Codecov Report

Merging #4984 (c1f2a85) into master (b5b2b11) will decrease coverage by 3.42%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #4984      +/-   ##
==========================================
- Coverage   76.91%   73.49%   -3.42%     
==========================================
  Files         606      606              
  Lines       53770    53770              
==========================================
- Hits        41355    39518    -1837     
- Misses      12415    14252    +1837     

Flags with carried forward coverage won't be shown. Click here to find out more.

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

Would you happen to have any idea about this error?

https://github.com/espnet/espnet/actions/runs/4340074973/jobs/7578272513#step:18:2537

This error may happen only for enh tse task.

@ShigekiKarita

hmm no idea. Perhaps, multiple processes can write .coverage at the same time?

Thanks, Shigeki. Could you tell me how we generate .coverage in CI? I don't understand the setting basically.

Is the following command generating the .converage, right?

I tested running utils/run.pl JOB=1:10 JOB.log coverage run --append empty.py.

I couldn't reproduce Couldn't use data file '.../.coverage': no such table: meta, but it is sometimes successful, and sometimes is failed with sqlite3.OperationalError: no such table: coverage_schema or Couldn't use data file '.../.coverage': no such table: coverage_schema
Potentially, parallel running for coverage run --append causes the problems.

I'm not sure these errors come from the same reason, but I found that once .coverage was generated successfully, I didn't get the error after it.

I'll try to modify the script to generate .coverage at the first in a single process.