Alexander Köplinger [Wed, 29 Jul 2020 14:44:16 +0000 (16:44 +0200)]
[wasm] Remove unused icalls for Environment properties (#40079)
They are no longer needed after https://github.com/dotnet/runtime/pull/38996 since we return values in managed code now.
Geoff Kizer [Wed, 29 Jul 2020 11:31:25 +0000 (04:31 -0700)]
BindAsync -> ListenAsync (#40068)
BindAsync -> ListenAsync
Marie Píchová [Wed, 29 Jul 2020 10:41:42 +0000 (12:41 +0200)]
Do not test sockets handler fix on WinHttpHandler. (#40070)
Santiago Fernandez Madero [Wed, 29 Jul 2020 09:55:56 +0000 (02:55 -0700)]
WASM: Enable GetAssemblyNameTest_ValidAssembly and add hook to include custom files into vfs (#40032)
Vitek Karas [Wed, 29 Jul 2020 09:47:49 +0000 (02:47 -0700)]
Redirect DllImport of hostpolicy to the main executable when it's embedded (#40014)
Ryan Lucia [Wed, 29 Jul 2020 09:19:08 +0000 (05:19 -0400)]
[mono] Disable cfgdir config on Windows netcore builds (#40005)
This was already disabled with autotools, but I forgot to update winconfig.h as well.
Maxim Lipnin [Wed, 29 Jul 2020 09:10:35 +0000 (12:10 +0300)]
[wasm] Re-enable some tests in the System.Net.WebSockets.Tests namespace (#40004)
There is no repro for https://github.com/dotnet/runtime/issues/38807 at the moment so the related tests have been enabled.
Two tests have been disabled due to PNSE (see https://github.com/dotnet/runtime/pull/39346).
Fixes: https://github.com/dotnet/runtime/issues/38807
The test run result: `Tests run: 177, Errors: 0, Failures: 0, Skipped: 0.`
Marie Píchová [Wed, 29 Jul 2020 08:16:04 +0000 (10:16 +0200)]
Disabled outerloop test. (#40045)
Katelyn Gadd [Wed, 29 Jul 2020 05:57:13 +0000 (22:57 -0700)]
Update threadpool callback error messages to fix #38803 (#40000)
Katelyn Gadd [Wed, 29 Jul 2020 05:56:41 +0000 (22:56 -0700)]
Assert that we don't get the value we don't want instead of attempting to check for a specific error code. Fixes #39955 (#40019)
Kenneth Pouncey [Wed, 29 Jul 2020 05:05:31 +0000 (07:05 +0200)]
[browser][wasm] Configuring request options in Browser WebAssembly (#39182)
* [browser][wasm] Initial addition of configuring request options in Browser WebAssembly
* Fix key code. Not what was proposed
* Add TryGetValue<TValue> and Set<TValue> as per proposal
* Add missing obsolete attribute
* Address review comments
* Update tests to use Options and not obsolete Properties.
* Implement IDictionary<string, object?> explicitly
* Update tests to use Options and not obsolete Properties.
* Add HttpRequestOptions source to the System.Net.Http.Unit.Tests project to fix build.
* Address review comments - explicit
* Fix build error cannot convert from 'string' to 'System.Net.Http.HttpRequestOptionsKey<object>'
* Add tests for HttpRequestOptions
* Fix test build
* Add special case code for NETFRAMEWORK for API change.
* #endif out of place fix
* #endif out of place fix
Sergio Pedri [Wed, 29 Jul 2020 03:06:52 +0000 (05:06 +0200)]
Add Unsafe.IsNullRef and Unsafe.NullRef (#40008)
Co-authored-by: Adeel Mujahid <adeelbm@outlook.com>
Co-authored-by: Jan Kotas <jkotas@microsoft.com>
Tarek Mahmoud Sayed [Wed, 29 Jul 2020 02:15:05 +0000 (19:15 -0700)]
Fix creating RegionInfo object using names with different casing. (#40052)
yowl [Tue, 28 Jul 2020 23:16:45 +0000 (18:16 -0500)]
Negate find_object GetConservativeGC logic. (#39905)
Zoltan Varga [Tue, 28 Jul 2020 23:11:00 +0000 (19:11 -0400)]
[interp] Add null checks to STIND opcodes. (#40041)
Jan Kotas [Tue, 28 Jul 2020 23:10:38 +0000 (16:10 -0700)]
Cleanup mscorlib references under src\coreclr\src (#39994)
* s/MscorlibBinder/CoreLibBinder/
* s/g_Mscorlib/g_CoreLib/
* Remaining mscorlib->corelib renames
* Delete unreachable FNV NGen string
Elinor Fung [Tue, 28 Jul 2020 22:57:17 +0000 (15:57 -0700)]
Fix parsing of embedded .clsidmap in comhost (#40022)
* Fix parsing of embedded .clsidmap in comhost
* Update tests to use HostModel to create/embed clsidmap
Jarret Shook [Tue, 28 Jul 2020 22:38:43 +0000 (15:38 -0700)]
Disable based on #40034 (#40035)
monojenkins [Tue, 28 Jul 2020 22:35:56 +0000 (18:35 -0400)]
[debugger] Revert Removing unhandled_exception which was used for android (#40027)
This reverts commit
60643c36f5d4e349276e5dcce0e7fc3dc1d72f74.
To investigate error on android tests.
Co-authored-by: thaystg <thaystg@users.noreply.github.com>
Thays Grazia [Tue, 28 Jul 2020 21:38:41 +0000 (18:38 -0300)]
Manually merging driver.c from https://github.com/mono/mono/pull/20159 (#40031)
Dan Moseley [Tue, 28 Jul 2020 21:33:36 +0000 (14:33 -0700)]
Update visualstudio.md (#40047)
Alexander Köplinger [Tue, 28 Jul 2020 21:27:17 +0000 (23:27 +0200)]
[wasm] Add missing EMCC_FLAGS variable when building corebindings.o (#40024)
Egor Bogatov [Tue, 28 Jul 2020 21:23:56 +0000 (00:23 +0300)]
Intrinsify MemoryMarshal.GetArrayDataReference for interp (#39920)
dotnet-maestro[bot] [Tue, 28 Jul 2020 20:45:31 +0000 (13:45 -0700)]
Update dependencies from https://github.com/dotnet/icu build
20200728.2 (#40011)
Microsoft.NETCore.Runtime.ICU.Transport
From Version 5.0.0-preview.8.20373.1 -> To Version 5.0.0-preview.8.20378.2
Co-authored-by: dotnet-maestro[bot] <dotnet-maestro[bot]@users.noreply.github.com>
Bruce Forstall [Tue, 28 Jul 2020 20:31:41 +0000 (13:31 -0700)]
Enable dumping GC tables in release builds (#39946)
* Dump GC info always in release builds
Add support for dumping non-x86 info in release builds
* Revert DUMP_GC_TABLES change
* Fix formatting
Ryan Lucia [Tue, 28 Jul 2020 19:57:01 +0000 (15:57 -0400)]
[mono] Fix warnings on OSX netcore (#39997)
This commit brings us back to zero warnings for the default build configuration
Juan Hoyos [Tue, 28 Jul 2020 19:25:48 +0000 (12:25 -0700)]
Stop using transport packages for CoreCLR assets (#40002)
Tomas Weinfurt [Tue, 28 Jul 2020 19:14:04 +0000 (12:14 -0700)]
use empty server name is client did not specify one (#39671)
* use empty server name
* fix merge
* feedback from review
* add missing file
Alexander Nikolaev [Tue, 28 Jul 2020 18:03:37 +0000 (20:03 +0200)]
Increase HTTP/2 connection expiration test's timeouts (#40015)
Elinor Fung [Tue, 28 Jul 2020 17:21:56 +0000 (10:21 -0700)]
Error on multiple calling conventions in modopts (#39809)
David Wrighton [Tue, 28 Jul 2020 17:07:39 +0000 (10:07 -0700)]
Fix mdarrays (#39984)
- MDArrays are special and method references to the methods needs to be encoded with their type in more cases than normal methods
- Only generate the owner type if the memberref does not encode the same owner type
- Add tests for various multidimensional array scenarios to crossgen2smoke
Fix issue #38260
Paulo Morgado [Tue, 28 Jul 2020 16:19:18 +0000 (17:19 +0100)]
Fix typo on comment on ReflectionEnum::InternalHasFlag (#39995)
Cory Nelson [Tue, 28 Jul 2020 15:06:57 +0000 (08:06 -0700)]
Initial (partially-reviewed API) System.Net.Connections. (#39524)
Shay Rojansky [Tue, 28 Jul 2020 15:04:00 +0000 (18:04 +0300)]
Annotate System.ComponentModel.DataAnnotations for nullability (#39611)
Marie Píchová [Tue, 28 Jul 2020 14:58:43 +0000 (16:58 +0200)]
Expect 100 Continue Hang (#38774)
Inside sendRequestContentTask recognizes that it's invoked from a timer thread and if request content fails, unblocks SendAsyncCore and eventually propagates the exception from the content to the outside.
Fixes the issue for H2 as well.
Fixes #36717
Alexander Nikolaev [Tue, 28 Jul 2020 14:55:12 +0000 (16:55 +0200)]
System.Net.Sockets telemetry counters (#39708)
Counters in SocketsTelemetry measuring
`bytes-received` (Bytes Received)
The cumulative total number of bytes received by all Socket objects since the process started.
`bytes-sent` (Bytes Sent)
The cumulative number of bytes sent by all Socket objects since the process started.
`outgoing-connections-established` (Outgoing Connections Established)
The cumulative total number of outgoing connections established by Socket objects for stream sockets since the process started.
`incoming-connections-established` (Incoming Connections Established)
The cumulative total number of incoming connections established by Socket objects for stream sockets since the process started.
`datagrams-received` (Datagrams Received)
The cumulative total number of datagram packets received by all Socket objects since the process started.
`datagrams-sent` (Datagrams Sent)
The cumulative total number of datagram packets sent by all Socket objects since the process started.
Contributes to #37428
Anton Firszov [Tue, 28 Jul 2020 14:42:31 +0000 (16:42 +0200)]
Adapt RFC 6265 in path handling for cookies (#39250)
Resolves "Path related issues" of #26141
(described in #21440, #22372 and #25226) by adapting RFC 6265 for path management:
- Removes the restriction for the Path attribute (it's no longer expected to prefix the request path)
- Introduces the default-path calculation and path matching algorithms as defined in section 5.1.4 of the new RFC
- Adds integration tests for HttpClient based on user scenarios described in the issues
monojenkins [Tue, 28 Jul 2020 12:34:29 +0000 (08:34 -0400)]
Apple Silicon support (#39939)
<!--
Thank you for your Pull Request!
If you are new to contributing to Mono, please try to do your best at conforming to our coding guidelines http://www.mono-project.com/community/contributing/coding-guidelines/ but don't worry if you get something wrong. One of the project members will help you to get things landed.
Does your pull request fix any of the existing issues? Please use the following format: Fixes #issue-number
-->
Co-authored-by: vargaz <vargaz@users.noreply.github.com>
Juan Hoyos [Tue, 28 Jul 2020 11:19:58 +0000 (04:19 -0700)]
Add ICorDebugModule4 - Get Module mapping mode (#39865)
* Add IDL for ICorDebugModule4
* Add cached version of the MIDL generated stubs
* Implement ICorDebugModule4
* Align method naming with other methods in ICorDebug
Ryan Lucia [Tue, 28 Jul 2020 11:07:25 +0000 (07:07 -0400)]
Add brew update before brew bundle (#39999)
Every lane appears to be failing because of an issue with brew bundle expecting the latest homebrew but not automatically updating: https://github.com/Homebrew/homebrew-bundle/issues/751
So forcibly update, which is apparently good practice anyway: https://github.com/Homebrew/homebrew-bundle/issues/751#issuecomment-
664958735
Alexander Nikolaev [Tue, 28 Jul 2020 10:43:39 +0000 (12:43 +0200)]
[retry only] Additional HTTP/2 connections created when active streams limit is reached (#39439)
HTTP/2 standard commands clients to not open more than one HTTP/2 connection to the same server. At the same time, server has right to limit the maximum number of active streams per that HTTP/2 connection. These two directives combined impose limit on the number of requests concurrently send to the server. This limitation is justified in client to server scenarios, but become a bottleneck in server to server cases like gRPC. This PR introduces a new SocketsHttpHandler API enabling establishing additional HTTP/2 connections to the same server when the maximum stream limit is reached on the existing ones.
**Note**. This algorithm version uses only retries to make request choose another connection when all stream slots are occupied. It does not implement stream credit management in `HttpConnectionPool` and therefore exhibit a sub-optimal request scheduling behavior in "request burst" and "infinite requests" scenarios.
Fixes #35088
Kenneth Pouncey [Tue, 28 Jul 2020 09:21:41 +0000 (11:21 +0200)]
[browser][http] Fix HttpClientHandler "Supports*" properties regressions (#39889)
* [browser][http] Fix HttpClientHandler "Supports*" properties regressions
* Also needs implementing on SocketsHttpHandler.
On the SocketsHttpHandler we will default to `true` here.
```
internal bool SupportsAutomaticDecompression => true;
internal bool SupportsProxy => true;
internal bool SupportsRedirectConfiguration => true;
```
* Should be internal accessor
Peter Sollich [Tue, 28 Jul 2020 09:05:19 +0000 (11:05 +0200)]
Fix issue in sort_mark_list causing AVs in merge_mark_lists.
This happens during gen 2 GCs where at least one GC thread sees no surviving objects. The bug causes merge_mark_lists to attempt the merge when in fact we are not using the mark lists at all during gen 2 GCs.
Fix comment in sort_mark_list.
Vitek Karas [Tue, 28 Jul 2020 08:13:10 +0000 (01:13 -0700)]
Fix ILLink descriptors for Mono.CoreLib (#39898)
Alexis Christoforides [Tue, 28 Jul 2020 05:22:22 +0000 (01:22 -0400)]
[mono] Use DefaultDllSearchPaths attribute for PInvokes (#38975)
* [mono] Use DefaultDllSearchPaths attribute for PInvokes
So far we have been ignoring the flags contained in the attribute.
Attribute can be applied to a pinvoke method or an assembly. See https://docs.microsoft.com/en-us/dotnet/api/system.runtime.interopservices.defaultdllimportsearchpathsattribute.-ctor?view=netcore-3.1#System_Runtime_InteropServices_DefaultDllImportSearchPathsAttribute__ctor_System_Runtime_InteropServices_DllImportSearchPath_
* Update src/mono/mono/metadata/native-library.c
Co-authored-by: Aleksey Kliger (λgeek) <akliger@gmail.com>
* Minor fixes to comments and error handling
* Additional fixes
Co-authored-by: Aleksey Kliger (λgeek) <akliger@gmail.com>
Co-authored-by: Ryan Lucia <rylucia@microsoft.com>
Jan Kotas [Tue, 28 Jul 2020 04:58:32 +0000 (21:58 -0700)]
Update safemath.h
Fixes #39985
monojenkins [Tue, 28 Jul 2020 02:00:48 +0000 (22:00 -0400)]
[wasm] Use interp-only instead of interp-llvmonly as the default execution mode. (#39880)
interp-llvmonly caused mono_llvm_only to be set, causing bad behavior like
stack walks not working.
Co-authored-by: vargaz <vargaz@users.noreply.github.com>
Elinor Fung [Tue, 28 Jul 2020 01:12:25 +0000 (18:12 -0700)]
Fix releasing of file handle in binder tracing (#39974)
Zoltan Varga [Tue, 28 Jul 2020 00:20:28 +0000 (20:20 -0400)]
[jit] Add a intrinsic for ThrowForUnsupportedVectorBaseType<T>. (#39965)
Part of https://github.com/dotnet/runtime/issues/38718.
Bruce Forstall [Tue, 28 Jul 2020 00:12:15 +0000 (17:12 -0700)]
Fix superpmi.py script for asmdiffs (#39948)
Bruce Forstall [Mon, 27 Jul 2020 23:33:38 +0000 (16:33 -0700)]
Disable eventpipe test under GCStress due to test failure (#39978)
Issue: https://github.com/dotnet/runtime/issues/39932
Bruce Forstall [Mon, 27 Jul 2020 23:32:36 +0000 (16:32 -0700)]
Disable gcdump test under GCStress due to test failure (#39977)
Issue: https://github.com/dotnet/runtime/issues/39931
Juan Hoyos [Mon, 27 Jul 2020 23:01:04 +0000 (16:01 -0700)]
Remove GetIPCEventSendBufferContinuation (#39662)
* Fix write on failure to allocate buffer
* Remove GetIPCEventSendBufferContinuation altogether
Alexander Köplinger [Mon, 27 Jul 2020 22:59:51 +0000 (00:59 +0200)]
[mono] Fix building multiple os/arch combinations in the same working directory (#39970)
After adding the ICU shim code to the runtime we hit an issue when e.g. compiling for "desktop" mono in a working directory that already had Browser wasm artifacts in it.
This is due to automake putting the intermediate compilation artifacts at the same location where a file was referenced from.
This means that e.g. when using `../../../libraries/Native/Unix/System.Globalization.Native/pal_icushim.c` in the Makefile.am we'd get the intermediaries in `artifacts/obj/mono/libraries/` instead of somewhere in `artifacts/obj/mono/OSX.x64.Debug`.
Later on we'd get the following error because it was reusing the existing .o file from another architecture:
```
Undefined symbols for architecture x86_64:
"_gPalGlobalizationNative", referenced from:
_c_qcalls in libmonoruntimesgen.a(libmonoruntimesgen_la-native-library-qcall.o)
```
Fix this by symlinking the source files into the build directory.
Eric StJohn [Mon, 27 Jul 2020 22:37:01 +0000 (15:37 -0700)]
Fix issue with Crossgen2 package excluding Icon (#39972)
The latest packaging targets rely on ProjectDefaults inclusion of package icon,
but this item was being removed by the Crossgen2 project's targets.
Also updating the version of packaging tools to see if we have any other failures.
Maryam Ariyan [Mon, 27 Jul 2020 21:10:59 +0000 (14:10 -0700)]
Keep only Abstractions dependency on Configuration.Binder (#39461)
David Cantu [Mon, 27 Jul 2020 21:02:32 +0000 (14:02 -0700)]
Add nullability annotations to files in XPath folder (#39821)
* Add nullability annotations to files in XPath folder
* Address feedback from krwq
* Make changes based on recently merged annotations
* Mark resolver parameter in SelectSingleNode as nullable
Maryam Ariyan [Mon, 27 Jul 2020 19:18:57 +0000 (12:18 -0700)]
Fix JSON possible dupe key issue (#39860)
Maxim Lipnin [Mon, 27 Jul 2020 19:07:11 +0000 (22:07 +0300)]
[wasm] Mark DiagnosticSource.Switches.Tests with an active issue (#39952)
dotnet-maestro[bot] [Mon, 27 Jul 2020 17:59:25 +0000 (19:59 +0200)]
[master] Update dependencies from Microsoft/vstest mono/linker dotnet/icu (#39697)
Co-authored-by: dotnet-maestro[bot] <dotnet-maestro[bot]@users.noreply.github.com>
Ganbarukamo41 [Mon, 27 Jul 2020 15:26:22 +0000 (00:26 +0900)]
Fix Half comparison (#39773)
* Fix Half comparison
* Fixes the comparison to be correct when both numbers are negative
* Add relevant tests
* Add parentheses for clarity
* Add tests validating comparison between negative numbers for Single/Double as well
Viktor Hofer [Mon, 27 Jul 2020 14:38:45 +0000 (16:38 +0200)]
Fix build on some libs projects when metadata isn't available (#39960)
* Fix build on some libs projects when metadata isn't available
* Update references.targets
Maxim Lipnin [Mon, 27 Jul 2020 12:08:16 +0000 (15:08 +0300)]
[wasm] Remove System.ValueTuple.Tests test suite from the exclusion list (#39957)
Shay Rojansky [Mon, 27 Jul 2020 09:09:50 +0000 (12:09 +0300)]
Annotate System.Data.{Odbc,OleDb} for nullability (#39597)
* Annotate System.Data.Odbc for nullability
* Annotate System.Data.OleDb for nullability
* Suppress nullable warnings on netcoreapp < 3.0
* Mark System.Data.DataSetExtensions as null-annotated
* Fix null-check bug in System.Data.OleDb
Tomáš Rylek [Mon, 27 Jul 2020 06:58:01 +0000 (08:58 +0200)]
Mark the multifolder GC2 test as GC stress incompatible until fixed (#39942)
Bruce Forstall [Mon, 27 Jul 2020 03:47:32 +0000 (20:47 -0700)]
Disable reverseouter test under GCStress (#39937)
Issue: https://github.com/dotnet/runtime/issues/38801
Bruce Forstall [Mon, 27 Jul 2020 03:46:59 +0000 (20:46 -0700)]
Disable processinfo for ilasm round-trip testing (#39936)
Issue: https://github.com/dotnet/runtime/issues/39935
Bruce Forstall [Sun, 26 Jul 2020 18:39:41 +0000 (11:39 -0700)]
Disable Microsoft.Extensions.Caching.Memory.CapacityTests.DoNotAddIfSizeOverflows (#39930)
Issue: https://github.com/dotnet/runtime/issues/33993,
https://github.com/dotnet/runtime/issues/39241
Josh Schreuder [Sun, 26 Jul 2020 18:38:15 +0000 (04:38 +1000)]
Mark Assembly.CodeBase / Assembly.EscapedCodeBase as obsolete (#31127) (#39261)
Scott Xu [Sun, 26 Jul 2020 16:56:57 +0000 (00:56 +0800)]
Use consistent access modifiers in AsnValueReader.cs
Matt Kotsenas [Sat, 25 Jul 2020 18:39:57 +0000 (11:39 -0700)]
Remove unused locals in System.Numerics.Tensors (#39578)
Viktor Hofer [Sat, 25 Jul 2020 18:34:14 +0000 (20:34 +0200)]
Fix CoreLib path resolution (#39921)
Fixes issue described in https://github.com/dotnet/runtime/pull/39891#issuecomment-
663767530.
Egor Chesakov [Sat, 25 Jul 2020 18:33:55 +0000 (11:33 -0700)]
[Arm64] Don't generate hardware intrinsic method bodies (#39753)
* Don't generate hardware intrinsic method bodies on Arm64 in zapinfo.cpp
* Treat Vector64 and Vector128 methods as intrinsics on Arm64 in zapinfo.cpp
David Wrighton [Sat, 25 Jul 2020 18:30:34 +0000 (11:30 -0700)]
Address slowdowns in type loader performance (#39841)
Hugh Bellamy [Sat, 25 Jul 2020 18:28:58 +0000 (19:28 +0100)]
Remove CoreMangLib duplicate tests (#36612)
* Cleanup duplicated Type tests
* Baseline mono test failures
Vitek Karas [Sat, 25 Jul 2020 18:11:14 +0000 (11:11 -0700)]
Add linker annotation to LazyDebugView to avoid warnings (#39899)
In reality this is only to get rid of the warnings, the class is never instatiated by the managed code, so the annotation has no effect on including more code.
When in debugger the class is instatiated with the same T of an existing Lazy<T> which has the same annotation, so the T will always fulfill the annotation's requirements.
Mike McLaughlin [Sat, 25 Jul 2020 04:01:54 +0000 (21:01 -0700)]
MacOS managed debugging perf fix (#39804)
* MacOS managed debugging perf fix
This PR adds PAL_OpenProcessMemory, PAL_ReadProcessMemory and PAL_CloseProcessMemory that
encapsulate the remote read memory functions for both MacOS and Linux.
The DBI code that reads memory using the runtime IPC messages to use these new functions
and fallback to the IPC message if the open fails.
On MacOS, this won't have any affect on VSCode performance until vsdbg/vsdbgui are signed
with the "com.apple.security.cs.debugger" entitlement.
Sample entitlements.plist file:
```
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
<key>com.apple.security.cs.debugger</key>
<true/>
</dict>
</plist>
```
Example of locally signing with entitlement command:
codesign -s <local-cert> --entitlements entitlements.plist
Fix FreeBSD build. Remove FEATURE_DATATARGET4 for macOS so managed debugging uses faster OOP unwind for HelpMethodFrames
Eugene Rozenfeld [Sat, 25 Jul 2020 03:55:49 +0000 (20:55 -0700)]
Fix `fgUpdateChangedFlowGraph`. (#39878)
`fgComputeDoms` has an assumption that the flow graph
has no unreachable blocks. It's checked with this assert:
https://github.com/dotnet/runtime/blob/
ad325b014124b1adb9306abf95fdac85e3f7f2f4/src/coreclr/src/jit/flowgraph.cpp#L2342
This assert fired when testing #39474 (`Convert Math{F}.CoreCLR methods from FCall to QCall`)
when we are updating the flow graph after inserting GC polls.
This change switches `fgUpdateChangedFlowGraph` to call `fgComputeReachability`,
which both removes unreachable blocks and calls `fgComputeDoms`.
pin-icount reported a 0.0043% throughput improvement, which is within noise level.
David Wrighton [Sat, 25 Jul 2020 03:31:58 +0000 (20:31 -0700)]
Fix CanCastTo behavior (#39915)
* Fix missing cases in CanCastTo
* Update issues.targets to start running the previously broken tests
Maryam Ariyan [Sat, 25 Jul 2020 01:26:56 +0000 (18:26 -0700)]
Avoid using DateTime - Switch to DateTimeOffset (#39916)
David Wrighton [Sat, 25 Jul 2020 00:12:19 +0000 (17:12 -0700)]
Fix issue in type equivalence involving arrays (#39914)
- Fix issue where single dimensional arrays and multidimensional arrays of rank 1 are considered equivalent
Krzysztof Wicher [Fri, 24 Jul 2020 23:18:45 +0000 (01:18 +0200)]
Nullable: System.Xml, part 5 (XmlDocument) (#39691)
* Nullable: System.Xml, part 5 (XmlDocument)
* apply feedback
Maryam Ariyan [Fri, 24 Jul 2020 23:11:53 +0000 (16:11 -0700)]
AnsiParser improvement (#39729)
Egor Chesakov [Fri, 24 Jul 2020 22:48:32 +0000 (15:48 -0700)]
Fix issues with JitStressRange (#39754)
* Add 10 when digit is [a-fA-F] in ConfigMethodRange::InitRanges in utils.cpp
* Zero out compMethodHashPrivate earlier in Compiler::compInit() in compiler.cpp
Jan Kotas [Fri, 24 Jul 2020 22:29:24 +0000 (15:29 -0700)]
Delete redundant MSVC compiler options (#39901)
Eric StJohn [Fri, 24 Jul 2020 22:17:43 +0000 (15:17 -0700)]
Ensure CoreCLR pkgproj include packaging props (#39906)
Since these projects don't directly consume the nupkg they didn't
automatically get props when we added it.
Sergey Andreenko [Fri, 24 Jul 2020 21:45:48 +0000 (14:45 -0700)]
Fix lcl fld addr. (#39424)
* Add a repro test.
* Small ref.
Combine long chains of `addr->OperGet() == GT_* ||` with `GT_LCL_VAR_ADDR` using `OperIs` to simplify future changes.
* Add an assert that would fire in the repro test.
`emitter::emitHandleMemOp` has special logic for contained `memBase` and but the last block does not
expect a contained node. A contained node doesn't produce a register so it is not correct to use result
of `GetRegNum()` from a contained node as a valid register.
However, adding an assert to `GetRegNum()` that `!this->isContained` is a bigger task that is out of this PR.
* Assert that `LCL_FLD_ADDR` is not contained in `genPutArgStk(Split)`
We have contained `LCL_VAR_ADDR` support there but make sure that contained `LCL_FLD_ADDR` can't reach it.
* Contain `GT_LCL_FLD_ADDR` under HW_INTRINSIC.
This is an additional optimization that makes future changes simpler.
* Add contained checks.
In all these places we expect `LCL_VAR_ADDR` to be contained.
If we had gotten a `LCL_VAR_ADDR` that is not contained we would have instantiated `LCL_VAR_ADDR` twice:
in the register and the parent instruction.
The register value would have been unused.
* Support `FLD_ADDR` where `LCL_ADDR` is supported.
However, fire an assert if we think that this path is unreachable for now.
* Delete asserts in the reachable blocks.
We have coverage for this asserts in the following tests:
hwintrinsic 478: Ssse3_ro
instr 11645: Runtime_39403
instr 1028 : Aes_ro
hwintrinsic 716: pmi of Microsoft.Diagnostics.Tracing.TraceEvent
* Review response.
* Add repro cases.
Delete the rest `assert(!"don't expect GT_LCL_FLD_ADDR");`.
* Use `GetLclOffs` from `LclVarCommon`.
* missed file.
Alexander Köplinger [Fri, 24 Jul 2020 20:31:59 +0000 (22:31 +0200)]
Remove setting ICUCORE for macOS in System.Globalization.Native CMake config (#39900)
Follow-up to #39833.
We don't link against ICU on macOS but load it via `dlopen()` at runtime and compilation uses headers from homebrew.
That means `HAVE_SET_MAX_VARIABLE` and `HAVE_UDAT_STANDALONE_SHORTER_WEEKDAYS` will always be defined and we don't need the ICUCORE variable.
* PR feedback
Co-authored-by: Adeel Mujahid <adeelbm@outlook.com>
Eugene Rozenfeld [Fri, 24 Jul 2020 20:19:43 +0000 (13:19 -0700)]
Fix GC Poll inlining. (#39881)
* Don't inline GC polls in cold basic blocks.
* Allow GC poll inlining in basic blocks with `BBF_LOOP_PREHEADER` or `BBF_RETLESS_CALL` set.
This fixes one of the assert seen in https://github.com/dotnet/runtime/pull/39474#issuecomment-
662033060
Contributes to resolving #39726.
Jeremy Koritzinsky [Fri, 24 Jul 2020 19:30:37 +0000 (12:30 -0700)]
Update to unified Alpine build image version (#39903)
Update our Alpine build image so we can unify for a single image for our builds.
Fixes #2030
Andy Ayers [Fri, 24 Jul 2020 17:59:16 +0000 (10:59 -0700)]
Jit: fix some sources of x64 chk/rel diffs (#39888)
Break IGs for perfscores the same way in chk and rel. Also, fix the xarch
peephole to work across extended IGs.
Closes #39845.
Brian Sullivan [Fri, 24 Jul 2020 17:21:25 +0000 (10:21 -0700)]
New fix for duplicate methods in the AltJit jit-diff output (#39751)
Alan West [Fri, 24 Jul 2020 16:25:05 +0000 (09:25 -0700)]
Fix ActivitySource.StartActivity when start time is provided (#39884)
Viktor Hofer [Fri, 24 Jul 2020 16:02:32 +0000 (18:02 +0200)]
Post #35606 cleanup (#39891)
* Post #35606 cleanup
Adeel Mujahid [Fri, 24 Jul 2020 15:35:17 +0000 (18:35 +0300)]
Convert math intrinsics to named intrinsics (#39730)
* Convert math intrinsics to named intrinsics
* Annotate Floor and Ceiling with [Intrinsic]
monojenkins [Fri, 24 Jul 2020 14:42:10 +0000 (10:42 -0400)]
[debugger] Removing unhandled_exception which was used for android. (#39377)
Trying to remove the usage of unhandled_exception function to make debugger handle with an exception when running on android.
We should ignore the try catch that is in an WRAPPER_SUBTYPE_ICALL_WRAPPER, then we can walk through all the callstack and find any try catch in managed code if it exists.
Co-authored-by: thaystg <thaystg@users.noreply.github.com>
Eirik Tsarpalis [Fri, 24 Jul 2020 13:56:12 +0000 (14:56 +0100)]
Add CBOR property-based tests (#39828)
* Add CBOR property-based tests
* address feedback
Kenneth Pouncey [Fri, 24 Jul 2020 09:50:41 +0000 (11:50 +0200)]
[browser][tests] Activate System.Private.Uri.Functional.Tests (#39817)
Local:
```
info: test[0]
Tests run: 883, Errors: 0, Failures: 0, Skipped: 1. Time: 188.7140968s
```
Zoltan Varga [Fri, 24 Jul 2020 09:23:38 +0000 (05:23 -0400)]
[runtime] Fix some more gc tracking problems in create_cattr_named/typed_arg. (#39856)
Natalia Kondratyeva [Fri, 24 Jul 2020 07:18:00 +0000 (10:18 +0300)]
Remove leading dot check for cookie domain (#39781)
Minimal fix for domain-related cookie issues of #26141
To fully comply with RFC 6265, one should remove deprecated cookie
properties, such as Version, from public API. So only the stated
issues with leading dot were addressed now.
Also note that the leading dot was not stripped from the domain even
though RFC 6265 proposed it. This behavior was chosen because
browsers like Chrome and Edge also don't strip the leading dot.