Crystal: Can't format `NamedTuple(foo: Int32 bar: Bool)`

[bew]

The formatter can't format the following code: (@MakeNowJust :smiley:)

p NamedTuple(foo: Int32 bar: Bool).types

(there is no comma and it compiles! https://carc.in/#/r/3wvz)

---

NOTE: This syntax only works for the generic type, not in actual named tuples.
I think the syntax can be confusing, I'd suggest to remove it (by making the comma mandatory).

[MakeNowJust]

I think it should not be compiled and this is a simple parser bug.
-- 土曜日, 21 4月 2018, 00:29午後 +09:00 from Benoit de Chezelles [email protected] :

The formatter can't format the following code: ( @MakeNowJust 😃)
p NamedTuple(foo: Int32 bar: Bool).types

(there is no comma and it compiles! https://carc.in/#/r/3wvz )

NOTE: This syntax only works for the generic type, not in actual named tuples.
I think the syntax can be quite confusing, I'd suggest to remove it (by making the comma mandatory).
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub , or mute the thread .

[bew]

I don't think that's a parser bug, just a feature that's never used and that looks weird to me with the rest of the language.

[ghost]

Just to add, this bug actually goes a bit deeper. Because when you compile you will not see any error, but when you are using Sublime for Crystal (or a console that runs crystal tool format on save), the formatter will echo out an Error: expecting ,, not IDENT, and ask the user to run a stack-trace. This will cause a small delay when saving files on WSL. Once the , was added, the delay went away, and the console flash is now instant.

[MakeNowJust]

"parser bug" means and I guess the author of NamedTuple parser does not intend to accept such a syntax. Is there a spec for this?

土曜日, 21 4月 2018, 00:40午後 +09:00 from Benoit de Chezelles [email protected] :

I don't think that's a parser bug, just a feature that's never used and that looks weird to me with the rest of the language.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub , or mute the thread .

[bew]

Good point, I can't find a spec for that