| Tue, 16 Jun 2015 18:19:03 +0200 |
Jaroslav Tulach |
Removing com.oracle.truffle.* modules as they have been moved do truffle repository. Modifying suite.py to consume Truffle as a binary.
|
| Tue, 16 Jun 2015 16:30:34 +0200 |
Roland Schatz |
Support derived references in HotSpotReferenceMap.
|
| Tue, 16 Jun 2015 12:19:53 +0200 |
Roland Schatz |
Ignore undefined-bool-conversion warning on clang/linux.
|
| Tue, 16 Jun 2015 16:14:57 +0200 |
Doug Simon |
MatchStatementSet providers are JVMCI providers
|
| Tue, 16 Jun 2015 15:26:43 +0200 |
Jaroslav Tulach |
Include dependencies of a library when it is put into -processorpath
|
| Tue, 16 Jun 2015 11:25:50 +0200 |
Doug Simon |
relax verify_defs_make in jvmci.make for conditional exports
|
| Mon, 15 Jun 2015 11:36:11 +0200 |
Gilles Duboscq |
mx: Add --parsable option to archive
|
| Fri, 12 Jun 2015 16:51:32 +0200 |
Gilles Duboscq |
mx: improve first usage experience:
|
| Thu, 11 Jun 2015 16:17:16 +0200 |
Gilles Duboscq |
mx: Use purpose while selecting jdk for IGV and projects. Make jdk selection messages nicer
|
| Thu, 11 Jun 2015 15:38:42 +0200 |
Gilles Duboscq |
mx: Fix typos and error if mx/env does not exist
|
| Mon, 15 Jun 2015 14:25:29 +0200 |
Stefan Anzinger |
[SPARC] Make frem and drem SSA LIR ready
|
| Mon, 15 Jun 2015 13:49:38 +0200 |
Stefan Anzinger |
Change load in NullCheckOp to ldub which neglects alignment issues
|
| Mon, 15 Jun 2015 13:01:35 +0200 |
Stefan Anzinger |
Don't apply os.path.realpath to the boot classpath
|
| Mon, 15 Jun 2015 13:45:35 +0200 |
Christian Wirth |
Also fix compiler warning (thus, error) on Windows.
|
| Mon, 15 Jun 2015 11:55:38 +0200 |
Stefan Anzinger |
Use real path (canonicalized) when create relative paths for bootclasspath
|
| Thu, 11 Jun 2015 12:21:15 +0200 |
Josef Eisl |
LinearScanWalker#allocLockedRegister: only spill current interval if the first use pos with ShouldHaveRegister is bigger than the next register use pos.
|
| Mon, 08 Jun 2015 15:19:36 +0200 |
Josef Eisl |
LinearScanLifetimeAnalysisPhase: make registerPriorityOfOutputOperand a member and move SSA specific stuff to SSALinearScanLifetimeAnalysisPhase.
|
| Mon, 08 Jun 2015 15:03:25 +0200 |
Josef Eisl |
SSAMoveResolver: support stack2stack-only mappings.
|
| Mon, 08 Jun 2015 17:51:58 +0200 |
Josef Eisl |
mx gate: add UnitTestsSSA task.
|
| Sun, 14 Jun 2015 01:37:31 -0700 |
Chris Seaton |
Restore deleted Graal version tags.
|
| Fri, 12 Jun 2015 17:57:03 -0700 |
Christian Wimmer |
Fix MetaUtil.toInternalName
|
| Fri, 12 Jun 2015 17:01:15 -0700 |
Tom Rodriguez |
Delete some dead code and fixed exact arraycopy check
|
| Sat, 13 Jun 2015 01:32:44 +0200 |
Doug Simon |
fixed C++ compiler warning
|
| Sat, 13 Jun 2015 01:01:21 +0200 |
Doug Simon |
re-enabled checking that jvmci.make is in sync with suite*.py
|
| Sat, 13 Jun 2015 00:59:15 +0200 |
Doug Simon |
fixed pylint issue
|
| Sat, 13 Jun 2015 00:58:03 +0200 |
Doug Simon |
replaced class file patching with use of jre/lib/jvmci/graal.properties to set graal.version system property
|
| Sat, 13 Jun 2015 00:25:30 +0200 |
Doug Simon |
show diff if jvmci.make generation modifies an existing file
|
| Sat, 13 Jun 2015 00:22:48 +0200 |
Doug Simon |
add properties from <jdk>/jre/lib/jvmci/*.properties files to system properties
|
| Fri, 12 Jun 2015 18:02:58 +0200 |
Stefan Anzinger |
Introduce clean target in jvmci.make
|
| Fri, 12 Jun 2015 17:13:45 +0200 |
Andreas Woess |
change phi node name template back to an ascii string to ease searching
|
| Fri, 12 Jun 2015 14:46:18 +0200 |
Andreas Woess |
Truffle: fix delete/removeProperty regression
|
| Fri, 12 Jun 2015 15:23:50 +0200 |
Andreas Woess |
Truffle: remove Shape#copyOverPropertiesInternal
|
| Fri, 12 Jun 2015 17:32:52 +0200 |
Roland Schatz |
Temporarily disable jvmci.make gate check.
|
| Fri, 12 Jun 2015 17:12:08 +0200 |
Roland Schatz |
Remove unused method.
|
| Fri, 12 Jun 2015 16:13:33 +0200 |
Doug Simon |
jvmci.make should have a newline at the end
|
| Fri, 12 Jun 2015 15:47:54 +0200 |
Doug Simon |
Merge.
|
| Fri, 12 Jun 2015 14:53:05 +0200 |
Stefan Anzinger |
Add parameter description to jvmci.make
|
| Fri, 12 Jun 2015 14:52:08 +0200 |
Stefan Anzinger |
Convert addresses in ArrayCopyCallNode to word size
|
| Fri, 12 Jun 2015 14:42:29 +0200 |
Jaroslav Tulach |
Verifying we can use List<Object> via ReadOnlyArrayList in execute(VirtualFrame) without any slowdown.
|
| Fri, 12 Jun 2015 15:34:11 +0200 |
Doug Simon |
fixed subtle makefile evaluation order issue in the verification that only existing JVMCI services and options are exported in make/defs.make
|
| Fri, 12 Jun 2015 15:33:11 +0200 |
Doug Simon |
switched to use of simple makefile assignment in jvmci.make and removed unnecessary semicolons
|
| Fri, 12 Jun 2015 13:19:03 +0200 |
Doug Simon |
merge fix
|
| Fri, 12 Jun 2015 13:06:27 +0200 |
Doug Simon |
Services cannot be sub-classes or instantiated
|
| Fri, 12 Jun 2015 13:05:40 +0200 |
Doug Simon |
don't export com.oracle.jvmci.options.Options as it has no JVMCI providers
|
| Fri, 12 Jun 2015 13:04:35 +0200 |
Doug Simon |
verify that only existing JVMCI services and options are exported in make/defs.make
|
| Fri, 12 Jun 2015 01:25:58 +0200 |
Doug Simon |
don't add command used to generate jvmci.make to the top of jvmci.make as it confuses the gate check that this file is in sync with suite*.py
|
| Fri, 12 Jun 2015 01:19:57 +0200 |
Doug Simon |
replaced Service marker interface with non-standard META-INF directory names to differentiate JVMCI providers from standard service providers
|
| Fri, 12 Jun 2015 01:06:36 +0200 |
Doug Simon |
mx: add check to gate the jvmci.make and suite.py are in sync
|
| Fri, 12 Jun 2015 01:02:57 +0200 |
Doug Simon |
ignore generated NetBeans project configs in jvmci/
|
| Thu, 11 Jun 2015 20:46:30 +0200 |
Christian Humer |
Truffle: direct all log output to GraalTruffleRuntime#log
|
| Fri, 12 Jun 2015 11:32:39 +0200 |
Roland Schatz |
Unit tests for counted loop detection.
|
| Fri, 12 Jun 2015 11:08:45 +0200 |
Roland Schatz |
Fix trip count of counted loops with stride > 1.
|
| Thu, 11 Jun 2015 20:40:11 -0700 |
Christian Wimmer |
Preserve stateBefore when creating newInstanceNode and newArrayNode during canonicalization
|
| Thu, 11 Jun 2015 16:20:13 -0700 |
Christian Wimmer |
Limit inlining depth during partial evaluation to avoid StackOverflowError, provide useful error message instead
|
| Thu, 11 Jun 2015 15:14:14 -0700 |
Christian Wimmer |
Allow subclasses of DynamicNewArrayNode and DynamicNewInstanceNode
|
| Thu, 11 Jun 2015 13:03:54 -0700 |
Tom Rodriguez |
TypeSwitchNode should be more careful about deleting successors
|
| Thu, 11 Jun 2015 12:15:19 -0700 |
Tom Rodriguez |
Map from method name to rule should be per type not per annotation processor
|
| Thu, 11 Jun 2015 09:51:50 -0700 |
Christian Wimmer |
Merge
|
| Wed, 10 Jun 2015 17:23:19 -0700 |
Christian Wimmer |
Move instantiation checks from methods substitutions to the lowering of DynamicNewInstanceNode and DynamicNewArrayNode, so that Unsafe.allocateInstance and Array.newInstance can be intrinsified to a single node.
|
| Thu, 11 Jun 2015 17:16:53 +0200 |
Andreas Woess |
Truffle: minor refactoring
|
