fsharp_space_before_uppercase_invocation=true breaks method calls

[lydell]

Issue created from fantomas-online

Code

Log.Logger <-
    LoggerConfiguration()
        .Destructure.FSharpTypes()
        .WriteTo.Console()
        .CreateLogger()

Result

Log.Logger <-
    LoggerConfiguration ()
        .Destructure.FSharpTypes ()
        .WriteTo.Console ()
        .CreateLogger ()

Problem description

The output code causes the compilation errors:

error FS0039: The type 'Unit' does not define the field, constructor or member 'Destructure'.
error FS0501: The object constructor 'LoggerConfiguration' takes 0 argument(s) but is here given 1. The required signature is 'LoggerConfiguration() : LoggerConfiguration'.

Extra information

• The formatted result breaks by code.
• The formatted result gives compiler warnings.
• I or my company would be willing to help fix this.

This seems to have been introduced in 4.4.0-beta-003.

Options

Fantomas Master at 02/05/2021 16:56:34 - 0cbe3df

    { config with
                SpaceBeforeUppercaseInvocation = true }

[knocte]

Wow, shouldn't this be an F#-compiler bug?

[auduchinok]

Wow, shouldn't this be an F#-compiler bug?

@knocte Nope, it's per-spec. The application without space inside has a higher precedence than a normal one, otherwise it'd be very hard to come up with rules for applications and access-via-dot syntax, and cases like the following ones would be messed up:

f(x, y).P
f (x, y).P

[knocte]

cases like the following ones would be messed up:

hey thanks for the info, but what do you mean exactly by that example? how is f (x,y).P different than f(x,y).P? genuinely curious

[auduchinok]

how is f (x,y).P different than f(x,y).P?

Here the first example applies (x, y).P to f, while the second one accesses P member on the result of f(x, y).

[knocte]

Truly interesting, I guess I hadn't come up with this because the fluent style is less visible in functional languages (because returning the same element in each call attracts a programming style that involves mutability). And what would f (x,y) .P be?

[auduchinok]

Well, the tuple example is not very correct, and my point needed a syntax example where parens were required. 🙂
Imagine accessing member on a function application result like this:

f(g x).P
f (g x).P

[auduchinok]

And what would f (x,y) .P be?

f (x,y) .P is the same as f (x,y).P, since dot access has higher precedence than "normal" (i.e. with space) application.

The operators and applications part of this table below can give more info: https://docs.microsoft.com/en-us/dotnet/fsharp/language-reference/symbol-and-operator-reference/#operator-precedence

[knocte]

Ok thanks!