Age | Commit message (Collapse) | Author | Files | Lines |
|
* Fail to explicitly tail call on x86 unix.
* Correctly return 100
* Correct return value
* Add noway assert in morphTailCall to avoid morphing slow tail calls on unix.
* Address feedback
|
|
Skip JIT/superpmi/superpmicollect/superpmicollect.sh
|
|
Lists:
JIT.jit64.opt.cse.hugeSimpleExpr1.hugeSimpleExpr1
GC.Regressions.v2.0-rtm.494226.494226.494226
GC.Scenarios.ServerModel.servermodel.servermodel
JIT.jit64.opt.cse.HugeField1.HugeField1
JIT.jit64.opt.cse.HugeField2.HugeField2
JIT.jit64.opt.cse.hugeexpr1.hugeexpr1
|
|
* Hardcode LKG version of ucrtbased.dll
* Only use hardcoded UCRT version when search path fails
* Add missing single-quotes
|
|
|
|
* Update dependencies from https://github.com/dotnet/corefx build 20190626.2
- Microsoft.Bcl.AsyncInterfaces - 1.0.0-preview7.19326.2
- Microsoft.NETCore.Platforms - 3.0.0-preview7.19326.2
- Microsoft.Private.CoreFx.NETCoreApp - 4.6.0-preview7.19326.2
* Reenable globalization tests
|
|
Currently, corefx testing is done on Windows/x64.
Expand testing to include:
1. Linux/x64, Linux/arm, Linux/arm64, Windows/x86 platforms.
Windows/arm should be added after corefx official builds
publish tests and test manifest. Windows/arm64 should be
added after corefx tests are published, and we have
sufficient hardware.
2. All JIT stress modes that are used in normal coreclr testing.
Testing is split into 4 Azure DevOps Pipelines: coreclr-corefx,
coreclr-corefx-jitstress, coreclr-corefx-jitstressregs, and
coreclr-corefx-jitstress2-jitstressregs.
In addition, we now use the corefx meta-package Microsoft.Private.CoreFx.OOB
instead of a list of individual assemblies when constructing
the testhost.
|
|
PowerShell is useless without this argument in default Windows configurations.
|
|
* Throw unsupported platform for native vcarargs on Arm32 Windows
* Re-enable baseservices vararg test
|
|
digit if one exists. (#25096)
* Updating Dragon4 to ensure the number buffer always provides a significant digit if one exists.
* Changing System.Number.RoundNumber to not round up floating-point numbers.
* Re-enabling the RealFormatterTestsBase CoreFX tests
* Updating Number.RoundNumber to take a isCorrectlyRounded parameter and to use IEEE compliant rounding for floating-point numbers.
* Change SinglePrecisionCustomFormat to 7, ensuring it matches the value used in netcoreapp2.1
|
|
* Remove IFormattable
* Hardcode characters
* Add tests
|
|
|
|
* Update dependencies from https://github.com/dotnet/corefx build 20190621.1
- Microsoft.Bcl.AsyncInterfaces - 1.0.0-preview7.19321.1
- Microsoft.NETCore.Platforms - 3.0.0-preview7.19321.1
- Microsoft.Private.CoreFx.NETCoreApp - 4.6.0-preview7.19321.1
* Disable failing tests
|
|
Fix Issue #25134 - AssertionProp incorrectly removes cast from uint
|
|
(#25276)
* Add EventPipe Processor Number support and make NetTrace the default format
The EventPipe header now has a Processor Number field. On windows we query the correct value, on other OSes we currently have a -1 placeholder, but the field is written to the format regardless.
NetTrace is now the default format when using the environment variable, and the format must be explicitly configured when using the IPC channel or managed API. A parallel change in the diagnostics repo is changing dotnet-trace and dotnet-counter to specify nettrace format which means .NET devs should see nettrace almost exclusively from now on. If for whatever reason it is needed, NetPerf remains available if a scenario explicitly requests to use it.
* PR feedback + attempting to fix broken tests
|
|
Fix #25245
|
|
* Address nullable feedback for System.Diagnostics.Tracing and StackTrace
* Disable corefx tests that need to be updated
|
|
* Disable unloading incompatible test for runincontext
These tests do various things that make them incompatible with running
in an unloadable assembly load context using the runincontext tool.
I've attached comments to the UnloadabilityIncompatible element
describing the specific problems of each test.
I've also found a problem in the parameter passing in the
runincontext.sh where it was always passing 7 parameters to each test
and tests that were checking the number of parameters were failing due
to that. This change fixes it by passing just the parameters that were
passed to the runincontext.sh.
* Added printing unhandled exception details if it happened in the test
* Added 10ms wait to the GC collection loop to also give time to async
stuff to completely exit threadpool threads
|
|
|
|
|
|
|
|
This test was checking for "Parameter name:" substring in the
ArgumentNullException message. But a recent change #25185 modified
that string to just "Parameter".
We also have a coverage for this in corefx tests, so I am removing
this test.
|
|
Update run-corefx-tests.py for new 3.0.0 directory name
|
|
|
|
(#25214)
* Address nullable feedback for System.Threading.Overlapped
* Address nullable feedback for System.Memory
* Address nullable for System.Runtime.InteropServices
* Update xml comments and make CultureInfo nullable to match String.ToLower/Upper feedback
* PR feedback: revert accepting null comparer in BinarySearch
* Disable test that needs to be updated in corefx
|
|
Add additional check for the GT_UNSIGNED flag
+ Ran clang-format
+ Code review feedback, use IsUnsigned()
|
|
* Change ArgumentException message to one line
* Disable 2 tests
* Remove duplication in AggregateException.ToString()
* Inner exceptions on new lines
* Caps letter
* Typo
|
|
Disable failing tests. Also enable tests for changes that propagated through the system.
Fixes dotnet/corefx#38535
|
|
* Fix contained LEA handling
This adds an LEA case to both `LinearScan::BuildOperandUses` and `CodeGen::genConsumeRegs`.
Fix #25039
|
|
First approximation of a fix for #25057.
This has two problems:
* We're checking for any byref-like typed fields. Types that don't actually contain interior pointers but were marked as `ref struct` will fail to load when not aligned properly.
* We're not doing the deep validation that we do for reference types to make sure the `ByReference<T>` field doesn't overlap with another non-byreference field.
Question is whether we're okay with those limitations, or whether we need a better fix. Better fix would likely entail inefficiently walking over the fields à la `FindByRefPointerOffsetsInByRefLikeObject` (doing the more efficient thing that we do for object references below would require a GCDesc representation of byrefness).
Contributes to #25057.
|
|
* Allow PEImageLayout to be loadable for R2R images loaded from streams.
* Add regression test coverage
* Change CheckILOnly to return true for R2R images (similar to the ILOnly API)
* Early out for R2R images
* Remove R2R check
|
|
* Refactor IJW test infra. Add some tests for some interesting simple native varargs.
* Add missing include of IjwHelper.cs
* Fix Cmake test project name.
* PR Feedback.
* Fix compilation error.
* Deploy msvcp*d.dll as part of the copy-local'd CRT.
|
|
* IndexOfAny(values.Length == 0) == -1
* Skip tests
|
|
(#20900)
Fix an issue that is reproduced when:
1) a method requires GSSecurityCookie (e.g. when allocating unsafe buffer using localloc) and
2) one of the method parameters is "vulnerable" and requires a shadow variable to be created and used instead of the original parameter variable and
3) the method parameter has small signed integer type (i.e. int8 or int16) and
4) the value of the parameter is converted to unsigned type (e.g. uint8 or uint16).
|
|
* abort build-tests if failed to restore stress dependencies
* Fix arm condition.
* Add arm64 condition.
* Skip runtime dependencies for arm32/arm64.
|
|
|
|
It doesn't work on OSX and Alpine.
It shows random failures on Windows as well, that we do not want to see in our PR/CI outerloop testing.
|
|
There was a typo for Unix (it is COMPlus_GCStress, not COMPlus_GcStress, but COMPlus_GcStressOnDirectCalls was correct). That was why the test passed on OSX.
|
|
* Enable coreclr-runincontext in azure pipelines
|
|
If a call site fails tail call stress validation, don't consider it for
implicit tail calling either.
In normal jitting we defer this level of validation until impImportCall
to avoid duplicating work.
This avoids an assert when we have invalid IL.
Fixes #25027.
|
|
This test creates a native thread on which it calls a managed callback. It uses
std::thread to create the thread. The problem is that on MUSL based Linux distros,
the default stack size for secondary threads is 80kB, which is not enough for this
test.
Since std::thread has no way to set the thread stack size, the fix is to use
pthreads on Unix and keep the std::thread for Windows only.
|
|
CoreFX tests are now used to test CoreCLR on Windows/x64 and Linux/x64.
PR testing includes normal and no-tiered modes. CI testing will include
all JIT stress modes that apply to CoreCLR testing. Many tests have been
disabled to allow the jobs to (mostly) run error-free.
Support for other platforms requires fixes to the testhost generation.
|
|
|
|
* Switch PInvokes into libcoreclr to QCalls
* Register QCalls in runtime. Remove pal exports from libcoreclr
* Alphabetize new ecalllist.h entries
* Remove W suffixes from native side of QCall
* Interop signatures are in nested classes (which doesn't show up here) in the root namespace. So we don't want to pass a namespace in here
* Convert the interop test in R2R to use a local native library instead of calling into the CoreCLR PAL via DllImport.
|
|
parameters (#25054)
* fix generic attribute bug
* add missing test cases
|
|
* Debug.Print should call Debug.WriteLine
- to behave similar to .NET Framework
* temporarily disable tests
- Waiting for coreclr/fx changes to propagate
|
|
csc.exe ourselves. (#24342)
* Use CMake's C# support to build DacTableGen instead of manually invoking csc.exe ourselves.
* Fix x86 failures.
* Disable DAC generation when building with NMake Makefiles and issue an error since the CMake C# support is VS-only. We don't actually support building with NMake (only configure) so this is ok.
* Clean up rest of the macro=1's
PR Feedback.
* Fix Visual Studio generator matching.
* Explicitly specify anycpu32bitpreferred for DacTableGen so the ARM64 build doesn't accidentally make it 64-bit
* Fix bad merge
|
|
* Add COM object lifetime control tests to CoreCLR.
* Combine cleanup control test into NETClientPrimitives.
|
|
|
|
I was not able to repro any failures, maybe they were fixed already. Lets see what ci thinks.
|