Frank Thomas comments

Results 290 comments of


                                            Frank Thomas

Unacceptable compilation time

That caveat only applies if you have lots of static data that you want to check at compile-time. In my projects the vast majority of the validation refined performs happens...

Unacceptable compilation time

@howyp I've not measured how compilation time changed between 0.8.7 and 0.9.0 due to implicit resolution. That would be interesting to know. What I know is that if the macro...

RFE: A version of Inference[P, C] that fails for implicit chaining

`Inference` was designed before we had singleton-ops and the prospect of having them integrated into the compiler. The value-level `Boolean` field was an easy way to do something interesting in...

Draft macroless ==>

I applause the macroless `==>`! One quick question: Is `RefinedLT` only required to implement `equalValidateInference` or are there other reasons or instances it is required for?

My thoughts with regards to `equalValidateInference`: If we accept that the `Inference#isValid` field was a mistake, then `equalValidateInference` was a mistake too. The instance was added without explanation in https://github.com/fthomas/refined/commit/f1812d88042b143b826b8ce423d13501a6c1a197...

Pureconfig not published for scala3

I'd like to keep this open until refined-pureconfig is published for Scala 3.

Feature request: Add runtime exception check instead of error "compile-time refinement only works with literals"

my tl;dr: - This is an interesting idea because it is a safer variant of [`unsafeFrom`](https://github.com/fthomas/refined/blob/e32047bb5a0c12d162d592e533a6089e129f6baf/core/shared/src/main/scala/eu/timepit/refined/internal/RefinePartiallyApplied.scala#L21). - We should not change the semantics of `refineMV`. - By default all functions...

Feature request: Add runtime exception check instead of error "compile-time refinement only works with literals"

I started working on this here: https://github.com/fthomas/refined/compare/topic/unsafeFrom2

Feature request: Add runtime exception check instead of error "compile-time refinement only works with literals"

@AndreasKostler Currently I've little time to work on this. But in principle I'm still okay with changing `unsafeFrom` as outlined in https://github.com/fthomas/refined/issues/188#issuecomment-232068707. The only downside of this approach I'm aware...

Feature request: Add runtime exception check instead of error "compile-time refinement only works with literals"

@ysusuk If we change `unsafeFrom` as outlined in https://github.com/fthomas/refined/issues/188#issuecomment-232068707, you could invoke `createBucketT` with `createBucketT(unsafeFrom(x))` where `x` is of type `String` and it would perform a compile-time check if `x`...