re: <>: ...
# pex
re: i feel pretty strongly that SOURCE_DATE_EPOCH is a misfeature when compared to them having specified a particular date that tools should use
IMO, leave a NB near the deterministic datetime explaining why we don't use it.
...oh. hm.
or we could have pants guarantee to specify it in all cases where we invoke pex? not loving that, as it seems unnecessarily fragile
Okay. I am okay with removing it if we are willing to revisit the decision if we get feedback from users that they care about this feature. Otherwise, simpler works for me. It makes our docs simpler, can make the function a constant, and simpler unit test
ok, thanks.
I would like to be a good Linux citizen, although it is also worth noting that the intent of the reproducible builds website is to make tools reproducible for security, that you know the binary was not tampered with. That is not really what we’re after, and you’re right it’s probably a misfeature for that reason
Buck and Bazel don’t seem to respect that env var
i expect for exactly this reason
you would then not be able to reproduce a build between osx/windows and linux without additional effort
the only benefit i see to the env var is that if you have two different tools that might produce the same file, you could then convince them to produce it with the same timestamp
👍 1
but... that is very, very tenuous