emscripten-core / emscripten-core/emscripten
test_futimens is flaky on ARM64 macOS.
- Dominant language
- C++
- Stars
- 27.6k
- Forks
- 3.6k
- Avg merge
- 1d 1h
- Merged PRs (30d)
- 105
Description
on Apple Silicon macOS M1, the core test `test_futimens` (in both parameters `test_futimens_nodefs` and `test_futimens_rawfs`) fails occassionally, e.g.
[thinlto0.test_futimens_rawfs](http://clbri.com:8010/api/v2/logs/63940/raw_inline)
[llvmlibc.test_futimens_rawfs](http://clbri.com:8010/api/v2/logs/47305/raw_inline)
[corez.test_futimens_rawfs](http://clbri.com:8010/api/v2/logs/32155/raw_inline)
[thinlto1.test_futimens_nodefs](http://clbri.com:8010/api/v2/logs/29544/raw_inline)
[ltoz.test_futimens_nodefs](http://clbri.com:8010/api/v2/logs/29276/raw_inline)
[core2ss.test_futimens_nodefs](http://clbri.com:8010/api/v2/logs/27612/raw_inline)
[ltos.test_futimens_rawfs](http://clbri.com:8010/api/v2/logs/25910/raw_inline)
with a timestamp difference.
Reading the code, I am not quite able to find what could cause such a timestamp divergence. There does not seem to be room for such behavior as far as I could tell.
The test is not flaky on any other bots that I run (Windows or Linux). Specifically, it is not flaky on my macOS x64 bot either, only on macOS M1 bot.
Contributor guide
Assessment
This issue has not been assessed yet.