Thinking about the transformation chaining API for BasicTextField2. Which one of these do you prefer...

# compose

[Zach Klippenstein (he/him) [MOD]]

Thinking about the transformation chaining API for BasicTextField2. Which one of these do you prefer (see thread)

[Zach Klippenstein (he/him) [MOD]]

1:

inputTransformation = maxLengthInChars(6)
  .then(allCaps(Locale.current))

where you can define your own transformations as

fun allCaps(…): InputTransformation = …

[Zach Klippenstein (he/him) [MOD]]

2:

inputTransformation = InputTransformation
  .maxLengthInChars(6)
  .allCaps(Locale.current)

where you define your own transformations as

fun InputTransformation.allCaps(…): InputTransformation =
  this.then(…)

[Zach Klippenstein (he/him) [MOD]]

2 is a copy of the Modifier style, and requires a bit more ceremony to write custom transformations but is cleaner to use.
1 is simpler to write custom transformations, but harder to discover (no auto-complete) and more verbose to use

[Zach Klippenstein (he/him) [MOD]]

The currently-shipping code defines transformations as extension on InputTransformation.Companion, which is kind of the worst. Forces you to write InputTransformation. before every one, even within the same chain.

[annsofi]

I’d say 2 because discoverability is more important to me at least

[efemoney]

Same I agree with 2.

[Dario Ahdoot]

Same. 2

[Zach Klippenstein (he/him) [MOD]]

Wow alright it’s tight but I think 2 is just slightly ahead

[saket]

I was thinking of voting for 1 because the chaining of transformations with then() makes it clear that they're applied in order, but I immediately remembered that 2 matches how we think of reactive operators and modifiers