- 24 10月, 2022 5 次提交
-
-
由 Stephen Toub 提交于
This reverts commit 8cb4e934.
-
由 Jakob Botsch Nielsen 提交于
repPrintObjectDescription was not null-terminating when the map was null. There is SPMI error trap on the JIT side so we can just make it fail on missing data. repGetStringLiteral should not succeed on missing data, it is used for optimization purposes.
-
由 Ahmet Ibrahim AKSOY 提交于
* Eliminate temporary -1 on Socket.Unix.cs
-
由 Stephen Toub 提交于
* Add ReadOnly{Observable}Collection/Dictionary.Empty * Revert XamlLoadPermission.cs change to fix downlevel build
-
由 Jan Kotas 提交于
-
- 23 10月, 2022 5 次提交
-
-
由 Mitchell Hwang 提交于
Akin to #73067
-
由 Vlad Brezae 提交于
* [mono][interp] Improve GETITEM_SPAN intrinsic Field offsets are known at compile time, they don't need to be included in the code stream. * [mono][interp] Add version of getitemspan that doesn't use indirection * [mono][interp] Add general MonoSpanOfVoid struct and use it in interp opcodes
-
由 SingleAccretion 提交于
Not just the promoted ones.
-
由 Adeel Mujahid 提交于
-
由 Connor Bullard 提交于
Co-authored-by: NDan Moseley <danmose@microsoft.com>
-
- 22 10月, 2022 10 次提交
-
-
由 Jakob Botsch Nielsen 提交于
* JIT: Use metadata names for SIMD intrinsic method recognition eeGetMethodName is only expected to be used for diagnostics. * Reorder checks in impSIMDIntrinsic * Suggested refactoring * Avoid unnecessary metadata lookups in getMethodNameFromMetadata * Fix nested lookup * Fix repGetMethodNameFromMetadata
-
由 Elinor Fung 提交于
-
由 Will Smith 提交于
-
由 Maksym Koshovyi 提交于
-
由 github-actions[bot] 提交于
-
由 Mitchell Hwang 提交于
Android library tests were not running on CI in runtime-extra-platforms. This PR brings back the jobs that ran library tests on Android device/simulator removed in #76328 They are updated to follow the similar changes made to branch out runtime-extra-platforms in the aforementioned PR.
-
由 Jakob Botsch Nielsen 提交于
* Stop using GenTree::ReplaceWith * GenTreeRetExpr now directly contains its substitution members, which is a tree either from the inline candidate, a local (due to spill temps) or the call itself (due to inliner failure). In the former case, it also contains the BasicBlock the child tree comes from to ensure that its flags can be propagated once and for all during substitution. * Make various members strongly typed, e.g. GenTreeRetExpr::gtInlineCandidate is now a GenTreeCall*, InlineCandidateInfo::retExpr is now a GenTreeRetExpr*. This is possible since we are no longer using GenTree::ReplaceWith. Fix #75827
-
由 Ankit Jain 提交于
* [wasm] Fix Wasm.Build.Tests with stable 7.0.0 Set the manifest package version to use (`7.0.0` instead of `7.0.0-ci`) early. Fixes https://github.com/dotnet/runtime/issues/76908 . * Misc fixes based on review feedback Fixes https://github.com/dotnet/runtime/issues/73758 . * Update WorkloadBuildTasksDir to remove 'publish' in the path * workloads-testing.targets: ignore symbols.nupkgs when looking for runtime packs * Fix WBT to have the correct version for WORKLOAD_PACKS_VER * [wasm] Emit an error if PackageVersionForWorkloadManifests property is not set * workloads-testing: Get the PackageVersion at build time .. because it is not set at evaluation time. So, get it in a target, and move the WorkloadIds to a target too. * Ensure the version is set before generating the runscript for WBT
-
由 Jan Kotas 提交于
-
由 Andrew Au 提交于
-
- 21 10月, 2022 20 次提交
-
-
由 Alhad Deshpande 提交于
[ppc64le] Implementation of mono_arch_get_delegate_virtual_invoke_impl method for ppc64le architecture (#77308) * [ppc64le] Fixed thunk address 8 byte alignment issue * Fixed FSharp crash issue * [ppc64le] Implementation of mono_arch_get_delegate_virtual_invoke_impl method for ppc64le architecture
-
由 Andrew Au 提交于
-
由 Jan Vorlicek 提交于
This function is never used.
-
由 Andrew Au 提交于
-
由 Adeel Mujahid 提交于
-
由 Zoltan Varga 提交于
The concrete method cannot be looked up so 'cmethod' is still the interface method. Related: https://github.com/dotnet/runtime/pull/76724.
-
由 Jan Kotas 提交于
-
由 Jan Vorlicek 提交于
It turned out that my recent change to redefine __int64 on Unix is actually correct for macOS only. For other 64 bit Unixes, the __int64 should be defined as long. This change fixes it.
-
由 Koundinya Veluri 提交于
A race condition when the object header is transitioning to a sync block can cause a concurrent recursive `TryEnter` to return false. Fixes https://github.com/dotnet/runtime/issues/76660
-
由 Aaron Robinson 提交于
* Remove uses of PAL's swprintf_s from mono.
-
由 Gregory LEOCADIE 提交于
* Add new version of the ContentionStart event and a LockCreated event The new ContentionStart includes a lock ID, associated object ID, and the holding thread ID. The LockCreated event includes a lock ID and associated object ID. Co-authored-by: NKoundinya Veluri <kouvel@microsoft.com>
-
由 Egor Bogatov 提交于
-
由 Jakob Botsch Nielsen 提交于
* Fix dotnet-pgo dump * Fix merge compare-mibc and create-mibc too * A couple more fixes
-
由 Maryam Ariyan 提交于
-
由 Jan Vorlicek 提交于
Currently the ld linker links incorrectly the compiled thunktemplates.S for ARM64 and puts wrong offsets into all the LDR instructions. Only the LLVM lld linker works correctly. I have looked into it and it turns out the assembler was putting in relocations that the linker then replaced by fixed offsets. That's where the GNU ld made the mistake. I have figured out a slight modification of expressing the addresses in the data page referenced by those LDR instructions so that the assembler doesn't produce any relocs and so the linker cannot cause any problems.
-
由 Andrew Au 提交于
-
由 Adeel Mujahid 提交于
* Trim file type bits from mode header * Define and use ValidUnixFileModes * Rev System.Formats.Tar.TestData version * Move ValidUnixFileModes to shared helpers
-
由 CyrusNajmabadi 提交于
-
由 Chuck Ries 提交于
Adding [DebuggerStepThrough] to these constructors allows debuggers to avoid stepping into them and stopping regardless of Just My Code settings. This is desirable for debuggers as there is little benefit to stopping in these constructors as they are primarily infrastructure inserted by the compiler.
-
由 Daniel Ramos 提交于
* Initial fix for null reference exception * Took away lazy and put back private constructor * Added parent property and made handler thread safe
-