[lit] Running lit with bazel fails for Python 3.11

[fzakaria]

I noticed that running llvm's lit is failing for Python 3.11 but passes for Python 3.10 -- I am not sure why but I wanted to capture my investigation.

Here is a script of me running lit with bazel on Python 3.10 successfully and then switching to Python 3.11 and failing.

❯ python -V
Python 3.11.5

❯ which python
/usr/local/google/home/fmzakari/code/github.com/openxla/stablehlo/venv3.11/bin/python

❯ bazel run @llvm-project//llvm:lit
INFO: Analyzed target @llvm-project//llvm:lit (0 packages loaded, 0 targets configured).
INFO: Found 1 target...
Target @llvm-project//llvm:lit up-to-date:
  bazel-bin/external/llvm-project/llvm/lit
INFO: Elapsed time: 0.137s, Critical Path: 0.01s
INFO: 1 process: 1 internal.
INFO: Build completed successfully, 1 total action
INFO: Running command line: bazel-bin/external/llvm-project/llvm/lit
Traceback (most recent call last):
  File "/usr/local/google/home/fmzakari/.cache/bazel/_bazel_fmzakari/17bce12c4b47a4a2fc75249afee05177/execroot/stablehlo/bazel-out/k8-fastbuild/bin/external/llvm-project/llvm/lit.runfiles/llvm-project/llvm/utils/lit/lit.py", line 3, in <module>
    from lit.main import main
ModuleNotFoundError: No module named 'lit'

❯ deactivate

❯ source venv310/bin/activate

❯ which python
/usr/local/google/home/fmzakari/code/github.com/openxla/stablehlo/venv310/bin/python

❯ python -V
Python 3.10.0

❯ bazel run @llvm-project//llvm:lit
INFO: Analyzed target @llvm-project//llvm:lit (1 packages loaded, 2 targets configured).
INFO: Found 1 target...
Target @llvm-project//llvm:lit up-to-date:
  bazel-bin/external/llvm-project/llvm/lit
INFO: Elapsed time: 0.186s, Critical Path: 0.01s
INFO: 2 processes: 2 internal.
INFO: Build completed successfully, 2 total actions
INFO: Running command line: bazel-bin/external/llvm-project/llvm/lit
usage: lit [-h] [--version] [-j N] [--config-prefix NAME] [-D NAME=VAL] [-q] [-s] [-v] [-vv] [-a]
           [-o PATH] [--no-progress-bar] [--show-excluded] [--show-skipped] [--show-unsupported]
           [--show-pass] [--show-flakypass] [--show-xfail] [--gtest-sharding]
           [--no-gtest-sharding] [--path PATH] [--vg] [--vg-leak] [--vg-arg ARG] [--time-tests]
           [--no-execute] [--xunit-xml-output XUNIT_XML_OUTPUT]
           [--resultdb-output RESULTDB_OUTPUT] [--time-trace-output TIME_TRACE_OUTPUT]
           [--timeout MAXINDIVIDUALTESTTIME] [--max-failures MAX_FAILURES] [--allow-empty-runs]
           [--per-test-coverage] [--ignore-fail] [--max-tests N] [--max-time N]
           [--order {lexical,random,smart}] [--shuffle] [-i] [--filter REGEX] [--filter-out REGEX]
           [--xfail LIST] [--xfail-not LIST] [--num-shards M] [--run-shard N] [--debug]
           [--show-suites] [--show-tests] [--show-used-features]
           TEST_PATH [TEST_PATH ...]
lit: error: the following arguments are required: TEST_PATH

We are using LLVM out of tree in https://github.com/openxla/stablehlo using llvm_configure starlark rule.
We are using a pretty recent commit as of this issue posting: b2cdf3c commit.

[fzakaria]

I noticed lots of repositories seem to pin to Python3.10 which makes this issue hidden.

[sjain-stanford]

+1 we're hitting this after updating one of our dependencies which transitively uses a newer rules_python (with python 3.11 interpreter).

[keith]

#87022