New linter is going crazy on perfectly fine elmish code

[forki]

Originally reported to ionide. Please see https://github.com/ionide/ionide-vscode-fsharp/issues/997

[jgardella]

Hi @forki.

This rule is also based on the F# style guide, which suggests always using parentheses around tuple instantiations. As @MangelMaxime suggested in the linked issue, it can be disabled through the config.

However if I'm understanding your comments on the linked issue correctly, it sounds like this is a common format in elmish/SAFE stack code? I was not aware of this. Given that this is in the official style guide, I think it makes sense to have it as a default. If elmish/SAFE code differs from the style guide in this instance or others, maybe we could create a config file specifically for elmish/SAFE projects?

[forki]

Actually this is so established right now, that I think the "official style guide" needs to be changed. Adding parens here breaks basically all the historical code in light syntax and adds a lot of clutter. /cc @cartermp

[forki]

I opened up a pr on the styleguide. https://github.com/dotnet/docs/pull/10186

I don't want to be a dick here and will happily accept community decision.ut this one feels so completely anti-F# for someone like me who uses it like first inofficial code drops.

[jgardella]

I think it's a great discussion to have, thanks for starting it. Personally, I like parentheses around tuple instantiation (and even around tuple pattern matching) as I think it makes the code easier to read. But again I have not used elmish/SAFE so I may not be seeing the full picture. Let's continue the discussion over on your style guide PR.

[MangelMaxime]

Considering the changes requested by @forki on the guidelines. If this is accepted, we could add a mode "strict" to FSharpLint for people who want to always put parens around their tuples.

The idea being because it's mentioned as optional people could want to force them self to always use parens even for function return.

[forki]

Why would anyone ever do that? It's a light language mode - next thing we do is require people to put semicolons at the end of each line ;-)

Am Mo., 28. Jan. 2019, 22:16 hat Maxime Mangel notifications@github.com geschrieben:

Considering the changes requested by @forki https://github.com/forki on the guidelines. If this is accepted, we could add a mode "strict" to FSharpLint for people who want to always put parens around their tuples.

The idea being because it's mentioned as optional people could want to force them self to always use parens even for function return.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/fsprojects/FSharpLint/issues/296#issuecomment-458304387, or mute the thread https://github.com/notifications/unsubscribe-auth/AADgNAt1jpnCGTyyg4TswbYcdG5V_LYvks5vH2iWgaJpZM4aWYpI .

[MangelMaxime]

No idea, but just because we don't know doesn't me but can't provide an option to support it if some want to and because the code is already written for it. ^^

I believe in configuration against strict walls. Even, we good defaults a tool doesn't all suits the need for everyone.

[forki]

Yes config is good, but if we manage to settle on good default it's even better. I'm willing to do compromises.

Am Mo., 28. Jan. 2019, 22:53 hat Maxime Mangel notifications@github.com geschrieben:

No idea, but just because we don't know doesn't me but can't provide an option to support it if some want to and because the code is already written for it. ^^

I believe in configuration against strict walls. Even, we good defaults a tool doesn't all suits the need for everyone.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/fsprojects/FSharpLint/issues/296#issuecomment-458316623, or mute the thread https://github.com/notifications/unsubscribe-auth/AADgNGhIny0y-B9Hh_psoRmyS-Yv6bThks5vH3FogaJpZM4aWYpI .

[cartermp]

I think having something like a requirement for all tuples to be surrounded by parentheses could make a good setting, but the default should be to allow them at the return site like the docs suggest in light of elmish styles.

[forki]

Cool, so the suggestion to open up the restriction is merged to the styleguide.

[jgardella]

Great, happy we were able to reach a decision and update the guide. I'm thinking for FSharpLint we can make this setting configurable with the following options:

• Instantiation - require parentheses for all instantiations
• InstantiationExceptReturn - require parentheses for all instantiations, except when they are return value (elmish style)

[MangelMaxime]

For me having both of these options make sense indeed.