release subcommand should fail more gracefully (or it should succeed)

[audreyfeldroy]

Yesterday I tried to release my urlify package: https://github.com/audreyr/meteor-urlify

From my urlify/ directory, I entered:

$ mrt release .

I was prompted for my username/password. I entered in the right username but the wrong password (accidentally). My terminal froze up at the Password: prompt, and I couldn't CTRL+C out of it.

This morning I tried again, this time eagerly with the right password in hand :)

Instead of seeing the username/password prompt, I got this error:

urlify audreyr$ mrt release .
/usr/local/lib/node_modules/meteorite/lib/atmosphere.js:94
      throw "This tag has already been committed to the repo.";
      ^
This tag has already been committed to the repo.

[possibilities]

So obviously this isn't any good but you can get around it momentarily by using mrt publish .

I'm gonna open another ticket for hanging on wrong password.

I'll think over the bigger republishing issue and talk to Tom about it. In this particular area we generally ask WWBD (a.k.a What Would Bundler Do)? Off the top of my head I'd say the solution is release shouldn't complain if it's just recovering from a failed attempt. Since your tag is already in place it should be happy.

Thanks, more to come, let me know if using publish gets you there.

[possibilities]

Changed the name and added #21

[audreyfeldroy]

I'm trying mrt publish . now. It looks like it's hanging again on the password. I know I have the correct password this time because I was able to log in with it, and because I tried twice.

[tmeasday]

Oh, WWBD is that it would untag the branch if release failed (although I don't know how they 'unpush' it).

Perhaps release shouldn't complain if the tag is on the currently checked out revision. How hard would that be to do?

[possibilities]

I think I'll just chill it out a bit... no need to throw an error if the tag has already been committed. I'll do that and try to figure out what all this freezing/hanging.

[tmeasday]

@audreyr can you please take a look at my message in #21

[possibilities]

Still not failing more gracefully but failing less... closing.