(Empty/NoEmpty)-bodied named computation expression produces inconsistent formatting

[edgarfgp]

Issue created from fantomas-online

When formatting Empty-bodied named computation expression causes:

error FS0003: This value is not a function and cannot be applied.

Code

A() {} // Empty-bodied

B() { 0 }

Result

A () { } // Empty-bodied

B() { 0 }

Problem description

Please describe here the Fantomas problem you encountered.
Check out our Contribution Guidelines.

Extra information

• The formatted result breaks my code.
• The formatted result gives compiler warnings.
• I or my company would be willing to help fix this.
• I would like a release if this problem is solved.

Options

Fantomas main branch at 2024-10-29T08:28:00Z - b7dca79

Default Fantomas configuration

_{Did you know that you can ignore files when formatting by using a .fantomasignore file?
PS: It's unlikely that someone else will solve your specific issue, as it's something that you have a personal stake in.}

[nojaf]

Thank for reporting this, it is somewhat interesting that we are not making the correct Oak node I see. I'll take a look.

[nojaf]

The wrong AST is produced in this case:

                    SynModuleDecl.Expr(
                        expr =
                            SynExpr.App(
                                flag = ExprAtomicFlag.NonAtomic,
                                isInfix = false,
                                funcExpr =
                                    SynExpr.App(
                                        flag = ExprAtomicFlag.Atomic,
                                        isInfix = false,
                                        funcExpr = SynExpr.Ident(Ident("A", R("(1,0--1,1)"))),
                                        argExpr = SynExpr.Const(constant = SynConst.Unit, range = R("(1,1--1,3)")),
                                        range = R("(1,0--1,3)")
                                    ),
                                argExpr = SynExpr.Record(baseInfo = None, copyInfo = None, recordFields = [], range = R("(1,4--1,6)")),
                                range = R("(1,0--1,6)")
                            ),
                        range = R("(1,0--1,6)")
                    )

The SynExpr.ComputationExpr is not there, so you'd need to check is this is a parser bug or by design?

[edgarfgp]

Thanks Florian. Will look into it.

[edgarfgp]

@nojaf Looks like this is by design unfortunately. But in F# 9 the compiler re-writes the Untyped AST from SynExpr.Record to SynExpr.ComputationExpr adding an a dummySynExpr.Const (SynConst.Unit, range0))forA() {}`

See: https://github.com/edgarfgp/fsharp/blame/935b796dc841b6346f655421bb791c1764ab1570/src/Compiler/Checking/Expressions/CheckExpressions.fs#L8546

Do you think Fantomas can also do the same ?

[nojaf]

I see, yeah, we could cook something similar in Fantomas.

[nojaf]

Available in https://www.nuget.org/packages/fantomas/7.0.0-alpha-004