- User Since
- Oct 11 2016, 5:32 AM (153 w, 2 d)
Wed, Sep 18
This change appears to cause build bot failures under msan:
FAIL: LLVM :: DebugInfo/AArch64/asan-stack-vars.mir (19531 of 33579) ******************** TEST 'LLVM :: DebugInfo/AArch64/asan-stack-vars.mir' FAILED ******************** Script: -- : 'RUN: at line 1'; /b/sanitizer-x86_64-linux-fast/build/llvm_build_msan/bin/llc -O0 -start-before=livedebugvalues -filetype=obj -o - /b/sanitizer-x86_64-linux-fast/build/llvm-project/llvm/test/DebugInfo/AArch64/asan-stack-vars.mir | /b/sanitizer-x86_64-linux-fast/build/llvm_build_msan/bin/llvm-dwarfdump -v - | /b/sanitizer-x86_64-linux-fast/build/llvm_build_msan/bin/FileCheck /b/sanitizer-x86_64-linux-fast/build/llvm-project/llvm/test/DebugInfo/AArch64/asan-stack-vars.mir -- Exit Code: 2
Hey! Thank you very much for your investigations!
Tue, Sep 17
@samparker : right after r372126 there still seems to be a test failure remaining:
This caused some ASAN tests to start breaking:
Adding @djasper as this is really a re-implementation of https://reviews.llvm.org/D6833.
Agree with @MyDeveloperDay that in general we should aim to document all the known cases where a style doesn't work (with FIXMEs etc.), but not submit failing tests. We can have tests that demonstrate the current (bad) behavior together with a FIXME comment about what is the expected good outcome instead.
Fri, Sep 13
This was for experimental purposes, abandoning.
This looks good with the FIXME.
Tue, Sep 10
I'll need some more time to investigate the implications of this with respect to Objective-C disambiguation stuff.
Specifically, this may interact with funny ways with the heuristic outlined in (old) UnwrappedLineParser.cpp line 1453:
// In a C++ lambda a template type can only occur after an arrow. We use // this as an heuristic to distinguish between Objective-C expressions // followed by an `a->b` expression, such as: // ([obj func:arg] + a->b)
Mon, Aug 26
Fri, Aug 23
Aug 8 2019
- Apply review comments
Jul 25 2019
Jul 23 2019
May 10 2019
Sorry for the delay. I'll submit this for you.
May 7 2019
Apr 18 2019
Apr 12 2019
I'll merge this.
Apr 11 2019
Apr 8 2019
Apr 5 2019
Generally I'm against introducing new style flags or flag options unless a supported style requires it.
IMO this increases maintenance burden and can quickly lead to a big space of style flags that contain tricky incompatibilities.
I believe there is no such thing as an "short else statement". The else is part of the if statement and if it is present, I don't consider the whole if statement short. As such, IMO the bug is invalid.
Apr 4 2019
- Add a note about intent
Apr 3 2019
- Address review comments