Rahul Kumar [Thu, 23 Mar 2017 08:13:54 +0000 (01:13 -0700)]
Merge pull request #10064 from rahku/cgroup
memory limit in containers on linux
rahul [Thu, 9 Mar 2017 02:55:24 +0000 (18:55 -0800)]
memory limit in containers on linux
Gaurav Khanna [Thu, 23 Mar 2017 06:25:57 +0000 (23:25 -0700)]
Various fixes for generating portable RID packages (#10409)
Luis G [Thu, 23 Mar 2017 05:51:45 +0000 (02:51 -0300)]
Add handling of multiple CMake installations in probe-win (#9980)
* Add handling of multiple CMake installations in probe-win
Check if multiple CMake versions are installed, and use by default the first path available
Fixes #8221
* Fix typo in value accessor
* Restore previous null check in probe_win
* Remove redundant assignment
Tanner Gooding [Thu, 23 Mar 2017 05:44:40 +0000 (22:44 -0700)]
Removing the special handling in classlibnative for atan2(±∞, ±∞) and pow(-1.0, ±∞). (#10295)
* Removing the special handling in classlibnative for atan2(±∞, ±∞) and pow(-1.0, ±∞).
* Fixing up the logic for HAVE_COMPATIBLE_POW in the PAL layer.
Jan Kotas [Thu, 23 Mar 2017 05:43:36 +0000 (22:43 -0700)]
Fix fallout from switching corelib to use resx (#10407)
Fixes #https://github.com/dotnet/corefx/issues/17413
Gaurav Khanna [Thu, 23 Mar 2017 04:35:24 +0000 (21:35 -0700)]
Unify command line arg for portable RID build (#10408)
Jon Hanna [Thu, 23 Mar 2017 04:25:58 +0000 (04:25 +0000)]
Named Multiply and Divide methods on TimeSpan (#10366)
dotnet/corefx#16476
Swaroop Sridhar [Wed, 22 Mar 2017 22:22:46 +0000 (15:22 -0700)]
Add reflect.xml for a abstract* tests
For the abstract* tests in question:
The tests check for TypeLoad exception from a derived class that doesn't implement an abstract method in base class.
However, the Linker removed the abstract method from base, so there is no exception.
Therefore, add reflect.xml to override the linker's behavior and enable testing for TypeLoadException.
Swaroop Sridhar [Wed, 22 Mar 2017 22:22:04 +0000 (15:22 -0700)]
Exclude some tests from Linker testing.
Exclude some tests while testing via ILLINK.
The reasons are mentioned in issues.targets.
Swaroop Sridhar [Mon, 20 Mar 2017 18:20:12 +0000 (11:20 -0700)]
Linker Testing: Use a central target for copying reflection roots.
Copy the test.refect.xml file (if one exists) as part of the
AfterBuild job -- so that individual tests need not identify
the CopyReflectionRoots explicitly as a default target.
Drew Scoggins [Thu, 23 Mar 2017 00:21:13 +0000 (17:21 -0700)]
Merge pull request #9947 from DrewScoggins/xunitApi
Update CoreCLR to use xunit perf api
Sean Gillespie [Wed, 22 Mar 2017 23:39:14 +0000 (16:39 -0700)]
Opt-out of GC Stress for FEATURE_STANDALONE_GC (#10391)
Bruce Forstall [Wed, 22 Mar 2017 23:36:53 +0000 (16:36 -0700)]
Jump stubs (#10398)
Add document on CLR Jump Stubs
Atsushi Kanamori [Wed, 22 Mar 2017 22:28:14 +0000 (15:28 -0700)]
Move Type.IsSerializable and Type.DefaultBinder to shared file. (#10395)
This will pave the way for the mirrorbot to fix
https://github.com/dotnet/corert/issues/3061.
Jonghyun Park [Wed, 22 Mar 2017 22:25:19 +0000 (07:25 +0900)]
[x86/Linux] 16-byte aligned TheUMEntryPrestub (#10376)
Gaurav Khanna [Wed, 22 Mar 2017 20:54:41 +0000 (13:54 -0700)]
Switch tests to use OSX.10.12 RID (#10394)
Gaurav Khanna [Wed, 22 Mar 2017 20:30:49 +0000 (13:30 -0700)]
Enable Portable Windows RID (#10365)
* Enable Portable RID support for Windows and OSX.
Drew Scoggins [Tue, 21 Mar 2017 21:01:03 +0000 (14:01 -0700)]
Add updated timeout
Drew Scoggins [Fri, 3 Mar 2017 22:51:37 +0000 (14:51 -0800)]
Update CoreCLR to use xunit perf api
This updates the CoreCLR performance runs to use the new xunit
performance api. This will unblock several people who were wanting to
move tests forward from netstandard1.4 which is were we have been locked
because of using the old desktop runner.
Atsushi Kanamori [Wed, 22 Mar 2017 17:51:46 +0000 (10:51 -0700)]
Reconcile the "easy" System\Reflection .cs files with their CoreRT versions. (#10387)
Files to be moved the shared partition (after CoreRT side is in)
AmbiguousMatchException.cs
AssemblyAttributes.cs
AssemblyContentType.cs
AssemblyNameFlags.cs
Binder.cs
BindingFlags.cs
CallingConventions.cs
CustomAttributeFormatException.cs
DefaultMemberAttribute.cs
EventAttributes.cs
ExceptionHandlingClauseOptions.cs
FieldAttributes.cs
GenericParameterAttributes.cs
ICustomAttributeProvider.cs
ImageFileMachine.cs
InterfaceMapping.cs
IntrospectionExtensions.cs
InvalidFilterCriteriaException.cs
IReflectableType.cs
ManifestResourceInfo.cs
MemberFilter.cs
MemberTypes.cs
MethodAttributes.cs
MethodImplAttributes.cs
Missing.cs
ModuleResolveEventHandler.cs
ParameterAttributes.cs
ParameterModifier.cs
PortableExecutableKinds.cs
ProcessorArchitecture.cs
PropertyAttributes.cs
ReflectionContext.cs
ReflectionTypeLoadException.cs
ResourceAttributes.cs
ResourceLocation.cs
StrongNameKeyPair.cs
TargetException.cs
TargetInvocationException.cs
TargetParameterCountException.cs
TypeAttributes.cs
TypeFilter.cs
Hyeongseok Oh [Wed, 22 Mar 2017 17:32:11 +0000 (02:32 +0900)]
Change default CI job for ARM and clean-up (#10370)
- Change ARM CI job triggered by PR: ubuntu16.04 debug, ubuntu release, tizen debug/release
- Clean-up: remove emulator build/test
- Change ubuntu16.04 x64 default job name and trigger phrase: conflict triggerring with arm
Sean Gillespie [Wed, 22 Mar 2017 17:25:08 +0000 (10:25 -0700)]
Fix an issue where GCStress allocated objects using the Gen 0 alloc context (#10322)
* Fix an issue where GCStress allocated objects using the Gen 0 alloc context
* Remove a GCStress MaybeTrigger on LOH allocations that attempted to use the gen 0 alloc context
Andy Ayers [Wed, 22 Mar 2017 17:23:56 +0000 (10:23 -0700)]
Merge pull request #10371 from AndyAyersMS/FixDevirt
Fix several devirtualization issues
Gaurav Khanna [Wed, 22 Mar 2017 15:55:53 +0000 (08:55 -0700)]
Move OSX packages to use 10.12 as the RID (#10354)
Jan Kotas [Wed, 22 Mar 2017 15:03:09 +0000 (08:03 -0700)]
Merge pull request #10374 from dotnet/mirror-merge-9057970
Mirror changes from dotnet/corert
Atsushi Kanamori [Wed, 22 Mar 2017 11:37:12 +0000 (04:37 -0700)]
Move DefaultBinder.cs to the shared partition. (#10355)
Jeremy Kuhne [Tue, 21 Mar 2017 23:07:53 +0000 (16:07 -0700)]
FileStream redirection got changed. Add ThreadPoolBoundHandle redirect.
https://devdiv.visualstudio.com/DefaultCollection/DevDiv/_workitems?id=398142&_a=edit
[tfs-changeset: 1651640]
Stephen Toub [Wed, 22 Mar 2017 06:09:22 +0000 (02:09 -0400)]
Fix leak in Task.WaitAny(..., int) (#10331)
WaitAny is effectively built on top of WhenAny, creating a continuation from the supplied tasks and then blocking on that continuation. When a timeout is provided, it blocks with that timeout. But if it doesn't complete within the timeout, it ends up leaking the continuations it created into the constituent tasks. The fix is simply to force the returned continuation to complete, such that its continuation logic does all of the appropriate cleanup.
dotnet bot [Wed, 22 Mar 2017 05:04:42 +0000 (22:04 -0700)]
Update CoreClr, CoreFx to beta-25121-02, beta-25122-01, respectively (#10310)
Alex Perovich [Wed, 22 Mar 2017 05:02:47 +0000 (00:02 -0500)]
Switch coreclr corelib to use resx (#10372)
* Revert "Revert "Switch coreclr corelib to use resx (#10268)""
* Add resources.target override
This fixes the GenerateResource task running on the arm emulator runs.
* Fix run.sh to have a working Tools-Override
Sean Gillespie [Wed, 22 Mar 2017 05:00:12 +0000 (22:00 -0700)]
[Local GC] Forbid inclusion of gcscan.h from VM and DAC directories (#10332)
* Forbid inclusion of gcscan.h from VM and DAC directories
* Address code review feedback - hoist IsGCHeapInitialized check to an assert, since the heap should definitely be initialized if we are validating objects that ostensibly came from it
Andy Ayers [Tue, 21 Mar 2017 20:17:28 +0000 (13:17 -0700)]
Fix several devirtualization issues
When doing interface devirtualization, if the object type is canonical,
ensure that the owner type is too. The jit may present a mixed set
when inlining a shared method into a non-shared method. Ideally the
jit would also be able to present exact object types in such cases but
currently it cannot guarantee this. Closes #10311.
Adjust contracts to address some contract violations seen in desktop
testing. Make the helper non-static and fold in some of the info that
was passed from the caller to bring the desktop and CoreCLR implementations
closer.
Disallow interface devirt if the method is final but the class is not
exact or final, since derived classes can still override final methods
when implementing interfaces.
Don't try and devirtualize interface calls from com objects.
Add some related test cases.
Jan Kotas [Wed, 22 Mar 2017 00:45:57 +0000 (17:45 -0700)]
Merge pull request #10363 from jkotas/revert
Revert "Switch coreclr corelib to use resx (#10268)"
Jan Kotas [Tue, 21 Mar 2017 23:24:52 +0000 (16:24 -0700)]
Revert "Switch coreclr corelib to use resx (#10268)"
This reverts commit
12ef04c22b07f1e4ec5a63009298e4537a2c8e1a.
Bruce Forstall [Tue, 21 Mar 2017 22:27:07 +0000 (15:27 -0700)]
Fix Linux/x86 call alignment calculation and insertion (#10266)
* Fix Linux/x86 call alignment calculation and insertion
The existing system of calculating per-call alignment adjustments
did not consider the possibility of nested calls. On x86, some
arguments for a call might be pushed on the stack before a
nested call's arguments start being pushed. Thus, a call can't
consider only its own arguments when determining the stack level.
Instead, use the existing genStackLevel variable, updated
dynamically during code generation, to determine the baseline
stack alignment when a call first needs to push something on
the stack, or if it has no stack arguments, whether it needs to
align the stack before a call.
One wrinkle here is that the fgAddCodeRef() function for adding
throw blocks for array bounds checks, and other similar helpers,
need to know the stack level on entry, and this stack level is
computed during argument morphing. There is a tough phase ordering
problem here. So, we bail out and force an EBP frame if there are
any such throw helpers in the function. When using an EBP frame,
the helper block doesn't need to know the stack level -- it is
only needed for ESP frames, needed for unwinding. Note that this
bail out already existed if the same helper needed multiple different
stack levels on entry. We could do slightly better without too much
work by not bailing out for top-level calls with no stack arguments.
* Track max alignment added for nested calls during codegen
This information is needed by Linux/x86 for an assert about
maximum emitter stack depth.
* Make max stack align tracking available in release as well
The assert these are used in is a noway_assert.
* Formatting
Hanjoung Lee [Tue, 21 Mar 2017 22:00:56 +0000 (07:00 +0900)]
[x86/Linux] Invoke gcResetForBB() in genFuncletProlog (#10186)
Andy Ayers [Tue, 21 Mar 2017 20:08:01 +0000 (13:08 -0700)]
Merge pull request #10343 from parjong/fix/self_override
Fix bogus assert in resolveVirtualMethodHelper and update related ilproj
Bruce Forstall [Tue, 21 Mar 2017 20:01:52 +0000 (13:01 -0700)]
Add /silent option to crossgen (#10350)
This sets the NGenOptions.fSilent flag, which prevents displaying
the final output message.
This is useful for JIT asm diff generation.
Alex Perovich [Tue, 21 Mar 2017 19:55:39 +0000 (14:55 -0500)]
Switch coreclr corelib to use resx (#10268)
* Switch to resx
* Use roslyn rewriter to switch from GetResourceString to SR
* More GetResourceString changes
* Add missing resource
* Remove Environment.GetResourceString and replace final usages with SR
* Remove comment
* Fix spacing
* Replace final instances of Environment.GetResourceString
* Add another missing resource
* Add back Environment.GetResourceStringLocal because the runtime needs it
Viktor Hofer [Tue, 21 Mar 2017 19:36:25 +0000 (20:36 +0100)]
System.IO.Path.IsPathRooted does not check if valid drive letter on Windows (#10323)
Aditya Mandaleeka [Tue, 21 Mar 2017 19:01:53 +0000 (12:01 -0700)]
Merge pull request #10254 from sdmaclea/PR-FIX-SCRIPT-ERROR
Fix script ERRORLEVEL typo
Koundinya Veluri [Tue, 21 Mar 2017 18:47:32 +0000 (11:47 -0700)]
Temporarily disable frequently failing Sleep PAL tests (#10313)
See #8321
Ian Hays [Tue, 21 Mar 2017 17:37:51 +0000 (10:37 -0700)]
Merge pull request #10321 from WinCPP/issue-7060
Issue #7060 - Make StreamReader.ReadBuffer non-virtual
Joseph Tremoulet [Tue, 21 Mar 2017 15:06:24 +0000 (11:06 -0400)]
Merge pull request #10327 from mikedn/assertion-info
Refactor GenTree assertion information storage
Jonghyun Park [Tue, 21 Mar 2017 09:31:04 +0000 (18:31 +0900)]
[x86/Linux] Log UpdateRegDisplay calls (#10335)
Jonghyun Park [Tue, 21 Mar 2017 06:10:30 +0000 (15:10 +0900)]
Fix bogus assert in resolveVirtualMethodHelper and update related ilproj
Jonghyun Park [Tue, 21 Mar 2017 04:36:27 +0000 (13:36 +0900)]
[x86/Linux] Use 4-byte align for double in struct (#10340)
Jonghyun Park [Tue, 21 Mar 2017 04:26:19 +0000 (13:26 +0900)]
[x86/Linux] Correct update on ResumableFrame (#10339)
Jan Kotas [Tue, 21 Mar 2017 04:13:04 +0000 (21:13 -0700)]
Merge pull request #10341 from dotnet/mirror-merge-9056283
Mirror changes from dotnet/corert
Jan Kotas [Tue, 21 Mar 2017 02:30:56 +0000 (19:30 -0700)]
Delete moved file
Atsushi Kanamori [Mon, 20 Mar 2017 21:01:10 +0000 (14:01 -0700)]
NS2.0 - Complete System.Security namespace.
- Add GetObjectData() implementation to System.SecurityExcepton
- Clean up a bit and move to shared partition.
Jonghyun Park [Tue, 21 Mar 2017 00:14:02 +0000 (09:14 +0900)]
'root' as a special case in UTIL_IsExecuteBitsSet (#10246)
Tarek Mahmoud Sayed [Mon, 20 Mar 2017 23:19:48 +0000 (16:19 -0700)]
Fix MAC build (#10328)
This is temporary fix to unblock the CI and coreclr building for MAC. this change should be reverted after having the proper fix for how we load ICU on MAC
Jan Kotas [Mon, 20 Mar 2017 23:17:55 +0000 (16:17 -0700)]
Merge pull request #10325 from dotnet/mirror-merge-9055987
Mirror changes from dotnet/corert
Aditya Mandaleeka [Mon, 20 Mar 2017 23:01:13 +0000 (16:01 -0700)]
Merge pull request #10330 from omajid/fix-markdown-atx-headers
Fix markdown headings
Omair Majid [Mon, 20 Mar 2017 22:12:33 +0000 (18:12 -0400)]
Fix markdown headings
GitHub recently changed how atx headings (beginning with `#`) are
rendered. A space is now required between `#` and the following text
for the heading to be recognized.
Update headings in the docs to match the expected format.
See https://github.github.com/gfm/#atx-headings
Jan Kotas [Mon, 20 Mar 2017 21:28:18 +0000 (14:28 -0700)]
Delete moved files
Mike Danes [Sat, 11 Mar 2017 09:04:04 +0000 (11:04 +0200)]
Refactor GenTree assertion information storage
Stealing a bit from gtAssertionNum is ugly and error prone (e.g. bool HasAssertion() { return gtAssertionNum != 0; } is incorrect if the edge bit is set).
Fix another per code review
Joseph Tremoulet [Mon, 20 Mar 2017 21:13:14 +0000 (17:13 -0400)]
Merge pull request #10319 from JosephTremoulet/NullField
Permit null object in `GenTreeUseEdgeIterator`
Atsushi Kanamori [Mon, 20 Mar 2017 17:42:17 +0000 (10:42 -0700)]
Move Type.cs and Module.cs to Shared Partition.
Types.cs is split into the main file, Type.Helpers.cs and Type.Enum.cs.
All three will be shared.
- GenericTypeArguments - reconcile with CoreClr
(the code is identical provided that the underlying
Type provider implemented IsConstructedGenericType
correctly. But we can't assume that for third party types.)
- IsInterface and IsSerializable has some CoreClr
specific logic for RuntimeTypes - moving this out
to the unshared space for now.
- IsContextFulImpl()/IsMarshalByRefImpl() - corrected
implemention to match CoreClr.
- IsIntegerType - no CommonRuntimeTypes in CoreClr
so going back to typeof() checks - the runtime implemented
types override all this enum code anyway on both runtimes.
- GetRootElementType() - upgraded visibility to "internal"
as a concession to CoreClr.
- CoreCLR's IsVisible has a fast path - we'll keep that.
- Moved IsRuntimeImplemented() out of Internal.Reflection.Core.NonPortable
namespace so calls to it can be more easily shared.
- Module.MethodHandle was not implemented correctly for
non-runtime modules (it's supposed to return default(ModuleHandle)).
This makes it behave correctly and reconciles Module.cs
with the CoreCLR version.
Jan Kotas [Mon, 20 Mar 2017 21:03:14 +0000 (14:03 -0700)]
Merge pull request #10317 from dotnet/mirror-merge-9055826
Mirror changes from dotnet/corert
WinCPP [Mon, 20 Mar 2017 19:58:55 +0000 (01:28 +0530)]
Issue #7060 - Make StreamReader.ReadBuffer non-virtual
Joseph Tremoulet [Mon, 20 Mar 2017 19:39:46 +0000 (15:39 -0400)]
Permit null object in `GenTreeUseEdgeIterator`
We use `GT_FIELD` for static field loads, which have no instance pointer.
Update the iterator constructor to recognize this as a valid state rather
than fail an assertion.
Bruce Forstall [Mon, 20 Mar 2017 19:09:36 +0000 (12:09 -0700)]
Merge pull request #10315 from dotnet-bot/from-tfs
Merge changes from TFS
Jan Kotas [Sun, 19 Mar 2017 02:41:39 +0000 (19:41 -0700)]
Fix FileStream for ProjectN
[tfs-changeset: 1651325]
Bruce Forstall [Mon, 20 Mar 2017 18:45:47 +0000 (11:45 -0700)]
Remove message about falling to default for MethodContext::repGetEEInfo()
[tfs-changeset: 1651465]
Brian Sullivan [Mon, 20 Mar 2017 17:34:10 +0000 (10:34 -0700)]
Merge pull request #10301 from dotnet-bot/from-tfs
Merge changes from TFS
Gaurav Khanna [Mon, 20 Mar 2017 17:32:34 +0000 (10:32 -0700)]
Initial document on AssemblyLoadContext Design (#10271)
* Initial document on AssemblyLoadContext Design
Brian Sullivan [Mon, 20 Mar 2017 17:24:17 +0000 (10:24 -0700)]
Merge pull request #10262 from briansull/refactor2
Updates for the handling of Profile data in the JIT
Atsushi Kanamori [Mon, 20 Mar 2017 17:03:26 +0000 (10:03 -0700)]
Make GetModuleHandle() sharable. (#10309)
Gaurav Khanna [Mon, 20 Mar 2017 15:55:56 +0000 (08:55 -0700)]
Merge pull request #9445 from hqueue/arm/ci
[ARM/CI] ARM CI script for arm and armel
Atsushi Kanamori [Mon, 20 Mar 2017 15:18:45 +0000 (08:18 -0700)]
Reconciles Type.cs with CoreRT version for move to shared partition. (#10306)
* Clone files and add resource strings.
* Subset each partial class.
* String=>string and Environment.GetResourceString => SR.cs
Doing this upfront with reduce the upcoming diffs - hopefully.
* Add IsRuntimeImplemented() emulator.
* One method was put in the wrong file. Correcting.
* Converted Type.Enum.cs to CoreRT style member by member.
* Converted Type.Helpers.cs to CoreRT style member by member.
* Internalize __Filter.cs into Type.Helpers.cs
* Pretransform to reduce diffs.
Removed contracts and comments,
"abstract public" -> "public abstract"
* Converted Type.cs to CoreRT style member by member.
* Eh.. rather not share IsInterface and IsSerializable than have that #if CORECLR.
* Transplant the CoreRt files (now just a reordering.)
Steve MacLean [Mon, 20 Mar 2017 15:17:05 +0000 (11:17 -0400)]
[Arm64/Unix] Set PROCFS_* available in tryrun (#10255)
Bruce Forstall [Mon, 20 Mar 2017 02:46:08 +0000 (19:46 -0700)]
Desktop side of the change to remove COR_JIT_EE_VERSION
1. Bring over the JIT-EE interface file changes,
2. Stop building ctpjit.dll, and delete its build rules,
3. Stop building skipjit.dll.
[tfs-changeset: 1651372]
Tarek Mahmoud Sayed [Sun, 19 Mar 2017 23:19:56 +0000 (16:19 -0700)]
Invariant globalization (#10264)
* Invariant Globalization Work
* Convert the testing Exceptions to asserts
* Remove un-needed comment
* Fix typos
* Fix unrelated typo
* Address the PR feedback
* More feedback addressing
* More feedback addressing
* Fix Linux break
* More feedback addressing
* cleanup
Stephen Toub [Sun, 19 Mar 2017 12:06:20 +0000 (08:06 -0400)]
Merge pull request #10291 from dotnet-bot/master-UpdateDependencies
Update CoreClr to beta-25119-02 (master)
dotnet-bot [Sun, 19 Mar 2017 08:08:34 +0000 (08:08 +0000)]
Update CoreClr to beta-25119-02
Bruce Forstall [Sun, 19 Mar 2017 06:35:05 +0000 (23:35 -0700)]
Fix code to build with older VC++ compiler (#10281)
John Bottenberg [Sun, 19 Mar 2017 04:33:58 +0000 (21:33 -0700)]
Remove potential overflow in SafeBSTRHandle (#10182)
Jonghyun Park [Sun, 19 Mar 2017 04:32:57 +0000 (13:32 +0900)]
Fix a typo in exceptoinhandling.cpp (#10237)
Maoni Stephens [Sun, 19 Mar 2017 04:30:55 +0000 (21:30 -0700)]
I hoisted the accounting for LOH into allocate_large_objects to avoid missing counting in certain stages during bgc when we go through bgc_loh_alloc_clr instead of adjust_limit_clr (#10284)
Atsushi Kanamori [Sat, 18 Mar 2017 22:15:43 +0000 (15:15 -0700)]
Fix for Module.GetMethod(string) (#10283)
Stephen Toub [Sat, 18 Mar 2017 21:40:09 +0000 (17:40 -0400)]
Merge pull request #10280 from dotnet-bot/master-UpdateDependencies
Update CoreClr to beta-25118-03 (master)
Jan Kotas [Sat, 18 Mar 2017 20:57:24 +0000 (13:57 -0700)]
Merge pull request #10272 from dotnet/mirror-merge-9051754
Mirror changes from dotnet/corert
Joseph Tremoulet [Sat, 18 Mar 2017 18:06:54 +0000 (14:06 -0400)]
Merge pull request #10216 from JosephTremoulet/SpanBench
Update SpanBench test
dotnet-bot [Sat, 18 Mar 2017 16:37:23 +0000 (16:37 +0000)]
Update CoreClr to beta-25118-03
Joseph Tremoulet [Thu, 16 Mar 2017 00:43:01 +0000 (20:43 -0400)]
Update SpanBench test
- Re-enable the tests that were disabled in
e859c309.
- Re-work how tests are invoked from the command line to require less
boilerplate
- Now each test has a single [Benchmark] entrypoint
- That entrypoint invokes its test's single inner-loop by wrapping it
in a lambda that it passes to a new helper Invoke method (shared
across all tests) which handles the xunit vs. command-line
differences..
- Main finds the entrypoints by using reflection to search for the
[Benchmark] attributes, so the explicit list of stringified test
names is no longer needed and the command line will run the same
set of tests that xunit-perf does.
- The new SpanAPI tests now get invoked when this tests is run from the
command line as well.
- Add [NoInlining] to the SpanAPI tests' kernels.
- Add some heap writes and conditional writes to prevent deadcode
optimization from eliminating tests' kernels (assuming we don't do
interprocedural constant propagation or deadcode, or store sinking and
final value calculation, or PRE...).
- Split the Index SpanAPI tests into one version that uses a
loop-invariant index and thus should get hoisted out of the loop, and
another version that uses variant indices.
Pat Gavlin [Sat, 18 Mar 2017 15:41:40 +0000 (08:41 -0700)]
Merge pull request #10054 from pgavlin/GenTreeUseEdgeIterator
Improve GenTreeUseEdgeIterator.
Jan Kotas [Sat, 18 Mar 2017 15:22:40 +0000 (08:22 -0700)]
Delete moved files
Bruce Forstall [Sat, 18 Mar 2017 15:12:52 +0000 (08:12 -0700)]
Remove JIT-EE version ifdefs (#10273)
* Remove COR_JIT_EE_VERSION define
Currently, we maintain two versions of the JIT-EE interface in the
source tree: the current version, and .NET 4.6 version. This is
managed by the `COR_JIT_EE_VERSION` define. Any new changes to the JIT-EE
interface must be put under `#if COR_JIT_EE_VERSION > 460`. As of now,
there are quite a few of these `#if` statements in the tree.
We started doing this so we could build and ship a CTP (Community
Technology Preview) version of the JIT from the current source tree that
runs against earlier versions of the runtime that are widely available
to customers, so those customers can try out the new JIT and give us
early feedback. This was done during the RyuJIT/x64 bring-up before
shipping .NET 4.6, to get as much feedback as possible before replacing
JIT64. After 4.6 shipped, we re-based the old version of the JIT-EE
interface to the 4.6 version, and started adding the above `#if`
conditions, with the thought that if we ever shipped another CTP, it
would be expected to run on top of 4.6. We have not shipped such a
thing, and we currently have no plans to build another CTP in the same
way we did before. (Early adopters can try out the "live" JIT in the
.NET Core open source project, although that JIT can't be used on the
desktop.)
So, to simplify the code, and reduce the cost of adding to or changing
the JIT-EE interface, we are removing the `COR_JIT_EE_VERSION` define.
Stephen Toub [Sat, 18 Mar 2017 12:32:46 +0000 (08:32 -0400)]
Merge pull request #10278 from dotnet-bot/master-UpdateDependencies
Update CoreClr to beta-25118-02 (master)
Atsushi Kanamori [Sat, 18 Mar 2017 11:37:06 +0000 (04:37 -0700)]
Move ParameterInfo.cs and Assembly.cs to shared partition. (#10270)
* Clone the files.
* Split the members between shared and nonshared files.
* Replaced ParameterInfo with CoreRt text member by member.
* Replaced Module.cs with CoreRt text member by member.
* Remove dead filter code.
* Replaced Assembly.cs with CoreRt text member by member.
* Transplanted CoreRt files.
This is a reordering without any other changes.
Verified by sorting the old and new files and
diffing.
* Migrated ParameterInfo.cs and Assembly.cs to shared directory.
Leaving Module.cs in place for now as one line is going to
have to change over in CoreRt land.
ModuleHandle ModuleHandle => new ModuleHandle(this);
* Cleanup.
* Move the final internal methods to RuntimeParameterInfo.
dotnet-bot [Sat, 18 Mar 2017 08:56:56 +0000 (08:56 +0000)]
Update CoreClr to beta-25118-02
Bruce Forstall [Sat, 18 Mar 2017 07:18:26 +0000 (00:18 -0700)]
Merge pull request #10269 from BruceForstall/FixBuildAll
Fix build.cmd "all" to work with pass-through args again
dotnet bot [Sat, 18 Mar 2017 06:27:38 +0000 (23:27 -0700)]
Update CoreClr, CoreFx to beta-25117-08, beta-25118-01, respectively (#10267)
Pat Gavlin [Thu, 9 Mar 2017 02:45:17 +0000 (18:45 -0800)]
Improve GenTreeUseEdgeIterator.
Instead of executing a rather long set of conditionals on each call
to `GenTreeUseEdge::operator++`, decide on the function that will be
used to advance the iterator's state in the constructor.
Gaurav Khanna [Sat, 18 Mar 2017 04:35:47 +0000 (21:35 -0700)]
Merge pull request #10220 from dotnet/revert-10206-MultiQueue
Revert "Add support for Multiqueing jobs in Helix"
Andy Ayers [Sat, 18 Mar 2017 04:04:56 +0000 (21:04 -0700)]
Merge pull request #10274 from AndyAyersMS/FixMoreCasing
Fix yet another casing issue.