cd92319223
Optimize string length one comparisons by comparing char codes only. Implemented for one byte strings only at the moment. Will add other string types/mixed compares later.
...
R=fschneider@google.com
Review URL: https://codereview.chromium.org//105143011
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@31347 260f80e4-7a28-3924-810f-c04153c831b5
2013-12-20 20:11:38 +00:00
e994d8b697
Revert r31326.
...
TBR=johnmccutchan@google.com ,
Review URL: https://codereview.chromium.org//118133004
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@31332 260f80e4-7a28-3924-810f-c04153c831b5
2013-12-20 12:21:09 +00:00
fe726e1971
Use parallel move resolver for optimized try-catch.
...
This CL has no change in functionality and is purely clean up and
refactoring. Instead of manually generating the moves at throwing
instructions in try-catch, construct a parallel move and use the resolver
to emit the native code. This eliminates a lot of duplicated code
from all platforms.
It will also allow to re-use stack space that is currently allocated
separately for each individual try-catch in a function.
I added a few more unrelated minor changes in various parts of the VM
* Simpilify guard code generation
* Resolve refactoring TODO in deoptimization
* Improve names
R=johnmccutchan@google.com
Review URL: https://codereview.chromium.org//119213002
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@31326 260f80e4-7a28-3924-810f-c04153c831b5
2013-12-20 09:41:42 +00:00
a576fec3c0
Allow the native resolver to setup whether it needs the Dart API scope to
...
be setup automatically or not when a native function is invoked.
This would allow the embedder to treat some native functions as leaf
functions whose invocation can be very light.
R=hausner@google.com , zra@google.com
Review URL: https://codereview.chromium.org//117723002
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@31286 260f80e4-7a28-3924-810f-c04153c831b5
2013-12-19 17:47:56 +00:00
f7149b05c8
Fix bug in double unboxing.
...
The state bit that record if a field is unboxed or not has to be checked
in unoptimized code as well, since that code does not change when deoptimizing.
I also compacted the kind_bits bit field in the Field class to start from bit 0.
R=srdjan@google.com
Review URL: https://codereview.chromium.org//106623006
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@31229 260f80e4-7a28-3924-810f-c04153c831b5
2013-12-18 15:46:39 +00:00
6abd6e326e
Add mutable double boxes for fields.
...
This allows the optimizing compiler to generate unboxed loads/stores
to fields containing double values. The double value is stored
in a reusable double object.
Unboxed loads/stores are generated for optimized code. Unoptimized code
allocates a new double on loads. To avoid performance regressions
for fields that are only written few times (e.g. only in the constructor)
I put a heuristic in place that
compares the usage count of setters and getters. Unboxed operations
are only generated if the setter is invoked a significant amount of
times (threshold is 10% of getter invocations).
The CL is so big because it changes the way LocationSummmary
is allocated: We now have a bit to generate different summaries
for optimized and unoptimized code.
R=srdjan@google.com
Review URL: https://codereview.chromium.org//99573005
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@31164 260f80e4-7a28-3924-810f-c04153c831b5
2013-12-16 15:11:31 +00:00
5aa02dce31
Small cleanup and more test coverage for guarded fields.
...
Some cases of field guards for guarded length were not covered by our tests.
Otherwise I removed redundant conditions that are either impossible
or always true from the code generator for GuardFieldInstr.
R=johnmccutchan@google.com
Review URL: https://codereview.chromium.org//104583002
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@30887 260f80e4-7a28-3924-810f-c04153c831b5
2013-12-05 11:17:06 +00:00
a8f549dbc3
Merge sin(a), cos(a) into one instruction. On IA32 use x87 fsincos operation, on x64 call runtime routine whihc merges sin/cos intos sincos.
...
TODO: Implement for ARM and MIPS.
R=fschneider@google.com
Review URL: https://codereview.chromium.org//92433002
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@30827 260f80e4-7a28-3924-810f-c04153c831b5
2013-12-03 22:00:27 +00:00
cf53988cf6
Use range information to omit checks in TRUNCDIV/MOD operations.
...
R=fschneider@google.com
Review URL: https://codereview.chromium.org//81623003
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@30673 260f80e4-7a28-3924-810f-c04153c831b5
2013-11-26 18:18:24 +00:00
8b0b3fde4d
Generalize if-conversion to arbitrary smi comparisons.
...
Until now only == and != comparisons were supported.
Further changes are:
Remove flag FLAG_new_identity_spec: It is not needed anymore.
Invoke Canonicalize another time before branch optimizations.
Fixes support for TestSmiInstr with the branch-simplifier pass.
R=kmillikin@google.com
Review URL: https://codereview.chromium.org//78733002
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@30622 260f80e4-7a28-3924-810f-c04153c831b5
2013-11-25 12:57:21 +00:00
3165e204e3
Merge TRUNCDIV and MOD into one instruction. Icorporated feedback from CL https://codereview.chromium.org/68663003
...
R=fschneider@google.com
Review URL: https://codereview.chromium.org//79653002
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@30536 260f80e4-7a28-3924-810f-c04153c831b5
2013-11-21 17:56:54 +00:00
05b8012333
Remove dead code from IfThenElse and StrictCompare in the compiler.
...
The code for these cases is either completely unused or can only be hit with
canonicalization and constant propagation turned off. Since this is not tested,
and does not bring any performance benefit for the default configuration it is
best to just remove it.
R=srdjan@google.com
Review URL: https://codereview.chromium.org//62203018
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@30393 260f80e4-7a28-3924-810f-c04153c831b5
2013-11-19 10:38:37 +00:00
88726f7f70
Streamline code generator for branches and comparisons.
...
This is another refactoring step. It turn all comparisons
to be generated after the same pattern and removes duplicated
code from each of the different comparisons (smi, mint, double, etc.)
A comparison as an expression now follows the same patterns as a comparison
inside a branch. It just adds a common epilogue to materialize the boolean
result.
In another CL this common epilogue will be generated by reusing the IfThenElse
IL instruction.
R=kmillikin@google.com
Review URL: https://codereview.chromium.org//64483002
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@30243 260f80e4-7a28-3924-810f-c04153c831b5
2013-11-13 18:21:27 +00:00
ee554e199d
Cleanup of code generation for StrictCompare.
...
The case of two constant operands does not occur unless certain optimizations
are turned off. It is not necessary to have a special handling in the code
generator for this case.
R=kmillikin@google.com
Review URL: https://codereview.chromium.org//60883003
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@30153 260f80e4-7a28-3924-810f-c04153c831b5
2013-11-11 13:01:54 +00:00
b97c8e894c
Inline integer modulo operation.
...
R=zra@google.com
Review URL: https://codereview.chromium.org//61123003
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@30075 260f80e4-7a28-3924-810f-c04153c831b5
2013-11-07 21:50:26 +00:00
b324e06153
VM: Fix identical comparisons with bigints.
...
The optimizing compiler did not properly preserve registers across
the runtime call that occurs when using identical with bigints.
BUG=http://dartbug.com/14903
TEST=tests/language/vm/regress_14903_test.dart
R=srdjan@google.com
Review URL: https://codereview.chromium.org//64723002
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@30058 260f80e4-7a28-3924-810f-c04153c831b5
2013-11-07 17:27:16 +00:00
b15bd6fe04
Cleanup of branch code generation (no change in functionality).
...
This CL is the first step in refactoring the way branches and comparisons
are generated.
1. Move helper functions from flow_graph_compiler_xyz.cc to
intermediate_language_xyz.cc.
2. Remove IL class ControlInstruction. It was only implemented by BranchInstr.
All functions provided are moved to BranchInstr.
3. When generating branch code for comparisons, pass the successor labels explicitly
instead of getting them from the branch. This will allow us to provide different labels
when materialize a bool value of a comparison.
4. Move some common code for IfThenElseInstr from the platform-specific files
into intermediate_language.cc and simplify it.
The goal is to enable if-conversion of arbitrary comparisons. Right now,
the code for == is hard-coded in IfThenElseInstr (and duplicated, too). This
means that e.g. "a < b ? 0 : 1" cannot be optimized.
As a result, IfThenElseInstr can be used to materialize the boolean value of
a comparison. This way the complication of having ComparisonInstr both as a
normal instruction and as wrapped inside a BranchInstr can be simplified.
Comparisons would no longer appear as plain instructions in the IL, but only
wrapped inside either a Branch or an IfThenElse(true, false).
R=zra@google.com
Review URL: https://codereview.chromium.org//62133002
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@30042 260f80e4-7a28-3924-810f-c04153c831b5
2013-11-07 11:47:04 +00:00
aa789a7b27
Merge (x & y) == 0 pattern to emit a single test instruction.
...
This is based on a previous uncommitted CL by vegorov@
(https://codereview.chromium.org/14251023/ ).
It is rebased and fixes a bug in the MIPS implementation: BranchOnCondition
requires the comparison result in fixed registers CMPRES1/CMPRES2.
Remove register alias TMP1 (=TMP) and CMPRES (=CMPRES1). It is confusing to
have them around and easy to forget that they are actually the same.
R=srdjan@google.com , zra@google.com
Review URL: https://codereview.chromium.org//59613005
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@29966 260f80e4-7a28-3924-810f-c04153c831b5
2013-11-06 12:13:29 +00:00
ca82c461a0
Impelments IfThenElse instruction for ARM and MIPS.
...
R=fschneider@google.com
Review URL: https://codereview.chromium.org//58003003
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@29933 260f80e4-7a28-3924-810f-c04153c831b5
2013-11-05 20:45:25 +00:00
29c01044bc
This is the final breaking change in dart:typed_data needed for Dart 1.0. We need this change because the ECMAScript SIMD specification only includes Int32x4 and Int32x4List and our types must match.
...
R=srdjan@google.com
Review URL: https://codereview.chromium.org//56023004
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@29849 260f80e4-7a28-3924-810f-c04153c831b5
2013-11-04 21:56:59 +00:00
9031c80c88
Change == into an instance call to allow polymorphic inlining of ==.
...
In unoptimized code equality is now just another instance call.
The optimizer replaces it with a specialized implementation based on static
type information and type feedback.
Many of the manual optimizations of == in the optimizer are now just handled
by the generic inliner, plus polymorphic inlining of == calls is now possible.
This also eliminates the need for a lot of duplicated code in the backend.
I adapted the inlining heuristics to compensate for the slightly larger
inital flow graph size.
Review URL: https://codereview.chromium.org//27307005
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@29800 260f80e4-7a28-3924-810f-c04153c831b5
2013-11-04 11:32:52 +00:00
5f63715d47
SIMD shuffle API changes
...
R=srdjan@google.com , zra@google.com
Review URL: https://codereview.chromium.org//51373004
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@29723 260f80e4-7a28-3924-810f-c04153c831b5
2013-11-01 07:09:54 +00:00
8a0fd8ce4d
Add Uint32x4List to typed_data
...
R=asiva@google.com , srdjan@google.com , vsm@google.com
Review URL: https://codereview.chromium.org//51333005
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@29639 260f80e4-7a28-3924-810f-c04153c831b5
2013-10-31 04:57:38 +00:00
3fa7a61efc
Implement latest spec changes regarding malformed types (see issue 14006):
...
- a malformed type used in a is, as, or catch type test results in a dynamic
type error being thrown.
- a malformed type is not mapped to dynamic anymore.
For now, this applies to type tests in checked mode as well, but discussion is
still on-going.
R=hausner@google.com
Review URL: https://codereview.chromium.org//53583003
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@29608 260f80e4-7a28-3924-810f-c04153c831b5
2013-10-30 23:30:32 +00:00
df1e5247fe
Adds support for ARM softfp calling convention.
...
The cross-compiler with the Android NDK doesn't support
the 'hard' calling convention, so we need both. I've
added a preprocessor define to the gyp file for the
standalone vm indicating support for 'hard' so that I
don't have to modify the Dartium gyp files.
R=srdjan@google.com
Review URL: https://codereview.chromium.org//52803002
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@29577 260f80e4-7a28-3924-810f-c04153c831b5
2013-10-30 18:57:55 +00:00
253fc69ecc
Record a correct deoptimization environment for slow-path code.
...
Slow-path code is not marked as a call, so the deoptimization environment
built by the register allocator can contain unspilled registers. However,
all registers are spilled before the call. Compute the spill slot for all
registers and change the ones in the deoptimization environment to be
spilled.
BUG=https://code.google.com/p/dart/issues/detail?id=13995
R=fschneider@google.com
Review URL: https://codereview.chromium.org//26823006
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@29239 260f80e4-7a28-3924-810f-c04153c831b5
2013-10-25 10:17:10 +00:00
1efeb05d4d
Fixes ARM SIMD IL bug.
...
Issue 14341
R=fschneider@google.com
Review URL: https://codereview.chromium.org//37413002
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@29092 260f80e4-7a28-3924-810f-c04153c831b5
2013-10-23 15:56:35 +00:00
b83bda9e0e
Add new style of string interpolation optimization: new nodes, working constant folding.
...
TODO: partial constant fold of incoming arguments.
R=kmillikin@google.com
Review URL: https://codereview.chromium.org//27192005
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@28741 260f80e4-7a28-3924-810f-c04153c831b5
2013-10-16 18:12:08 +00:00
e375053d2f
Pattern match on generated code to find edge counters.
...
In unoptimized code, use platform-specific pattern matching on generated
code to find edge counter arrays. Previously we searched pointer offsets,
but that does not work on platforms that encode the edge counters as indexes
into an object pool (i.e., x64, ARM, MIPS).
BUG=
R=fschneider@google.com , zra@google.com
Review URL: https://codereview.chromium.org//24744002
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@28085 260f80e4-7a28-3924-810f-c04153c831b5
2013-10-01 10:22:47 +00:00
a329b0e6bc
Dart VM: Simplify code generation for equality operators.
...
By inserting the necessary checks for null inside the callee
at the AST level, the code generation of == operations can be
greatly simplified.
This is a performance-neutral change and a step for allowing
generic inlining of arbitrary == methods. So far we could only
inline them for a common set of types in the flow graph
optimizer.
R=srdjan@google.com
Review URL: https://codereview.chromium.org//24203004
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@28084 260f80e4-7a28-3924-810f-c04153c831b5
2013-10-01 10:10:50 +00:00
1fb64d6d95
Cleanup of some comparisons on ARM.
...
R=regis@google.com
Review URL: https://codereview.chromium.org//23622037
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@27574 260f80e4-7a28-3924-810f-c04153c831b5
2013-09-17 16:28:06 +00:00
94ca56a774
Fix a bug in block reordering/block compaction.
...
The compiler assumes (a) the graph entry is followed in the block order by
its normal entry, and (b) the normal entry block is non-empty. There is not
necessarily true, so stop assuming it.
BUG=https://code.google.com/p/dart/issues/detail?id=13101
R=fschneider@google.com
Review URL: https://codereview.chromium.org//23766021
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@27377 260f80e4-7a28-3924-810f-c04153c831b5
2013-09-11 12:32:58 +00:00
0c0bb73f35
Fix large performance regression in Box2D, MD5, and SHA1 benchmarks caused by https://codereview.chromium.org/22915008/
...
R=kmillikin@google.com
Review URL: https://codereview.chromium.org//23684036
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@27288 260f80e4-7a28-3924-810f-c04153c831b5
2013-09-09 12:39:52 +00:00
e571ecdf87
Tests for GuardField length check along with bug fixes
...
R=regis@google.com , srdjan@google.com
Review URL: https://codereview.chromium.org//22915008
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@27279 260f80e4-7a28-3924-810f-c04153c831b5
2013-09-08 21:44:42 +00:00
aa8b03e7d7
For trigonometric functions call to C-libraries: they are faster than x87 operations (3% on Box2D). This also enables quicker trigonometrics on ARM and MIPS. TODO: Make InvokeMathCFunction more flexible so that it can handle both static and instance calls
...
R=zra@google.com
Review URL: https://codereview.chromium.org//22853009
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@27256 260f80e4-7a28-3924-810f-c04153c831b5
2013-09-06 18:14:53 +00:00
40013fe1c0
Simplify compilation of relational operators.
...
Instead of having a separate IL instructions relational comparisons
are built as normal instance calls initially. When optimized, we replace
the instance call with a smi/double/mint comparison instruction.
This enables generic inlining of relational operator calls and simplifies
code generation, too.
Merging comparisons with branches is done in the optimizing compiler's
branch simplification phase.
R=kmillikin@google.com
Review URL: https://codereview.chromium.org//23757016
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@27058 260f80e4-7a28-3924-810f-c04153c831b5
2013-09-03 14:02:53 +00:00
e174a4c18e
Fix a compiler bug caused by Utils::IsPowerOfTwo treating zero as a power of two.
...
The IsPowerOfTwo function is used together with ShiftForPowerOfTwo. Both function
do not work with zero. This caused the optimizing compiler to generate invalid code
for the expression
x ? 0 : 0
where it assumed that if one of the constants is a power-of-two, it can
be computed by (1 << n). We check for 0 in a number of places, but instead
I decided to fix Utils::IsPowerOfTwo itself and remove unnecessary checks
for the zero case.
TEST=tests/language/vm/if_conversion_vm_test.dart, runtime/vm/utils_test.cc
R=kmillikin@google.com
Review URL: https://codereview.chromium.org//23604024
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@27033 260f80e4-7a28-3924-810f-c04153c831b5
2013-09-03 09:28:34 +00:00
97891ce193
If a rethrow is encountered in the catch block, then its handler must be marked that it needs stacktrace. Fix issue r12584.
...
R=iposva@google.com
Review URL: https://codereview.chromium.org//23691024
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@26963 260f80e4-7a28-3924-810f-c04153c831b5
2013-08-31 00:26:47 +00:00
8171570100
Fix the previously ineffective assert checking the number of arguments passed to
...
a runtime entry in the VM.
R=asiva@google.com
Review URL: https://codereview.chromium.org//23672011
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@26826 260f80e4-7a28-3924-810f-c04153c831b5
2013-08-28 23:07:14 +00:00
e06ea63b60
Mark exception handlers if they have a stacktrace specified. Do not build a stacktrace if the handler has no stacktrace.
...
R=asiva@google.com , hausner@google.com
Review URL: https://codereview.chromium.org//23445012
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@26823 260f80e4-7a28-3924-810f-c04153c831b5
2013-08-28 22:42:21 +00:00
00f20f3ffc
Collect edge count profiling data and reorder basic blocks.
...
In unoptimized code, collect edge counts for all edges. Use the counts
to reorder basic blocks in optimized code. Basic blocks are reordered
by using Pettis and Hansen's "Profile Guided Code Positioning" bottom-up
algorithm.
R=fschneider@google.com , regis@google.com , srdjan@google.com
Review URL: https://codereview.chromium.org//18111006
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@26773 260f80e4-7a28-3924-810f-c04153c831b5
2013-08-28 09:17:07 +00:00
4f80ef09df
Change Bool::Get to return a handle instead of a pointer to a raw object.
...
Replace occurrences of 'expr ? Bool::True() : Bool::False()' with simply
'Bool::Get(expr)' and analogously for its negation. The pointer to the raw
object can still be extracted with 'Bool::Get(expr).raw()'.
R=iposva@google.com , srdjan@google.com
Review URL: https://codereview.chromium.org//23480005
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@26770 260f80e4-7a28-3924-810f-c04153c831b5
2013-08-28 08:38:56 +00:00
255a28369c
Fix corner cases for double pow operation.
...
R=regis@google.com , zra@google.com
Review URL: https://codereview.chromium.org//23072044
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@26662 260f80e4-7a28-3924-810f-c04153c831b5
2013-08-26 17:38:51 +00:00
d816dd325f
Adds Uint32x4 add/sub for ARM.
...
R=johnmccutchan@google.com
Review URL: https://codereview.chromium.org//23366003
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@26544 260f80e4-7a28-3924-810f-c04153c831b5
2013-08-22 18:24:20 +00:00
03111ecbeb
Fix performance regression on Tracer: recognize _doublePow in Math.
...
R=johnmccutchan@google.com
Review URL: https://codereview.chromium.org//23068020
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@26470 260f80e4-7a28-3924-810f-c04153c831b5
2013-08-21 20:36:08 +00:00
c8b370ab50
Initial support for length guards on final fields.
...
R=regis@google.com , srdjan@google.com , zra@google.com
Review URL: https://codereview.chromium.org//22851003
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@26433 260f80e4-7a28-3924-810f-c04153c831b5
2013-08-21 15:32:45 +00:00
6d11a8fa15
Add signMask getter to Float32x4
...
R=srdjan@google.com , zra@google.com
Review URL: https://codereview.chromium.org//22947006
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@26394 260f80e4-7a28-3924-810f-c04153c831b5
2013-08-20 21:35:53 +00:00
7c1e9d8298
Add add and subtract operations to Uint32x4
...
R=srdjan@google.com
Review URL: https://codereview.chromium.org//23018009
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@26358 260f80e4-7a28-3924-810f-c04153c831b5
2013-08-20 14:34:06 +00:00
266b9cb305
Add double unary operation (negate). This is a replica of the destrpyed CL https://chromiumcodereview.appspot.com/23219002/
...
Review URL: https://codereview.chromium.org//22999011
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@26169 260f80e4-7a28-3924-810f-c04153c831b5
2013-08-14 23:45:13 +00:00
2793f7ad63
Fix bug with optimized try-catch on ARM/MIPS.
...
The constant pool pointer must be restored before any parallel
moves at the catch-entry block.
In addition, this CL removes CatchEntryInstr and folds its
functionality into CatchBlockEntryInstr. Until now every catch-block
started with CatchBlockEntryInstr followed by a CatchEntryInstr.
This simplifies a lot of code in the compiler and avoids issues
with allocator-move code inserted between the two instructions.
BUG=https://code.google.com/p/dart/issues/detail?id=12291
TEST=tests/language/try_catch4_test.dart
R=regis@google.com , srdjan@google.com
Review URL: https://codereview.chromium.org//22590002
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@25918 260f80e4-7a28-3924-810f-c04153c831b5
2013-08-08 08:20:10 +00:00
srdjan@google.com