Having added a more robust test for "reproducibility" of output in #6, it has become apparent that zip archives generated under Windows are not binary-identical to those generated under Linux, even when "reproducible" mode is enabled.
At the moment, the test generates an xfail under Windows.
I haven't taken the time to dig deep enough to figure out what the difference is. Maybe the zip mode bits?
Having added a more robust test for "reproducibility" of output in #6, it has become apparent that zip archives generated under Windows are not binary-identical to those generated under Linux, even when "reproducible" mode is enabled.
At the moment, the test generates an xfail under Windows.
I haven't taken the time to dig deep enough to figure out what the difference is. Maybe the zip mode bits?