Doug Beardsley

+1 for this. I was looking at using groundhog the other day and embedded sum types were exactly what I needed.

That works for `$XDG_CONFIG_HOME/git/ignore`, but it doesn't work for `~/.gitconfig`. I think that one is complicated enough that it requires built-in support.

Yeah, when I wrote [the fake package](https://github.com/mightybyte/fake) I also had a similar feeling that it complimented `armor` very nicely. I wanted `armor` to be agnostic to your choice of test...

Nothing concrete. I tried to make them both as self-contained and composable as I could. The [`testArmorMany`](https://github.com/mightybyte/armor/blob/master/src/Armor.hs#L115) function was created specifically to facilitate this kind of interoperability. I'm definitely willing...

I'd like to add another vote for this feature. I was just about to create an issue saying exactly the same thing. At the moment this seems like the most...

The problem here was that the single quote after `read-keyset` was a "smart quote" that got auto-substituted in place by the editor. There are certainly other places in the chain...

Hmmm, this might also be related to the problems with code like this: https://gist.github.com/buckie/2ff437310d4112132340640c6f4505dd

I'd like to add my vote for this as well. I have an option parser that includes this: argument (maybeReader $ Just . T.pack) ( metavar "foo" completeWith (getFooCompletions cfg)...

@w3rs I would suggest using a single `many` argument and then simply take the last item from the resulting list and strip off the last item of the list and...

I believe that is expected because the transaction went in and gas got charged. Just the transfer failed.