Closed danielbayley closed 8 years ago
I don't really see a valid use case for this. What's wrong with pushing a patch release? If you made a big mistake, unpublish and republish? If something failed during a publish, you can try republishing the tag via apm publish --tag <tag>
.
You're claiming that bumping version numbers and pushing tags is something that is error prone, how would this not suffer from the same problems? How would this proposed --amend
flag work?
Thanks in advance!
Ah, I wasn't aware that apm publish --tag <tag>
would basically do this (i.e attempt the publish again without an additional version bump or adding another commit…) I guess --amend
could just be an alias for apm publish --tag <latest tag>
in that case… or not?
That's what patch releases are for. You shouldn't be modifying already-published releases of any software because that undermines the idea of having a known stable configuration.
Fair enough, I get that.
@thedaniel I only really suggested this after borking the test
and language-jxa
packages though (deleted the GitHub repo before apm unpublish
), which I would really like deleted from the database if possible (see #221).
Something like
apm publish --amend
might be useful for really small fixes perhaps? Especially since bumping the version number & pushing tags is so error prone at the moment (see #315 for example…)