Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Test suite failures on aarch64-darwin #43614

Closed
samuela opened this issue Dec 31, 2021 · 7 comments
Closed

Test suite failures on aarch64-darwin #43614

samuela opened this issue Dec 31, 2021 · 7 comments

Comments

@samuela
Copy link
Contributor

samuela commented Dec 31, 2021

After downloading the official dmg from the website, I was surprised to find that the test suite does not pass on aarch64-darwin (aka M1 aka apple silicon aka macOS on ARM etc). I ran bin/julia --check-bounds=yes --startup-file=no --depwarn=error share/julia/test/runtests.jl from the Contents/Resources/julia directory. To summarize I'm seeing the following failures:

  • segfault in julia/test/strings/basic.jl:228
  • segfault in julia/test/iterators.jl:358
  • segfault in julia/test/mpfr.jl:931
  • test failures in julia/test/errorshow.jl
Test Failed at /Users/samuelainsworth/Downloads/Julia-1.7.app/Contents/Resources/julia/share/julia/test/errorshow.jl:781
  Expression: occursin(r"repeats \d+ times", bt_str)
   Evaluated: occursin(r"repeats \d+ times", "\nStacktrace:\n [1] single_repeater()\n   @ Main.Test14Main_errorshow ~/Downloads/Julia-1.7.app/Contents/Resources/julia/share/julia/test/errorshow.jl:771")
Test Failed at /Users/samuelainsworth/Downloads/Julia-1.7.app/Contents/Resources/julia/share/julia/test/errorshow.jl:789
  Expression: occursin(r"the last 2 lines are repeated \d+ more times", bt_str)
   Evaluated: occursin(r"the last 2 lines are repeated \d+ more times", "\nStacktrace:\n [1] pair_repeater_a()\n   @ Main.Test14Main_errorshow ~/Downloads/Julia-1.7.app/Contents/Resources/julia/share/julia/test/errorshow.jl:772")
  • test failure in julia/test/stacktraces.jl:51
Test Failed at /Users/samuelainsworth/Downloads/Julia-1.7.app/Contents/Resources/julia/share/julia/test/stacktraces.jl:51
  Expression: StackTraces.lookup(C_NULL) == [StackTraces.UNKNOWN] == StackTraces.lookup(C_NULL + 1) == StackTraces.lookup(C_NULL - 1)
   Evaluated: Base.StackTraces.StackFrame[ip:0x0] == Base.StackTraces.StackFrame[ip:0x0] == Base.StackTraces.StackFrame[ip:0x1] == Base.StackTraces.StackFrame[_dyld_private at julia:?]
  • segfault in julia/stdlib/v1.7/SparseArrays/test/sparse.jl:166
  • segfault in julia/stdlib/v1.7/Printf/test/runtests.jl:5
  • segfault in julia/stdlib/v1.7/SuiteSparse/test/cholmod.jl:318
  • error during test in julia/test/testdefs.jl:21
Error During Test at /Users/samuelainsworth/Downloads/Julia-1.7.app/Contents/Resources/julia/share/julia/test/testdefs.jl:21
  Got exception outside of a @test
  LoadError: On worker 21:
  SystemError: shm_open() failed for /jl008570nxTwrg2kyN0MULbsX4Ue: Permission denied
  • Test runner seems to hang after providing the test summary for InvasiveLinkedList followed by a KILLING BY QUICK KILL WATCHDOG warning/message

You can view the full logs here: https://gist.github.com/samuela/b24439841ecd52cc201dd1a60d3b1d6f. The whole test suite ran in about 20 minutes for me on the 10-core M1 Max.

Related: #36617

@giordano
Copy link
Contributor

giordano commented Dec 31, 2021

I was surprised to find that the test suite does not pass on aarch64-darwin

You shouldn't be surprised: this platform is in tier-3 support, which by definition means

Julia may or may not build. If it does, it is unlikely to pass tests.

There are already plenty of tickets reporting issues for apple silicon, including #43321 which is a similar ticket reporting a grab bag of problems. Some of them had been already addressed before you opened this issue. For example, shm_open is already fixed by #43516, Base.StackTraces.lookup(C_NULL - 1) is addressed by #43612

@samuela
Copy link
Contributor Author

samuela commented Dec 31, 2021

Understood. How should I go about reporting these failures then? Should I create separate GH issues for each one (skipping those that have already been reported ofc)?

@giordano
Copy link
Contributor

Is there anything different than #43321? This looks a duplicate of that ticket, no?

@samuela
Copy link
Contributor Author

samuela commented Dec 31, 2021

I don't see any of the segfaults reported in #43321, but perhaps I'm missing something?

@giordano
Copy link
Contributor

Many segfaults on the M1 are related to #41440

@samuela
Copy link
Contributor Author

samuela commented Dec 31, 2021

Ah, understood. I'll try rerunning things on latest master. Out of curiosity, is there an ETA for getting aarch64-darwin support to tier 1 or 2?

@giordano
Copy link
Contributor

No

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants