Greg Pfeil

And if there’s a mechanism for looking up the info for local packages, then might as well include the license information for the package the report is based on, not...

This has been a bit of annoyance lately because of my CI process (and I’m open to suggestions of how I can improve things on my end). I commit the...

Has this been fixed? Looking at the code, https://github.com/repology/repology-updater/blob/7c58e1dca34452109dc51f6f38ca06faddbd77b9/repology/parsers/parsers/hackage.py#L88-L93 seems to explicitly handle preferred versions.

This still gets pretty-printed with a line-wrap ``` unison t1 : Request really.long.path.prefix.that.causes.a.line.wrap.Thing a -> () t1 = cases { really.long.path.prefix.that.causes.a.line.wrap.Thing.thing longlong names -> reallyReallyReally } -> () { _...

Unison now allows `|||` as an operator, but `|` is reserved for sum types. `|` and `||` are currently identified as keywords in the lexer, which results in error messages...

So, what is the bug to fix here? 1. Should we be trying to identify this case and print it in a way that will a) parse and b) have...

I don’t think linked emptyAbility.md is illustrating this particular issue (if it is, it just hasn’t been sufficiently minimized). It’s easy enough to replicate this issue, but the more fundamental...

Or … is this round-tripping not an actual problem? I don’t think there’s any context in which `f : a ->{} b` can’t be replaced by `g : a ->...

Ah, this is because when we reorder stanzas to bring `use` and types to the top, we don’t introduce any virtual semis. So, if they were the last item the...

I tried to replicate this, but had to water it down a little (`range` isn’t something I could find). You can see in this transcript output that `use Decoder fail`...