Adam Sitnik [Wed, 16 Dec 2020 15:57:33 +0000 (16:57 +0100)]
re-enable GetProcessesByName_ProcessName_ReturnsExpected test which used to fail on Nano (#46140)
Viktor Hofer [Wed, 16 Dec 2020 14:35:58 +0000 (15:35 +0100)]
Sort stable versions in pkgindex (#46135)
Adam Sitnik [Wed, 16 Dec 2020 14:27:18 +0000 (15:27 +0100)]
Add a few tests for Env Vars edge cases (#46039)
Maxim Lipnin [Wed, 16 Dec 2020 13:31:10 +0000 (16:31 +0300)]
Use MonoProjectRoot in the Mono sample project references (#46129)
dotnet-maestro[bot] [Wed, 16 Dec 2020 13:19:12 +0000 (14:19 +0100)]
[master] Update dependencies from dotnet/runtime (#45387)
* Update dependencies from https://github.com/dotnet/runtime build
20201130.7
runtime.native.System.IO.Ports , Microsoft.NETCore.ILAsm , Microsoft.NET.Sdk.IL , Microsoft.NETCore.DotNetHost , Microsoft.NETCore.DotNetHostPolicy , System.Text.Json
From Version 5.0.0-alpha.1.19563.3 -> To Version 6.0.0-alpha.1.20580.7
* Update dependencies from https://github.com/dotnet/runtime build
20201201.3
runtime.native.System.IO.Ports , Microsoft.NETCore.ILAsm , Microsoft.NET.Sdk.IL , Microsoft.NETCore.DotNetHost , Microsoft.NETCore.DotNetHostPolicy , System.Text.Json
From Version 5.0.0-alpha.1.19563.3 -> To Version 6.0.0-alpha.1.20601.3
* Replace obsoleted json serializer option
'JsonSerializerOptions.IgnoreNullValues' is obsolete: 'To ignore null values when serializing, set DefaultIgnoreCondition to JsonIgnoreCondition.WhenWritingNull.'
* Update dependencies from https://github.com/dotnet/runtime build
20201202.2
runtime.native.System.IO.Ports , Microsoft.NET.Sdk.IL , Microsoft.NETCore.DotNetHost , Microsoft.NETCore.DotNetHostPolicy , Microsoft.NETCore.ILAsm , System.Text.Json
From Version 5.0.0-alpha.1.19563.3 -> To Version 6.0.0-alpha.1.20602.2
* Update dependencies from https://github.com/dotnet/runtime build
20201203.3
runtime.native.System.IO.Ports , Microsoft.NETCore.ILAsm , Microsoft.NET.Sdk.IL , Microsoft.NETCore.DotNetHost , Microsoft.NETCore.DotNetHostPolicy , System.Text.Json
From Version 5.0.0-alpha.1.19563.3 -> To Version 6.0.0-alpha.1.20603.3
* Use PDFMT=PORTABLE option for ilproj
* Update dependencies from https://github.com/dotnet/runtime build
20201203.13
runtime.native.System.IO.Ports , Microsoft.NETCore.ILAsm , Microsoft.NET.Sdk.IL , Microsoft.NETCore.DotNetHost , Microsoft.NETCore.DotNetHostPolicy , System.Text.Json
From Version 5.0.0-alpha.1.19563.3 -> To Version 6.0.0-alpha.1.20603.13
* Update dependencies from https://github.com/dotnet/runtime build
20201204.10
runtime.native.System.IO.Ports , Microsoft.NETCore.ILAsm , Microsoft.NET.Sdk.IL , Microsoft.NETCore.DotNetHost , Microsoft.NETCore.DotNetHostPolicy , System.Text.Json
From Version 5.0.0-alpha.1.19563.3 -> To Version 6.0.0-alpha.1.20604.10
* Update dependencies from https://github.com/dotnet/runtime build
20201205.2
runtime.native.System.IO.Ports , Microsoft.NETCore.ILAsm , Microsoft.NET.Sdk.IL , Microsoft.NETCore.DotNetHost , Microsoft.NETCore.DotNetHostPolicy , System.Text.Json
From Version 5.0.0-alpha.1.19563.3 -> To Version 6.0.0-alpha.1.20605.2
* Update dependencies from https://github.com/dotnet/runtime build
20201206.6
runtime.native.System.IO.Ports , Microsoft.NETCore.ILAsm , Microsoft.NET.Sdk.IL , Microsoft.NETCore.DotNetHost , Microsoft.NETCore.DotNetHostPolicy , System.Text.Json
From Version 5.0.0-alpha.1.19563.3 -> To Version 6.0.0-alpha.1.20606.6
* Update dependencies from https://github.com/dotnet/runtime build
20201208.2
runtime.native.System.IO.Ports , Microsoft.NETCore.ILAsm , Microsoft.NETCore.DotNetHostPolicy , Microsoft.NET.Sdk.IL , Microsoft.NETCore.DotNetHost , System.Text.Json
From Version 5.0.0-alpha.1.19563.3 -> To Version 6.0.0-alpha.1.20608.2
* Update dependencies from https://github.com/dotnet/runtime build
20201209.2
runtime.native.System.IO.Ports , Microsoft.NETCore.ILAsm , Microsoft.NETCore.DotNetHostPolicy , Microsoft.NET.Sdk.IL , Microsoft.NETCore.DotNetHost , System.Text.Json
From Version 5.0.0-alpha.1.19563.3 -> To Version 6.0.0-alpha.1.20609.2
* Update dependencies from https://github.com/dotnet/runtime build
20201209.15
runtime.native.System.IO.Ports , Microsoft.NETCore.ILAsm , Microsoft.NETCore.DotNetHostPolicy , Microsoft.NET.Sdk.IL , Microsoft.NETCore.DotNetHost , System.Text.Json
From Version 5.0.0-alpha.1.19563.3 -> To Version 6.0.0-alpha.1.20609.15
* Add workaround for #45908
* Update dependencies from https://github.com/dotnet/runtime build
20201211.2
runtime.native.System.IO.Ports , Microsoft.NETCore.ILAsm , Microsoft.NETCore.DotNetHostPolicy , Microsoft.NET.Sdk.IL , Microsoft.NETCore.DotNetHost , System.Text.Json
From Version 5.0.0-alpha.1.19563.3 -> To Version 6.0.0-alpha.1.20611.2
* Update dependencies from https://github.com/dotnet/runtime build
20201211.23
runtime.native.System.IO.Ports , Microsoft.NETCore.ILAsm , Microsoft.NETCore.DotNetHostPolicy , Microsoft.NET.Sdk.IL , Microsoft.NETCore.DotNetHost , System.Text.Json
From Version 5.0.0-alpha.1.19563.3 -> To Version 6.0.0-alpha.1.20611.23
* Update dependencies from https://github.com/dotnet/runtime build
20201212.4
runtime.native.System.IO.Ports , Microsoft.NETCore.ILAsm , Microsoft.NETCore.DotNetHostPolicy , Microsoft.NET.Sdk.IL , Microsoft.NETCore.DotNetHost , System.Text.Json
From Version 5.0.0-alpha.1.19563.3 -> To Version 6.0.0-alpha.1.20612.4
* Update versions
* Remove PDBFMT option
* Apply ilasm .locals workaround for more files
See https://github.com/dotnet/runtime/issues/45908
* Revert "Apply ilasm .locals workaround for more files"
This reverts commit
54eb3ee53970b430b87ea2044475a2bbc872f20f.
* Revert "Add workaround for #45908"
This reverts commit
290631058301fc791d261395f436b3b875a003ca.
* Update versions
* Disable tests that fail on Mono due to https://github.com/dotnet/runtime/issues/46124
Co-authored-by: dotnet-maestro[bot] <dotnet-maestro[bot]@users.noreply.github.com>
Co-authored-by: Alexander Köplinger <alex.koeplinger@outlook.com>
Co-authored-by: Jan Kotas <jkotas@microsoft.com>
Alex Perovich [Wed, 16 Dec 2020 11:10:10 +0000 (03:10 -0800)]
Fix cross compiling linux managed assemblies on windows (#45975)
* Fix cross compiling linux managed assemblies on windows
This is needed to fix the source index.
* More targeted fix
* Update Configurations.props
Co-authored-by: Viktor Hofer <viktor.hofer@microsoft.com>
Maxim Lipnin [Wed, 16 Dec 2020 10:03:58 +0000 (13:03 +0300)]
Run wasm console sample on Helix (#45768)
- Included the WASM console sample project to the library test build. The sample can still be built/run locally using make file.
- Added zipping/copying the sample application to helix directory from where it can be consumed by CI.
- To run on Helix, tried to use a simple helix work item which wouldn't rely on generated RunTests.sh and just call the xharness command directly.
Relates to https://github.com/dotnet/runtime/issues/43865
Jan Kotas [Wed, 16 Dec 2020 09:46:42 +0000 (01:46 -0800)]
Misc fixes from runtimelab:NativeAOT (#46112)
Aleksey Kliger (λgeek) [Wed, 16 Dec 2020 07:42:02 +0000 (02:42 -0500)]
[mono] Make HelloWorld and wasm samples compile again (#46113)
* [mono] Make HelloWorld sample compile again
Since
296aaf818ec02101226845aea39ca14d32519cf2 it seems we need
`$(NetCoreAppToolCurrent)` as the `TargetFramework`
* Also fixup wasm samples
Koundinya Veluri [Wed, 16 Dec 2020 06:50:12 +0000 (22:50 -0800)]
Make the native thread pool initialization variable volatile (#46115)
- Necessary on arm32/arm64 such that if `IsInitialized()` returns true, initialized variables will also be observed correctly
- Fixes https://github.com/dotnet/runtime/issues/46084
Bruce Forstall [Wed, 16 Dec 2020 06:42:27 +0000 (22:42 -0800)]
Fix format of JIT-EE version GUID (#46116)
Eric Erhardt [Wed, 16 Dec 2020 03:33:00 +0000 (21:33 -0600)]
Fix remaining ILLink warnings in System.Reflection (#45984)
* Add RequiresUnreferencedCode to Delegate constructors
Contributes to #45623
* Add RequiresUnreferencedCode to some internal Reflection methods to address ILLink warnings.
* Address remaining ILLink warnings in Reflection.Emit.
* Respond to PR feedback
Annotate Delegate.BindToMethod with DynamicallyAccessedMembers
Fix up requires unreferenced code comments
* Update API Compat txt for Delegate.CreateDelegate attribute removal
* Fix Mono ILLink warnings
Anirudh Agnihotry [Wed, 16 Dec 2020 03:31:50 +0000 (19:31 -0800)]
use pack on csproj to build the packages (#45564)
* use pack on csproj to build the packages
* having projectreferences available during project evaluation and then having pack task handling is project isPackable or not
* default target as pack
* using default sdk tags and adding missing ilproj
* exclude shproj
* Do not use a recursive regex
* add csproj for projects with no build output
* adding comments about purpose of wrapper projects
Santiago Fernandez Madero [Wed, 16 Dec 2020 02:03:50 +0000 (18:03 -0800)]
Add another way to re-trigger pipelines in PR.
Ankit Jain [Wed, 16 Dec 2020 00:51:10 +0000 (19:51 -0500)]
[wasm][tests] Exit the script if `WasmTestRunner` build fails (#46048)
* [wasm][tests] Exit the script if `WasmTestRunner` build fails
If building the `WasmTestRunner.proj` on helix fails, then the error
gets ignored, and script execution continues. And that results in the
failure to be shown as:
```
/home/helixbot/work/
9CE408BA/w/
BF5C0A02/e/JIT/Methodical/eh/interactions/throw1dimarray_d/throw1dimarray_d.sh: line 136: cd: WasmApp: No such file or directory
/home/helixbot/work/
9CE408BA/w/
BF5C0A02/e/JIT/Methodical/eh/interactions/throw1dimarray_d/throw1dimarray_d.sh: line 137: ./run-v8.sh: No such file or directory
```
.. which is misleading. Looking at the execution log shows that the real error was:
```
Microsoft (R) Build Engine version 16.8.0+
126527ff1 for .NET
Copyright (C) Microsoft Corporation. All rights reserved.
AppDir: /home/helixbot/work/
9CE408BA/w/
BF5C0A02/e/JIT/Methodical/eh/interactions/throw1dimarray_d/WasmApp/
TestBinDir: /home/helixbot/work/
9CE408BA/w/
BF5C0A02/e/JIT/Methodical/eh/interactions/throw1dimarray_d
ArtifactsBinDir:
/home/helixbot/work/
9CE408BA/p/build/WasmApp.targets(11,5): error : Failed to load assembly reference 'System.Windows.Forms, Version=0.0.0.0, Culture=neutral, PublicKeyToken=null' for 'Except, Version=0.0.0.0, Culture=neutral, PublicKeyToken=null': Could not find assembly 'System.Windows.Forms, Version=0.0.0.0, Culture=neutral, PublicKeyToken=null'. Either explicitly load this assembly using a method such as LoadFromAssemblyPath() or use a MetadataAssemblyResolver that returns a valid assembly. [/home/helixbot/work/
9CE408BA/p/wasm-test-runner/WasmTestRunner.proj]
```
So, exit the script if the build fails.
* [wasm] Support RunOnly tests with wasm
Some test projects have `<CLRTestKind>RunOnly</CLRTestKind>`, eg.
`GitHub_26491_SingleReturnSynchronized`. In this particular project, it
doesn't have it own code/assembly, and just runs another assembly.
But the default script generated for the test adds the build command for
`WasmTestRunner.proj`, which fails here because there is no assembly to
bundle!
This manifested for
`JIT/Regression/JitBlue/GitHub_26491/GitHub_26491_SingleReturnSynchronized/GitHub_26491_SingleReturnSynchronized.sh` as:
```
BEGIN EXECUTION
Microsoft (R) Build Engine version 16.8.0+
126527ff1 for .NET
Copyright (C) Microsoft Corporation. All rights reserved.
AppDir: /home/helixbot/work/
AA96097C/w/
9CC508B9/e/JIT/Regression/JitBlue/GitHub_26491/GitHub_26491_SingleReturnSynchronized/WasmApp/
TestBinDir: /home/helixbot/work/
AA96097C/w/
9CC508B9/e/JIT/Regression/JitBlue/GitHub_26491/GitHub_26491_SingleReturnSynchronized
ArtifactsBinDir:
/home/helixbot/work/
AA96097C/p/build/WasmApp.targets(11,5): error : Could not find assembly '/home/helixbot/work/
AA96097C/w/
9CC508B9/e/JIT/Regression/JitBlue/GitHub_26491/GitHub_26491_SingleReturnSynchronized/GitHub_26491_SingleReturnSynchronized.dll' [/home/helixbot/work/
AA96097C/p/wasm-test-runner/WasmTestRunner.proj]
Test Harness Exitcode is : 1
```
This *test* wasn't failing earlier, because the build command would
fail, but the script continued execution. And the subsequent shell
script worked as expected. But now we exit if the build command fails,
causing the whole test to fail!
Solution: Honor `$(CLRTestKind) == RunOnly`, and don't emit the build
command in that case.
* Just to give this more context:
this was for `src/tests/JIT/Regression/JitBlue/GitHub_2649`
it has two projects:
A: `GitHub_26491_MultipleReturns.ilproj`
B: `GitHub_26491_SingleReturnSynchronized.ilproj`
Project A produces the test assembly. And we generate a wasm app for that.
Project B is just meant to run Project A's assembly. It does not need it's own wasm app. That's why it has
`<CLRTestKind>RunOnly</CLRTestKind>`
So, as the fix, we add the build WasmTestRunner.proj command to the generated shell script only if `CLRTestKind==BuildAndRun`.
Jeremy Koritzinsky [Wed, 16 Dec 2020 00:38:51 +0000 (16:38 -0800)]
Remove extraneous UsingTask (#46097)
We no longer use the CopyNupkgAndChangeVersion task, so we can remove it.
Jan Kotas [Wed, 16 Dec 2020 00:07:34 +0000 (16:07 -0800)]
Fix MethodDesc validation for FCalls (#45998)
Koundinya Veluri [Tue, 15 Dec 2020 23:06:20 +0000 (15:06 -0800)]
Follow-up fix for some PAL tests that compare sleep/wait times from one thread to another (#46060)
Follow-up fix for some PAL tests that compare sleep/wait times from one thread to another
Follow-up fix to https://github.com/dotnet/runtime/pull/45948, need to reset the global variable before creating the thread, as the test runs more than once
Jeremy Koritzinsky [Tue, 15 Dec 2020 21:48:24 +0000 (13:48 -0800)]
Fix building the crossdacs with ninja. (#46089)
Sung Yoon Whang [Tue, 15 Dec 2020 21:47:27 +0000 (13:47 -0800)]
Fix assert in SampleProfiler enabling on startup (#46045)
Jan Kotas [Tue, 15 Dec 2020 21:28:05 +0000 (13:28 -0800)]
Fix ilasm crash on empty local variable name (#46080)
Fixes #45908
Tom Deseyn [Tue, 15 Dec 2020 21:04:37 +0000 (22:04 +0100)]
SocketAsyncContext.Unix: fix double processing of AsyncOperations (#45683)
* SocketAsyncContext.Unix: fix double processing of AsyncOperations
When a socket event occurs, the pending operations gets triggered
to continue their work by calling the Process method.
The changes in https://github.com/dotnet/runtime/pull/37974
cause Process to be called twice on the same AsyncOperation.
When Process is called, the operation can complete, and the
AsyncOperation instance may be reused for a different operation.
* Remove processAsyncEvents
Ove Bastiansen [Tue, 15 Dec 2020 18:57:52 +0000 (19:57 +0100)]
adding null check on provider registration (#46002)
fix #45599
Kunal Pathak [Tue, 15 Dec 2020 18:47:35 +0000 (10:47 -0800)]
Enable crossgen SPMI collection for libraries (#46058)
* Better printing of files getting copied
* Exclude Core_Root folder for tests collection
* Perform libraries collection using crossgen
Eric StJohn [Tue, 15 Dec 2020 17:33:01 +0000 (09:33 -0800)]
Make EventLog work without .NETFramework (#45884)
* Make EventLog work without .NETFramework
* Use a project file instead of targets to create message DLL
* Add EventLogMessagesTests
* Address feedback
* Exclude EventLogMessage tests on net48
* Apply code review feedback
Also fix one test which would fail on machine without .NETFramework.
* Fix HelpLink test
Johan Lorensson [Tue, 15 Dec 2020 07:49:56 +0000 (08:49 +0100)]
Implementation of full EventPipeDiagnosticServer C library CoreClr shim. (#45354)
Break out build of eventpipe library into vm/eventing/eventpipe
(both for C++ as well as C library).
Introduce ability to build CoreClr with C++ or C version of
EventPipe/DiagnosticServer library. Beneficial during transition
to have both library useable, C++ library is still default,
but new cmake feature FEATURE_PERFTRACING_C_LIB can be defined
in order to build and use C library instead of C++ version.
Due to restrictions around standard headers in CoreClr,
moved out C library IPC PAL into debug_pal
(where C++ library has its IPC PAL). This unfortunately came
with some additional needed changes since library had dependencies
against runtime, so needed to make PAL sources independent of
majority of EventPipe library. Added same feature as for
EventPipe/DiagnosticServer library, possible to build debug_pal
with C or C++ IPC PAL sources.
Complemented a couple of CoreClr types with NoThrow methods
so they could be used in shim containers without need for
try/catch, CQuickArray::PushNoThrow, SHash::AddNoThrow.
Missing strtoull on non Windows, add PAL_strtoull
into PAL (PAL already included UTF16 version of the same functionality).
Jeremy Koritzinsky [Tue, 15 Dec 2020 06:26:30 +0000 (22:26 -0800)]
Enable non-blittable struct returns on UnmanagedCallersOnly (#45625)
* Allow non-primitive struct returns to not require a stub. Fixes #35928.
* Support propagating the UnmanagedCallersOnly calling convention to the JIT. Block UnmanagedCallersOnly in crossgen1 since the attribute parsing code isn't included.
* Support passing through the calling convention for UnmanagedCallersOnly in crossgen2
* Fix clang errors.
* Fix stack calculation.
* Fix usings
* Clean up jitinterface.
* Remove invalid assert.
* Fix up stdcall name mangling lookup.
* Fix flag condition.
* Use the register var type when copying from the register to the stack.
* Change flag check for readability.
* Rename variables to remove shadowing.
* Fix formatting.
* Create new getEntryPointCallConv method on the EE-JIT interface to handle all calling convention resolution and support extensible calling conventions.
* Remove unreachable code.
* Remove now unused getUnmanagedCallConv jitinterface method (replaced by getEntryPointCallConv).
* Fix formatting.
* Rename getEntryPointCallConv and only call it with a method when it's a P/Invoke or Reverse P/Invoke.
* Pass SuppressGCTransition through the getUnmanagedCallConv JIT-EE interface API.
* Refactor callconv handling so we can handle reverse P/Invokes with the callconv in the signature (and not in an UnmanagedCallersOnly attribute).
* Clean up whitespace.
* Pass MethodIL as the scope for the signature to enable propagating down annotations for calli's in P/Invokes.
* Remove usages of CORINFO_CALLCONV_FOO where FOO is an unmanaged callconv. move the definitions of those ones to the interpreter since that's the only place they're used.
* SuppressGC cleanup.
* Rename superpmi struct
* Add default condition to make clang happy.
* change enums to make clang happy.
* Remove CORINFO_CALLCONV_C and family from interpreter.
* Fix up handling of managed function pointers and remove invalid assert.
* Continue to use sigflag for suppressgc workaround.
* Clean up comment wording.
Signed-off-by: Jeremy Koritzinsky <jekoritz@microsoft.com>
* Remove more MethodIL passing we don't need any more
* Update src/coreclr/tools/Common/JitInterface/CorInfoImpl.cs
Co-authored-by: Jan Kotas <jkotas@microsoft.com>
* Fix SigTypeContext creation.
* Pass context by ptr.
* Fix formatting.
* Clear the Reverse P/Invoke flag when compiling inlinees. It's only needed on the entry-point functions.
Co-authored-by: Jan Kotas <jkotas@microsoft.com>
Bruce Forstall [Tue, 15 Dec 2020 03:37:27 +0000 (19:37 -0800)]
Fix GC vars dump at call sites for arm/arm64 (#46056)
Dan Moseley [Tue, 15 Dec 2020 03:20:09 +0000 (19:20 -0800)]
Enable File.Copy to EXFAT volumes when not running as root (#46027)
* Allow File.Copy to succeed when futimens fails with EPERM
* Manual test
* Update pal_io.c
Werner Mairl [Tue, 15 Dec 2020 02:16:09 +0000 (03:16 +0100)]
PR: Add test code coverage for Host.StopAsync (#40875)
* Add test code coverage for Host.StopAsync
Verifies StopAsync() doesn't throw CanceledException given a canceled token
Fix #40700
* implement using/dispose for hosts inside tests
Levi Broderick [Tue, 15 Dec 2020 01:01:29 +0000 (17:01 -0800)]
Fix bad case-insensitive ASCII equality check (#45928)
Tomáš Rylek [Mon, 14 Dec 2020 23:56:53 +0000 (00:56 +0100)]
Fix previously overlooked path to pgocheck.py (src/coreclr/src) (#46019)
Looking around the script I also noticed that __SourceDir
is not used anymore so I removed it.
Thanks
Tomas
Anirudh Agnihotry [Mon, 14 Dec 2020 22:40:50 +0000 (14:40 -0800)]
Updating darc dependencies and enabling closure (#45781)
* automatic version updates
* coreclr projects retargets to net5.0
* Removing manual addition of knownFrameworkReference
* disavling implicit references for sfx projects
* mono changes
* enable package testing
* installer changes
* installer tests ref pack changes
* remove the known framework reference for unit tests
Co-authored-by: Jeremy Koritzinsky <jkoritzinsky@gmail.com>
* TestInfraTargetFramework -> to net5.0
* set netcoremaxversion property for test utils
* retargeting mono sample apps to net5.0
Co-authored-by: Jeremy Koritzinsky <jkoritzinsky@gmail.com>
Co-authored-by: Alexander Köplinger <alex.koeplinger@outlook.com>
Buyaa [Mon, 14 Dec 2020 22:31:44 +0000 (14:31 -0800)]
Enable Supported attributes from sdk for non browser build, fix related warnings (#45504)
Ankit Jain [Mon, 14 Dec 2020 22:13:38 +0000 (17:13 -0500)]
[wasm] Move Wasm app building targets to new WasmApp.targets (#45977)
* [wasm] Don't quote multiple msbuild args
Else it ends up in the command line as:
`dotnet build foo.proj "/p:TargetOS=Browser /p:Configuration=Release"`
instead of
`dotnet build foo.proj /p:TargetOS=Browser /p:Configuration=Release`
* [wasm] Fix the name for MicrosoftNetCoreAppRuntimePackRidDir
We have `MicrosoftNetCoreAppRuntimePackDir=artifacts/bin/microsoft.netcore.app.runtime.browser-wasm/Release/`
.. and `MicrosoftNetCoreAppRuntimePack*Rid*Dir` as `$(MicrosoftNetCoreAppRuntimePackDir)/runtimes/browser-wasm/`.
But the wasm samples, and debugger test project files mixed up the
naming, though with correct paths. The following commit will move the
wasm build stuff to a single file, and this change will be useful to
have uniform, and correct naming.
* [wasm] Add new `WasmApp.targets`
Move the uses of `WasmAppBuilder`, and
`WasmLoadAssembliesAndReferences` tasks to a single targets file -
`src/mono/wasm/build/WasmApp.targets`.
- Any project that wants to use that, can import the file, and set up the
various properties before the target `WasmBuildApp` gets executed.
- `WasmBuildApp` is not run by default. The importing project will have
to do that.
- By default, the `WasmLoadAssembliesAndReferences` task is not run, and
the specified `@(WasmAssembliesToBundle)` are directly passed to
`WasmAppBuilder`.
- If the project needs assembly dependencies to be resolved, then
set `$(WasmResolveAssembliesBeforeBuild) == true`.
- Assemblies to be bundled with the app are set via
`@(WasmAssembliesToBundle)` (which optionally will have dependencies
resolved)
The various task inputs correspond to properties as:
```
AssemblySearchPaths : @(WasmAssemblySearchPaths)
Assemblies : @(WasmAssembliesToBundle)
AppDir : $(WasmAppDir)
MicrosoftNetCoreAppRuntimePackDir : $(MicrosoftNetCoreAppRuntimePackRidDir)
MainAssembly : $(WasmMainAssemblyPath)
MainJS : $(WasmMainJSPath)
InvariantGlobalization : $(WasmInvariantGlobalization)
SatelliteAssemblies : @(WasmSatelliteAssemblies)
FilesToIncludeInFileSystem : @(WasmFilesToIncludeInFileSystem)
DebugLevel : $(WasmDebugLevel)
```
This should be a step towards eventually having this build as a sdk.
* [wasm] Use the correct assembly path property
* Check for TargetOS==Browser
* Pass remaining WasmAppBuilder args - IcuDataFileName, and RemoteSources
* WasmAppBuilder: don't emit run-v8.sh in the task
* WasmTestRunner: use the new WasmApp.targets
* Include WasmMainAssemblyPath for dependency resolution, by default
* Remove unused run-v8.sh generation
* Add back run-v8.sh for console sample, WasmTestRunner, and lib tests
* WasmApp.targets: Use a default path for run-v8.sh
.. and control it's generation through `$(WasmGenerateRunV8Script)`
* run-v8.sh: use only the filename on the command line
* update README.md
* [wasm] Make the wasm targets conditional on `TargetOS==Browser`
.. instead of conditionally importing `WasmApp.targets`. This allows
other targets to depend on wasm targets, and execute only if it makes
sense.
Based on @lewing's feedback.
* [wasm][tests] fix check for when to generate run-v8.sh
* WasmTestRunner.proj: fix typo
* [wasm] Always generate the run-v8.sh script, for browser runs
The `Scenario==WasmTestOnBrowser` seems to fail for JIT tests, which
don't use the browser.
* [wasm] Fix tests
The real issue seems to be that the helix bundle doesn't have
`WasmApp.targets`. Fixing that, which should hopefully fix the tests!
* [wasm] Another attempt at fixing the tests
Mimic earlier behavior of having a `BundleTestWasmApp`, and making it
conditional on `TargetOS == Browser`.
* fix property reference
* [wasm] skip missing assemblies for the jit tests, same as the original code
Jeremy Koritzinsky [Mon, 14 Dec 2020 21:21:32 +0000 (13:21 -0800)]
Use netfxreference.props to provide the shims for the platform manifest. (#45573)
* Use netfxreference.props to provide the shims for the platform manifest.
* PR feedback.
David Wrighton [Mon, 14 Dec 2020 20:47:32 +0000 (12:47 -0800)]
Auto-generate all C++ header overrides of the entire jit interface (#45920)
* Auto-generate all C++ header overrides of the entire jit interface
- This reduces the amount of manual change required when updating the jit interface to just implementation changes
- Will cause compile errors if any api is removed from the interface, but not removed from various implementations
* Consistently specify override on the jit ee interface header
- Satisfies warnings on clang
Fan Yang [Mon, 14 Dec 2020 20:39:06 +0000 (15:39 -0500)]
Update testing.md (#46043)
skipstressdependencies switch was removed from the runtime tests build script.
LateApexEarlySpeed [Mon, 14 Dec 2020 16:15:59 +0000 (00:15 +0800)]
Create more informational assert collection log for FunctionalTests o… (#44350)
Egor Bogatov [Mon, 14 Dec 2020 12:31:52 +0000 (15:31 +0300)]
[mono] Fix race condition in Utils + remove duplicated code (#46029)
dotnet-maestro[bot] [Mon, 14 Dec 2020 11:29:08 +0000 (12:29 +0100)]
[master] Update dependencies from mono/linker (#45926)
Co-authored-by: dotnet-maestro[bot] <dotnet-maestro[bot]@users.noreply.github.com>
Miha Zupan [Mon, 14 Dec 2020 11:17:24 +0000 (12:17 +0100)]
Disable tests for 43931 and 46030 (#46031)
* Disable tests for 43931 and 46030
* Add [ActiveIssue] in comment
Miha Zupan [Mon, 14 Dec 2020 11:01:12 +0000 (12:01 +0100)]
Optimize percent-encoded UTF8 processing in Uri (#32552)
* Optimize percent-encoded UTF8 processing in Uri
* Rename charsConsumed to bytesConsumed
* Use ValueStringBuilder Append(char*, int) instead of Append(ROS<char>)
* Add tests for PercentEncodingHelper
* Use string literals instead of char.ConvertFromUtf32
* Use sizeof(uint) instead of 4
* Add missing license headers
* Improve codegen by using temporary local copy
* Correct Debug asserts
* Add ValueStringBuilder.Append(Rune)
* Improve hex decoding throughput
* Move VSB.Append(Rune) to a Uri-specific partial VSB file
* Add missing csproj link
* Add more comments documenting PercentEncodingHelper's logic
* Fix rebase conflicts
* Address PR feedback
John Call [Mon, 14 Dec 2020 10:56:01 +0000 (02:56 -0800)]
SortedDictionary Copy optimization (#45659)
* Sorted Dictionary Copy optimization
* Update src/libraries/System.Collections/src/System/Collections/Generic/SortedDictionary.cs
Co-authored-by: Jan Kotas <jkotas@microsoft.com>
* Feedback
Co-authored-by: Jan Kotas <jkotas@microsoft.com>
xtqqczze [Mon, 14 Dec 2020 01:46:25 +0000 (01:46 +0000)]
Remove nonsensical comments about fixed statement (#46013)
dotnet-maestro[bot] [Sun, 13 Dec 2020 23:59:00 +0000 (15:59 -0800)]
Update dependencies from https://github.com/dotnet/arcade build
20201210.4 (#46004)
Microsoft.DotNet.Build.Tasks.Archives , Microsoft.DotNet.Build.Tasks.Feed , Microsoft.DotNet.Build.Tasks.Installers , Microsoft.DotNet.Build.Tasks.Packaging , Microsoft.DotNet.Build.Tasks.SharedFramework.Sdk , Microsoft.DotNet.Build.Tasks.TargetFramework.Sdk , Microsoft.DotNet.CodeAnalysis , Microsoft.DotNet.GenAPI , Microsoft.DotNet.GenFacades , Microsoft.DotNet.XUnitExtensions , Microsoft.DotNet.Arcade.Sdk , Microsoft.DotNet.Helix.Sdk , Microsoft.DotNet.RemoteExecutor , Microsoft.DotNet.SharedFramework.Sdk , Microsoft.DotNet.VersionTools.Tasks , Microsoft.DotNet.XUnitConsoleRunner , Microsoft.DotNet.ApiCompat
From Version 6.0.0-beta.20609.9 -> To Version 6.0.0-beta.20610.4
Co-authored-by: dotnet-maestro[bot] <dotnet-maestro[bot]@users.noreply.github.com>
Jan Kotas [Sun, 13 Dec 2020 20:05:48 +0000 (12:05 -0800)]
Fix FC_NO_TAILCALL with newer compilers (#45999)
The C11 standard treats trivial infinite loops as undefined behavior and allows
compilers to optimize them out completely.
Add "volatile" on FC_NO_TAILCALL to defeat this optimization
monojenkins [Sun, 13 Dec 2020 07:47:18 +0000 (02:47 -0500)]
[aot] Quote the -Wl,install_name argument to clang. (#46000)
Fixes https://github.com/mono/mono/issues/20606.
Co-authored-by: vargaz <vargaz@users.noreply.github.com>
Steve MacLean [Sat, 12 Dec 2020 22:23:29 +0000 (17:23 -0500)]
Fix macos DAC & DBI search dependency paths (#45978)
* Fix macos DAC & DBI search dependency paths
* Only apply changes to macos-arm64
Bruce Forstall [Sat, 12 Dec 2020 20:52:27 +0000 (12:52 -0800)]
superpmi.py: Add support for easily collecting crossgen of assemblies (#45996)
Add a new `--crossgen` option to `collect` that will do a SuperPMI
collection by invoking crossgen on the set of assemblies specified
by the `-assemblies` argument.
Renamed the `-pmi_assemblies` argument to `-assemblies` so it can be
used by both `--pmi` and the new `--crossgen` option.
You can use both `--pmi` and `--crossgen` to collect both a PMI
and crossgen run across the same set of assemblies.
Bruce Forstall [Sat, 12 Dec 2020 20:50:14 +0000 (12:50 -0800)]
Remove superpmi.py asmdiff view with VS Code options (#45993)
The superpmi.py asmdiffs code is complicated by the existence of
various options to invoke VS Code to view diffs. Remove these options
(which I believe are unused, and shouldn't be in this location anyway)
and simplify the code.
Bruce Forstall [Sat, 12 Dec 2020 20:46:56 +0000 (12:46 -0800)]
Add a GC info option to superpmi.py asm diffs (#45992)
Add `--gcinfo` to an `asmdiffs` command to get asm diffs with GC
info dumped via the `COMPlus_JitGCDump` / `COMPlus_NgenGCDump` variables.
Improved the output of dasm/dump locations if there are no seen diffs.
xtqqczze [Sat, 12 Dec 2020 15:11:58 +0000 (15:11 +0000)]
Use C# 7.3 pattern-based fixed statement (#45969)
Documentation for `ValueStringBuilder.GetPinnableReference()` suggests using pattern-matching instead of the explicit method call.
Ankit Jain [Sat, 12 Dec 2020 04:07:02 +0000 (23:07 -0500)]
[wasm] provision-wasm: always install the "official" version (#45990)
if `EMSDK_PATH` was unset, or set to the default path, then
`provision-wasm` target would become a no-op. Instead, always install it
to the default path.
It can be useful to have the official version available, even if you
have a different one installed, so you can do comparisons etc. - Zoltan
Andy Ayers [Sat, 12 Dec 2020 01:40:11 +0000 (17:40 -0800)]
JIT: consistently mark method table accesses as invariant (#45854)
Introduce a new utility method to create the IR for method table access,
marking the resulting indir as invariant. This allows method table access
CSEs (which will be increasingly common with the advent of PGO-enabled
Guarded Devirtualization).
Add a workaround to assertion prop for method table to be able to see
through a CSE'd method table access.
Remove a few optimizer inhibitions related to colon cond.
Jan Kotas [Sat, 12 Dec 2020 01:22:23 +0000 (17:22 -0800)]
Delete non-supported handling of Windows-specific PDBs in ilasm (#45979)
I do not expect we will ever add the classic PDB support back in ilasm.
Fixes #45492
Dan Moseley [Sat, 12 Dec 2020 00:43:48 +0000 (16:43 -0800)]
Update issues-pr-management.md (#45693)
Change to present tense, as we're now merged.
Document workflow enabled in https://github.com/dotnet/runtime/issues/44775
Bruce Forstall [Fri, 11 Dec 2020 23:47:20 +0000 (15:47 -0800)]
Fix superpmi.py argument parsing (#45983)
1. Hoist out argument parsing/verification of target_os/target_arch/mch_arch
so they apply to more commands that require them, e.g., "list-collections".
2. Rename `-jit` argument to `-jit_name` to be more explicit, and differentiate
it from the pre-existing `-jit_path` argument.
3. Hoist verification of `-jit_ee_version` to reduce code duplication.
Fixes #45981
Sergey Andreenko [Fri, 11 Dec 2020 22:54:03 +0000 (14:54 -0800)]
Mark some pipelines as pr:none (#45973)
* Follow-up for #45931
* Add superpmi
Igor Velikorossov [Fri, 11 Dec 2020 21:48:09 +0000 (08:48 +1100)]
FontConverter restore order of font properties (#45746)
.NET Framework implementation of `FontConverter` provided a different
order of font properties, that can be observed by inspecting an instance
of a font in a `PropertyGrid`.
Restore the original sort order.
The .NET Framework implementation has another sort argument "Weight",
which however doesn't appear to be a property of `Font` type.
It is possible this property has existed at some point, or the converter
may have been expected to work for instances of `IFontDisp` object.
Either way presence or absence of "Weight" doesn't appear to make any
difference, hence it has not been ported across.
Fixes #45631
Sergey Andreenko [Fri, 11 Dec 2020 21:27:09 +0000 (13:27 -0800)]
Improve PUTARG_STK generation for Arm/Arm64. (#45824)
* Improve PUTARG_STK generation.
Allow non pointer-size put args for structs.
It was an old CQ issue but now it is an Arm64 apple requirement.
* Update src/coreclr/jit/codegenarmarch.cpp
* Update src/coreclr/jit/codegenarmarch.cpp
Co-authored-by: Andy Ayers <andya@microsoft.com>
Co-authored-by: Andy Ayers <andya@microsoft.com>
Jan Kotas [Fri, 11 Dec 2020 21:07:43 +0000 (13:07 -0800)]
Fix reporting of negative ISA uses in crossgen2 (#45892)
Fixes #45852
Brian Sullivan [Fri, 11 Dec 2020 20:58:06 +0000 (12:58 -0800)]
Fix for Issue 44895 (#45284)
* Fix for Issue 44895
- Fix: Don't allow an unwrapped promoted field of TYP_REF to be returned when we are expecting a TYP_STRUCT
Backout change in gtGetStructHandleIfPresent for GT_RETURN as it isn't needed for this fix
Deoptimize all GT_RETURN's with mismatched types for promoted struct fields.
* Allow both GT_ADDR and GT_ASG as a parent node
* Add second test case Repro2_44895.cs
* Change assert about Incompatible types to be a noway_assert in gtNewTempAssign
* Only use the smaller repro case for Runtime_44895.cs
* Added noway_assert in release build for an assignment of a TYP_REF to a TYP_STRUCT
* rerun jit-format
dotnet-maestro[bot] [Fri, 11 Dec 2020 19:12:54 +0000 (19:12 +0000)]
[master] Update dependencies from dotnet/xharness dotnet/arcade (#45956)
[master] Update dependencies from dotnet/xharness dotnet/arcade
Maxim Lipnin [Fri, 11 Dec 2020 17:43:05 +0000 (20:43 +0300)]
Run Mono Android sample on Helix (#45109)
- Included the android sample project to the library test build.
- Added copying resulting HelloAndroid.apk to helix test directory from where it can be consumed by CI.
- Set up the return value (42) as an expected exit code for xharness.
- Xharness for android doesn't support running an application w/o test runner yet
Related issue: #43865
monojenkins [Fri, 11 Dec 2020 15:12:35 +0000 (10:12 -0500)]
[metadata] Generate type-3 GUIDs for interfaces. (#43677)
I'm opening this although it's still WIP, for discussion, as I'm not sure what else I should be adding. Tests probably, of course, but also I was wondering if it should be plugged to RuntimeType.GUID / Marshal.GenerateGuidForType. I'm not entirely sure if that's desired as I've seen clues in the reference source that the GUID generation has issues.
I've left class GUID generation aside, as it doesn't seem so useful for COM interop, and the algorithm uses some Assembly properties to seed the GUID, and retrieving that is currently beyond my understanding of the code.
Also I'm a bit lost regarding which RuntimeType.GUID it should be added to: there's at least two places that looked appropriate : in `mono/netcore/System.Private.CoreLib/src/RuntimeType.Mono.cs` and in `mono/mcs/class/corlib/ReferenceSources/RuntimeType.cs`.
Co-authored-by: rbernon <rbernon@users.noreply.github.com>
Fan Yang [Fri, 11 Dec 2020 14:58:51 +0000 (09:58 -0500)]
Add test leg to run runtime tests on Android devices (#44560)
* Add test leg to run runtime tests on Android devices
* Add a new variable to describe the host OS of test wrappers for runtime tests
* Set win-x64 runtime for Android arm64
* Add missing %
* Use TestWrapperTargetsWindows when checking script targeting platform
* Fix variable usage format on windows
* Fix file path for windows
* Make file path portable
* Add debug info
* Add debug info to output instead of testOutput
* Add more debug info
* Add debug info in batch file
* Turn on echo for batch script
* Add another missing % and some debug info
* Fix test running command
* Remove extra double quotes
* Turn on verbose mode for xharness command and fix msbuild special characters
* Choose batch or bash script based on var TestWrapperTargetsWindows, when building Android apps
* Clean up
* Enable all tests
* Disable tests belonging to big work items
* Disable more long running work items
* Disable more long running tests
* Disable more long running tests and shorten the timeout threshold
* Increase timeout threshold
* For debugging, print out logs when passing
* Disable more tests to avoid long running tests
* Add Android tests to staging pipeline and revert previous debugging hack
Koundinya Veluri [Fri, 11 Dec 2020 13:26:10 +0000 (05:26 -0800)]
Fix some PAL tests that compare sleep/wait times from one thread to sleep/wait times on another thread (#45948)
- This affects some tests where the main thread T0 creates another thread T1, T0 sleeps for some time while T1 waits, then after waking up T0 interrupts T1's wait with an APC, and measures the expected sleep duration with the other thread's wait duration before interrupt
- In the failures T0 is likely starting to sleep for the interrupt delay before T1 has recorded the pre-wait timestamp, so the interrupt occurs sooner than intended and the expectation later fails
- Fix T0 to wait for T1 to record the pre-wait timestamp before sleeping for the interrupt delay
Hopefully fixes https://github.com/dotnet/runtime/issues/42669
Steve MacLean [Fri, 11 Dec 2020 13:01:58 +0000 (08:01 -0500)]
Fixes for VS preview (#45902)
* Fixes for VS preview
Stop using c++latest and designated initializers for Windows crossdac
Disable warning about overriding /TP with /TC
Maxim Lipnin [Fri, 11 Dec 2020 12:00:05 +0000 (15:00 +0300)]
Run Mono iOS sample on Helix (#44462)
- Included the ios sample project to the library test build. The sample still can be build/run locally using make file.
- Added copying resulting HelloiOS.app to helix directory from where it can be consumed by CI.
- Changed the sample to return 42 when building for CI, which is used as an expected exit code for xharness.
- Xharness for iOS supports running an application on a simulator w/o test runner inside, the sample is run by this approach.
- Checked locally that xharness can run HelloiOS.app successfully.
Related issue: #43865
Jan Kotas [Fri, 11 Dec 2020 11:55:08 +0000 (03:55 -0800)]
Add botr doc on managed type system (#45891)
monojenkins [Fri, 11 Dec 2020 08:06:38 +0000 (03:06 -0500)]
Add support for global aggressive-linining optimization. (#45822)
Currently we do support aggressive inlining per method using attributes. This commit adds a new global mono optimization flag that does similar thing, but whiout the need to annotate methods. There is still one big difference and that is the effect of cost. The new global optimization does not override cost evaluation, so if a method is to expensive to inline with regards to inline limit, it won't be inlined. In order to get the full "forced" inlining. method attributes should still be used.
This new optimization is not default and needs to be enabled using:
--optimize=aggressive-inlining
Co-authored-by: lateralusX <lateralusX@users.noreply.github.com>
Bruce Forstall [Fri, 11 Dec 2020 07:43:55 +0000 (23:43 -0800)]
Fix GC hole with STOREIND of LCL_VAR_ADDR/LCL_FLD_ADDR (#45818)
* Fix GC hole with STOREIND of LCL_VAR_ADDR/LCL_FLD_ADDR
We were updating the GC liveness of the ADDR node, but
`genUpdateLife()` expects to get the parent node, so no
liveness was ever updated.
There were 4 SPMI GC info diffs in the libraries, all
related to uses of `System.Collections.Immutable.ImmutableArray`
where we struct promote fields who are themselves single-element
gc ref structs that are kept on the stack and not in registers.
In all cases, the liveness of the stack local was not reflected
in codegen's GC sets, but it was reflected in the emitter's GC
sets, so it was marked as a GC lifetime. However, that lifetime
would get cut short if we hit a call site before the last use,
as calls (sometimes) carry the full set of live variables across
the call. So, variables not in this set (including the
"accidental" emitter-created GC lifetimes here) would get killed,
leaving a hole between the intermediate call and actual stack
local last use.
Fixes #45557
* Add unit test
* Add comment to `emitInsLoadInd()`
Andy Ayers [Fri, 11 Dec 2020 06:57:18 +0000 (22:57 -0800)]
JIT: implement return merging stress mode (#45917)
Under stress, optionally limit methods to having just one return block,
to stress the logic involved in merging returns.
Bruce Forstall [Fri, 11 Dec 2020 06:51:37 +0000 (22:51 -0800)]
Fix JitDisasmWithGC without JitDump (#45938)
Bruce Forstall [Fri, 11 Dec 2020 06:51:21 +0000 (22:51 -0800)]
Change GC stack slot negative hex offsets to show as negative (#45940)
In the GC output, instead of:
```
Stack slot id for offset -104 (0xffffff98) (frame) = 5.
```
display:
```
Stack slot id for offset -104 (-0x68) (frame) = 5.
```
This makes it easier to search for or visually match the
`-0x68` against asm like:
```
mov gword ptr [rbp-68H], rax
```
Kenneth Pouncey [Fri, 11 Dec 2020 04:37:16 +0000 (05:37 +0100)]
[browser][tests][outerloop] Standup OuterLoop WebSocket tests (#45470)
* [browser][tests][outerloop] Standup OuterLoop tests
* Reference issue not PR
* Add active issue for CloseDescription failing tests
* Add active issue for failures for invalid close codes failing tests
* Add active issue for failures when specifying subprotocol by URI Query
* Add active issue for Uncaught RuntimeError: memory access out of bounds
* Remove ActiveIssue and initialize with specific platform detection code
- Removes the necessity of setting up a new test.
- May want to go back to ActiveIssue for tracking purposes.
* Add ActiveIssue for CloseOutputAsync
* Add ActiveIssue for AbortTests
* Remove ActiveIssue and initialize with specific platform detection code
- Removes the necessity of setting up a new test.
- May want to go back to ActiveIssue for tracking purposes.
* Remove Active issue as it was fixed.
* Add ActiveIssue for CancelTests that is part of the Abort tests changes
* Update src/libraries/System.Net.WebSockets.Client/tests/ClientWebSocketOptionsTests.cs
Point to WebProxy support issue
Co-authored-by: Larry Ewing <lewing@microsoft.com>
Eric Erhardt [Fri, 11 Dec 2020 04:01:59 +0000 (22:01 -0600)]
Fix a few spelling mistakes (#45930)
paramterless => parameterless
Polymorhic => Polymorphic
Sergey Andreenko [Fri, 11 Dec 2020 03:07:49 +0000 (19:07 -0800)]
Make all pipelines triggerable via '/azp run name'. (#45931)
Brian Sullivan [Fri, 11 Dec 2020 02:37:54 +0000 (18:37 -0800)]
Fix for Issue 44762 (#45134)
* Fix for Issue 44762
Treat both SIMD and HWIntrinsic Memory Loads as Byref Exposed Loads that depend upon the global byref memory state.
In liveness added byref use for SIMD Memory Loads
Corrected the comment for GenTreeHWIntrinsic::OperIsMemoryLoadOrStore()
Added test case Runtime_44762.cs
* Moved the check for HWINTRINSIC node that have more than two args and/or use GT_LIST
Eric Erhardt [Fri, 11 Dec 2020 02:02:34 +0000 (20:02 -0600)]
P/Invoke function returning SafeSocketHandle throws MissingMethodException (#45911)
* P/Invoke function returning SafeSocketHandle throws MissingMethodException
The ILLinker is trimming the private default constructors of SafeHandle classes if they are not used by the library. Adding an XML descriptor file to preserve the private constructors of these SafeHandles so they can be used in P/Invokes in external code.
Fix #45633
Suppress P/Invoke test on Browser
Santiago Fernandez Madero [Fri, 11 Dec 2020 02:01:16 +0000 (18:01 -0800)]
Add workaround to remove runtime pack from download to use local runtime pack (#45922)
* Add workaround to remove runtime pack from download to use local runtime pack
* PR Feedback
* Add support for full msbuild
Eric Erhardt [Fri, 11 Dec 2020 01:59:30 +0000 (19:59 -0600)]
Remove ILLink warnings from empty console app (#45808)
* Remove ILLink warnings from empty console app
Also clean up some previously addressed warnings in the suppressions xml file.
Contributes to #45623
* Add trimming test for inherited attributes.
Santiago Fernandez Madero [Fri, 11 Dec 2020 01:05:16 +0000 (17:05 -0800)]
Update ServiceModel version on compat pack (#45904)
Koundinya Veluri [Fri, 11 Dec 2020 00:57:19 +0000 (16:57 -0800)]
Do YieldProcessor normalization in preemptive GC mode (#45913)
- The YieldProcessor normalization takes ~10 ms. In the finalizer thread start function, `Thread::HasStarted()` puts the thread into cooperative GC mode. Switch to preemptive GC mode for the normalization.
Fixes https://github.com/dotnet/runtime/issues/42515
Juan Hoyos [Fri, 11 Dec 2020 00:11:17 +0000 (16:11 -0800)]
Fix debugger attach+read on Linux 32 bit (#45506)
The debug PAL was undefining of macros that cause 32 bit POSIX syscalls to use 64 bit offsets. The offsets are defined in a signed manner, which means that any address over the 2GB range that we use will result in an incoherent call to pread and similar APIs. This has been the case for years, but it only affected ARMv7 systems as they are currently our only Unix 32 bit target. Customers started reporting being unable to attach to their processes using the debugger. The bug became a blocker as we changed the debugger startup path to use symbol exports in CoreCLR for our single file story, meaning that the DAC no longer used a compile time header but an export. That read resulted in an access on the >2GB virtual memory region when run in a device with enough memory available with a signed address.
Santiago Fernandez Madero [Thu, 10 Dec 2020 23:51:21 +0000 (15:51 -0800)]
Move iOS tests to staging pipeline (#45903)
Larry Ewing [Thu, 10 Dec 2020 22:31:19 +0000 (16:31 -0600)]
Rename file (#45923)
Steve Pfister [Thu, 10 Dec 2020 21:38:36 +0000 (16:38 -0500)]
Use FORCE_AOT define in the MonoAOTCompiler Task (#45898)
Fixes mono_aot_register_module not being found when AOT'ing.
monojenkins [Thu, 10 Dec 2020 21:06:15 +0000 (16:06 -0500)]
[reflection] Don't get parameter default values for wrapper methods (#45018)
The lookup assumes that the method came from metadata, which is not the case
for wrappers.
Fixes mono/mono#20600
Co-authored-by: lambdageek <lambdageek@users.noreply.github.com>
Viktor Hofer [Thu, 10 Dec 2020 20:05:34 +0000 (21:05 +0100)]
Use ridless PNSE configurations where applicable (#44230)
For projects that require platform support we should use a rid-less
configuration, i.e. $(NetCoreAppCurrent). That guarantees that PNSE
assemblies are available for new configurations that aren't represented
in the defined project's TargetFrameworks. This is especially important
for inbox assemblies.
I left the following assemblies as is as they are partial PNSEs:
- System.Security.Cryptography.Algorithms
- System.Net.NameResolution
System.Net.Requests and System.Net.WebClient are IL only but
intentionally marked as throwing on Browser.
Andrius [Thu, 10 Dec 2020 17:57:19 +0000 (19:57 +0200)]
Simplified MailAddressDisplayNameTest (#39625) (#39929)
* Simplified MailAddressDisplayNameTest (#39625)
* Address PR feedback
* Refactored tests
Co-authored-by: Andrius Lukočius <andrius.lukocius@gmail.com>
Co-authored-by: Stephen Toub <stoub@microsoft.com>
Co-authored-by: Marie Píchová <mapichov@microsoft.com>
Steve MacLean [Thu, 10 Dec 2020 16:10:10 +0000 (11:10 -0500)]
Fix macos-arm64 corrupted adhoc signature (#45860)
XCode linker adhoc signs binaries. Stripping the symbols
corrupts the signature. Replace the adhoc signature
after stripping the symbols.
Viktor Hofer [Thu, 10 Dec 2020 15:54:41 +0000 (16:54 +0100)]
Add Known* items to the local targeting pack resolution (#45874)
* Add Known* items to the local targeting pack resolution
Add the required Known* items so that the ProcessFrameworkReferences
logic in the SDK knows about these and is able to create Resolved* items
that we later update to point to our local packs.
This is required when building with an SDK that doesn't support the
project's TargetFramework yet.
Steve MacLean [Thu, 10 Dec 2020 15:54:28 +0000 (10:54 -0500)]
Fix SOS clrstack w^x issues on osx-arm64 (#45849)
* Fix clrstack w^x issues on osx-arm64
Viktor Hofer [Thu, 10 Dec 2020 15:08:39 +0000 (16:08 +0100)]
Create README.md for src/tasks (#45754)
Mitchell Hwang [Thu, 10 Dec 2020 14:45:05 +0000 (09:45 -0500)]
[wasm] Use dotnet-serve in Wasm Browser Sample (#45128)
* [wasm] Use dotnet serve in wasm browser sample
* [wasm] Remove wasm browser python server
* Check for global dotnet-serve
Co-authored-by: Mitchell Hwang <mitchell.hwang@microsoft.com>
Adeel Mujahid [Thu, 10 Dec 2020 13:52:12 +0000 (15:52 +0200)]
Aggregate adjacent memory sizes regardless of r+x (#45401)
* Aggregate adjacent memory sizes regardless of r+x
When adjacent memory ranges of same module differ by permission, the
line should not be skipped due to the lack of readability/executability
flags.
* Account for rows preceding the first one with r+w
* Account for last line
* Update src/libraries/Common/src/Interop/Linux/procfs/Interop.ProcFsStat.ParseMapModules.cs
Co-authored-by: Tom Deseyn <tom.deseyn@gmail.com>
* Update src/libraries/Common/src/Interop/Linux/procfs/Interop.ProcFsStat.ParseMapModules.cs
* Move module flag assignment after the final commit
* Improve readibility of moduleHasReadAndExecFlags
* Update src/libraries/Common/src/Interop/Linux/procfs/Interop.ProcFsStat.ParseMapModules.cs
* Update src/libraries/Common/src/Interop/Linux/procfs/Interop.ProcFsStat.ParseMapModules.cs
* Decouple line parsing from ProcessModule parsing
Co-authored-by: Tom Deseyn <tom.deseyn@gmail.com>
Co-authored-by: Eirik Tsarpalis <eirik.tsarpalis@gmail.com>
monojenkins [Thu, 10 Dec 2020 12:24:19 +0000 (07:24 -0500)]
Fix debugger late attach, suspend=n. (#45834)
Runtime workaround fix for incorrect VM resume when suspend policy none is in use. Real fix is in debugger-libs, mono/debugger-libs#264 but until that lands in VS installer (and tooling has been upgraded to VS2019) this workaround will prevent the error to be propagated back to client and VM will make sure the correct suspend state is maintained.
Fix also makes sure logging done within debugger engine/agent is not directly send to stdout/stderr on platforms where there is no console (or alternative console).
This is done to prevent scenarios where stdout/stderr has been redirected to for example the debug client, but logging into the same streams as part of debugger agent implementation will trigger new messages to redirect, flooding the message queue and suspending/resuming the VM for each message. This PR makes sure it is possible to get all logging done in debugger agent/engine (not just debug logging) into an alternative log implementation on platforms needing it.
Co-authored-by: lateralusX <lateralusX@users.noreply.github.com>
Ankit Jain [Thu, 10 Dec 2020 11:47:43 +0000 (06:47 -0500)]
[wasm] Add $(MSBUILD_ARGS) to more build targets (#45865)
This can be used to pass additional msbuild arguments when building, or running wasm make targets.
For example: `make -C src/mono/wasm run-tests-System.Collections.Concurrent MSBUILD_ARGS="/p:foo=bar /fl /flp:v=diag"`