- User Since
- Apr 11 2019, 1:09 PM (76 w, 14 h)
Reduce wait time to 5ms. (not really meaningful - we just use this as a way to prevent unrolling)
Wed, Sep 23
reduce wait time
Mon, Sep 21
Updated a few tests to work on Android.
Wed, Sep 16
Tue, Sep 15
Removed the check-support flag and always performs the check.
Tue, Sep 1
updated diff: skipped lsan/asan tests on pre-S images
Thu, Aug 27
Avoid accessing TLS variables for android when the thread-properties API is not present.
Wed, Aug 26
Use a more portable way to prevent loop unrolling
Aug 21 2020
[WIP] mov stuff to target-specific dirs
Aug 19 2020
Aug 18 2020
Aug 14 2020
Aug 13 2020
Dont run leakcheck for android if the thread-properties API is not present
(With LGTM from sbenza)
Aug 12 2020
- Remove the checking of non-power-of-2 sizes from is_lock_free for now (because the atomic_* ops don't handle them yet)
- Rename function to _c and add pragma redefine
Aug 5 2020
Fixed clangtidy warnings
Aug 4 2020
Removed use of GNU case range extension
(Triggered a bunch of clang-tidys)
Aug 3 2020
Jul 27 2020
Jul 24 2020
Would anyone mind Accepting this so I could clear this revision off my pending queue? Thanks!
(It's already committed so ... )
(Closing this ... assuming issue has been resolved)
'cycle' -> 'cycles'
Jul 23 2020
CheckStructHasMember didn't work (because cycles is a bit field) so updated it to just compile a small program that accesses .cycles.
Jul 22 2020
Check for existence of cycles before enabling LBR mode.
Use Cmake's CheckStructHasMember to check for cycles field.
Jul 21 2020
My point is that someone compiling for ARM will define _LIBCPP_ABI_ENABLE_SHARED_PTR_TRIVIAL_ABI thinking they are getting trivial weak_ptr and shared_ptr passed in registers, but they won't (and silently so). That deception is what I think should be flagged explicitly through an error of some sort. Do you agree?
Jul 20 2020
PP.S: weak_ptr test failed on ARM32. Fixed in https://reviews.llvm.org/D84200
P.S: This broke GCC build bot. Fixed forward in https://reviews.llvm.org/D84183
Jul 17 2020
Jul 16 2020
Jul 15 2020
Thanks, everyone, for your input (and patience)!
I'm planning on committing this at the beginning of next week. If anyone has some concern or additional testing they'd like to see happen, please let me know.
(Have run the check-cxx with -DLIBCXX_ABI_UNSTABLE=true)
Jul 14 2020
Added UNSUPPORTED comments on tests.
We'll need to add // UNSUPPORTED: clang-4, clang-5, clang-6, clang-7, clang-8, clang-9, clang-10 on the tests with a short comment explaining there was a segfault.
Jul 10 2020
Added unit test and the missing serialisation code
Jul 9 2020
Fixed warning [-Wmissing-field-initializers] in code
Removed empty line
Jul 8 2020
Return error for empty function
Jul 7 2020
one more test
prevent diagnostics for dependent types
Jul 6 2020
Add more tests
Dont skip checking even for templated type