Age | Commit message (Collapse) | Author | Files | Lines |
|
The code was incorrectly using signed conditions, which doesn't work for
server GC where the high ephemeral limit is 0xffffffffffffffff and so it
was checking if the address is less than -1, which is never the case. So
the card table was never updated.
|
|
* Enable Invoke and GetValue for ref-returning members
https://github.com/dotnet/corefx/issues/15960
Returned magic object is the object pointed to by
the ref. If the ref is null, NullReferenceException.
* Fix error message
* Found way to reuse InvokeUtil::CreateObject
* Remove comment, BYREF case.
|
|
[arm32] Disable corefx Microsoft.Win32.SystemEvents.Tests
|
|
[Arm64/Linux] Enable crosscomponent build in CI
|
|
Disallow Ubuntu arm32 non-PR jobs in non-master branches
|
|
|
|
Disallow Ubuntu arm32 non-PR jobs in non-master branches
|
|
Need to include the `-x64` in the output rid.
|
|
Unix/x64 ABI cleanup
|
|
preview1-26418-01, preview3-26418-01, master-20180418-0040, respectively (#17621)
|
|
|
|
|
|
|
|
* Ensure thumb bit is set when searching for debugger jit info on ARM
* Make GetCodeStartAddress return correct PCODE, rather than fixing it at the callee site
* Fix type in daccess.h
* Use PINSTRToPCODE rather than manually adding the thumb bit
|
|
Adjust non-PR triggers
|
|
operations (#17602)
|
|
Add linux-musl build leg
|
|
This saves the JIT from hardcoding the logic on where getTailCallCopyArgsThunk is implemented
|
|
1. Enable weekly Ubuntu arm32 GCStress jobs.
2. Fix bugs where both flow and non-flow jobs were being triggered for some configurations.
3. Convert Windows ARM push jobs to daily jobs to conserve limited hardware.
4. Convert Windows ARM64 every-four-hour jobs to daily jobs to conserve limited hardware.
5. Convert Windows ARM64 R2R baseline job from daily to weekly to conserve limited hardware.
6. For altjit jobs, only do push jobs for Checked builds; there's no need for more than that.
7. Add Windows and Ubuntu ARM R2R weekly jobs.
8. Add Windows ARM and ARM64, and Ubuntu ARM, weekly R2R JIT stress jobs.
Fixes #17572, #17241.
|
|
Eliminate `FEATURE_UNIX_AMD64_STRUCT_PASSING` and replace it with `UNIX_AMD64_ABI` when used alone. Both are currently defined; it is highly unlikely the latter will work alone; and it significantly clutters up the code, especially the JIT.
Also, fix the altjit support (now `UNIX_AMD64_ABI_ITF`) to *not* call `ClassifyEightBytes` if the struct is too large. Otherwise it asserts.
|
|
Fix minor issue with missing end quotes on command line
|
|
Tracking bug: https://github.com/dotnet/corefx/issues/29166
|
|
There is no need to store the dominator tree children in a hashtable, a vector is sufficient. Given the way the tree is constructed duplicates cannot be added.
|
|
|
|
(#17607)
|
|
|
|
|
|
* Fix dlbigleak
* cleanup dlstack
* Fix doublinknoleak
* Fix doublinkstay
* Fix dlcollect
* Fix doublinkgen
* Fix dlbigleakthd
|
|
* Adding failing string to Boolean and TimeSpan parse failure exceptions
* Avoid boxing char when using TryParse
* Removing unnecessary nameof expression
* Applying review feedback to remove ParseFailureKind from TimeSpanParse
* Added helper methods for TimeSpan parsing
|
|
|
|
Fix extra register-dependency on mem-form of vcvtsd/s2ss
|
|
|
|
master-20180417-0048, respectively (#17601)
|
|
[Arm64] Disable SIMD in crossgen (added as part of #14633)
|
|
|
|
- fix internal tests building
Signed-off-by: Petr Bred <bredpetr@gmail.com>
|
|
Fixes https://github.com/dotnet/coreclr/issues/17591
|
|
|
|
Use values that are more likely to match CPU cache sizes (e.g. L1 = 32K and L2 = 256K)
|
|
* Dict: Use Ref Local Reassignment (C# 7.3)
* Use lang version latest rather than 7.3
|
|
[Arm64] Add corefx_test_exclusions.txt
|
|
(#17593)
Signed-off-by: dotnet-bot-corefx-mirror <dotnet-bot@microsoft.com>
|
|
Clean up Ubuntu arm build machines
|
|
|
|
master-20180416-0039, respectively (#17580)
|
|
|
|
* [x86/Linux] Fix marshalling struct with 64-bit types
The System V ABI for i386 defines 4-byte alignment for 64-bit types.
* [Linux/x86] Fix marshalling tests in the case of System V i386 ABI
|
|
|
|
* Update KoreanLunisolarCalendar.cs
Corrects two issues with the conversion table for Gregorian to Korean lunisolar dates.
Issue 1: tables use Julian month/day instead of Gregorian month/day for entries prior to 1583.
Lmon and Lday have been converted from Julian month and day to Gregorian month and day for years prior to 1583.
This also required a new value for MIN_GREGORIAN_DAY (now 19) so the minimum date now supported by this class is 19 Feb 0918 Gregorian.
Issue 2: data discrepancy in years 1586, 1587, 1648, 1659, 1692, 1753 and 1754.
Table entries for 1586, 1587, 1648, 1659, 1692, 1753 and 1754 have been corrected to match Korea Astronomy and Space Science In-
stitute (KASI) data. Specifically:
•for years 1587, 1648, and 1754, the first day of year has been corrected
•for years 1586, 1587, 1659, 1692, 1753, and 1754, the lengths of each month have been corrected
Notes
A minor formatting change to the table has been made. The DaysPerMonth flag has been converted to use the binary literal to ensure that changes can be made and verified easily.
See: https://github.com/dotnet/coreclr/issues/17510 for full discussion of changes.
* Update KoreanLunisolarCalendar.cs
Added comment on sources of table data.
|
|
master-20180415-0051, respectively (#17540)
|