Basically, Repository.resume_write_group() internally _starts_write_group() and then resumes the given tokens.
However, tokens can fail to be resumed, which leaves things in an intermediate state. Namely, the _start_write_group finished, but wasn't aborted. Which leaves open the NewPack.
I don't know if this ever matters in practice, but at least on Windows it causes
def test_cannot_resume_aborted_write_group(self):
To give cleanup stipple because we can't cleanup the working dir because of the open file handle.
This gives a cleaner test suite run on Windows.
Basically, Repository. resume_ write_group( ) internally _starts_ write_group( ) and then resumes the given tokens.
However, tokens can fail to be resumed, which leaves things in an intermediate state. Namely, the _start_write_group finished, but wasn't aborted. Which leaves open the NewPack.
I don't know if this ever matters in practice, but at least on Windows it causes resume_ aborted_ write_group( self):
def test_cannot_
To give cleanup stipple because we can't cleanup the working dir because of the open file handle.