Revise F# formatting guide #25972
Revise F# formatting guide #25972
Conversation
|
@nojaf Could you review? This addresses https://github.com/dotnet/docs/issues/25917#issuecomment-913516706 but also rearranges the sections in the guide to follow a logical order. Please check these in particular
|
|
This was a bit hard to review, so I may have missed some things. As for the blank lines above comments, if the original code already had them Fantomas will do the right thing. |
|
@nojaf I'm going through the entire revised doc using the Fantomas preview online tools to check the formatting. Could you check this one? I was expecting Target.create "Build" (fun ctx ->
// code
// here
()) |
|
@nojaf Also this one, which formats the |
|
@nojaf There's also a discrepency with respect to named arguments - the style guide says "no space" |
|
@nojaf Here's another case of interest - records with XML doc comments on the fields. Looks a bit odd to place that brace on the last line |
|
@nojaf If you could check the above cases quickly that would be cool - if any of them are Fantomas bugs I can open issues. I think maybe some are the changes we've discussed either not yet implemented or not in preview tools |
|
Hey @dsyme, you are correct some stuff is not in the online tool yet (nor in a stable version). Target.create "Build" (fun ctx ->
// code
// here
())is still open in fsprojects/fantomas#1858. let function1 arg1 arg2 arg3 arg4 = arg1 + arg2 + arg3 + arg4 + arg2 + arg3 + arg4 + arg2 + arg3 + arg4Fantomas currently treats almost all infix operators the same. There is nothing in place for arithmetic operators and in general infix operators are quite challenging to get right. It is a bit on an ongoing dragon in the codebase. I'm open to some good heuristics there. let rainbow2 =
{ rainbow with
Boss = "Jeffrey"
Lackeys = [ "Zippy"; "George"; "Bungle" ] }Created fsprojects/fantomas#1876 let makeStreamReader x = new System.IO.StreamReader(path=x)we just format the shape of the argument AST here as is. Paren
(App
(NonAtomic, false,
App
(NonAtomic, true, Ident op_Equality, Ident path,
tmp.fsx (1,52--1,57) IsSynthetic=false), Ident x,
tmp.fsx (1,52--1,58) IsSynthetic=false),
tmp.fsx (1,51--1,52) IsSynthetic=false,
Some tmp.fsx (1,58--1,59) IsSynthetic=false,
tmp.fsx (1,51--1,59) IsSynthetic=false),
tmp.fsx (1,25--1,59) IsSynthetic=false),we should do something different when it is in the context of SynExpr.New. // Declaring additional members on PostalAddress
type PostalAddress =
{
/// The address
Address: string
/// The city
City: string
/// The zip code
Zip: string
}to preserve the original style here, you are looking for the fsharp_multiline_block_brackets_on_same_column setting perhaps? |
|
@nojaf Regarding records declarations (with/without XML doc comments):
This is tricky. Here's my thinking
Anyway, on summary on revising the guide I kind of feel that record definitions should get special treatment to support a nice look when XML doc commented. But I'm aware that this is a special case. Would be interested in your thoughts. |
|
Hmm, a bit tricky indeed. Although as I understand it there is a clear rule when to always use the wide record formatting:
This is a relatively easy thing to check in the syntax tree. Slightly related I've raised fsprojects/fantomas#1878. |
|
@nojaf Added fsprojects/fantomas#1879 where we can discuss |
|
I've updated the tool to use our |
|
Fantastic! @luisquintanilla @KathleenDollard could I have a review/approve when your ready? Thanks. |
|
@luisquintanilla This is ready, thank you!!
Thanks, all done
Updated |
|
@luisquintanilla Thanks! |
Summary
///documentation commentsPreview Link