Closed asummers closed 6 years ago
Agreed, I was just waiting to see if that last one was fixed in master.
Makes sense 😄
Still throughly confused why my PR fixed anything, I'll have to investigate that at some point.
If anything, I should have just gotten a better error message, but that last guy's issue was legitimately fixed in #192... Unsure how that's possible. Is it possible the original hex push wasn't pulled master?
I don't see how that is possible, because the version in mix.exs
is used by hex to publish it. So it wouldn't show up in hex.pm with that version number unless it were based on those sources.
I was also surprised it fixed anything, I wonder is it swallowing a warning now for them?
Push of rc.1 is literally blocked on this issue: https://github.com/hexpm/hexpm/issues/690
I buy the mix.exs argument. Just trying to think outside the box for that.
I'd share your suspicion of the error swallow except it seemed like the same error was printing the actual message after that point. I'm very confused.
Hex errors are no bueno :(
Additionally the rescue even being necessary is completely confusing because the lexing error was being explicitly thrown/caught.
I thought the lexing error was thrown by Erlang code? That would always be an exception that must be rescued.
https://github.com/jeremyjh/dialyxir/blob/master/lib/dialyxir/pretty_print.ex
lex/1
is already doing that though. The tuple was that one being thrown.
Then that was being explicitly caught in format_warning in
https://github.com/jeremyjh/dialyxir/blob/master/lib/dialyxir/formatter.ex
Ah, I see. Yeah I don't understand either then.
rc.1 is published
I know what the issue was; we didn't add the new files in src to the file list for Hex. So its only working when pulled via git. Will push a new RC shortly.
GOOD CALL!!!
@jeremyjh mind bumping? Should eliminate a bunch of the issues people are opening after the merge this morning.