Robert J. Simmons

Elm fragments get turned into Elm modules named Main

2

If `blah.txt` contains the following contents: ``` whatever = 6 + 12 ``` `elm-format` (version 0.8.5) will change the file to one that has different semantics: ``` % ./node_modules/.bin/elm-format blah.txt...

Running out of memory on a small input?

5

I don't know what could be going on with this seemingly innocent grammar: ``` @{% const mooLexer = require('moo').compile({ space: { match:/[ \t\n]+/, lineBreaks: true }, ident: { match: /[a-z][a-zA-Z0-9_]*/,...

Either one of two (backwards non-compatible) changes would make things easier for implementers: * Have the default `Tracer.startSpan` implementation call `Span.addTags` after calling the `Span` constructor. * Specify that addTags...

Gollum will turn `[[Page#wiki-anchor]]` into a link to `{base-path}Page#wiki-anchor`, but gollum-site appears to turn `[[Page#wiki-anchor]]` into a link to `{base-path}Page.html`, throwing away the #wiki-anchor portion.

Cidre MLB should support 'nonexhaustiveMatch ignore' annotations

2

I already annotate code that I expect to fail in libraries, it would be nice if CIDRE would suppress these errors where requested.

Map won't let me give an instantiation of the type it has

4

Truly a "Expected: ?.vector, got: ?.vector" esque error, courtesy of https://github.com/robsimmons/l10/blob/bug6/sml/splitting.sml#L67-70 (which does normal-sml-typecheck). If you can't tell, from the annotation, I played around with this one for awhile, to...

Spurious coverage error in normal ML checking (cmlib v0.4.0)

1

Ironically, I think it's in your code (pqueue-lazy-paring.sml). ``` - Cidre.make "/Users/rjsimmon/.smackage/lib/cmlib/v0.4.0/cmlib.mlb"; ``` gives me two warnings - one of them is where I'd added a nonexhaustive match ignore annotation...

No error for bad sorts in tuple patterns

1

``` (*[ datasort true = true ]*) fun f x = let (*[ val q: true ]*) val (q, _) = (false, 3) in q end ``` Rowan says the...

Performance: larger MDX files are unmanagably slow to parse

6

Bad news @wooorm - micromark/micromark#169, or something like it, is an issue here as well - I haven't yet grokked your edit maps, but either they don't solve the quadratic-complexity...

SEQ_CONCAT (n-ary) `concat [[a], [b], c] [d] [e, f] is [[a], [b], c, d, e, f]` Can be run backwards as long as there's only one free variable (prefix/postfix matching)...