -1000 -300 -100 -50 -20 +20 +50 +100 +300 +1000 +3000 +10000 +30000 +100000 +300000 tip