Enchancement #3008 #3228
Enchancement #3008 #3228
Conversation
|
Hi! This already looks pretty good, a few improvements that could be made:
|
|
By the way, to check that the modified test works correctly, you can run just the REPL printing tests by typing in sbt: dotty-compiler/test-only -- --tests=typePrinterTests |
| @@ -126,8 +126,8 @@ class RefinedPrinter(_ctx: Context) extends PlainPrinter(_ctx) { | |||
| def isInfixType(tp: Type): Boolean = tp match { | |||
| case AppliedType(tycon, args) => | |||
| args.length == 2 && | |||
| !Character.isUnicodeIdentifierStart(tycon.typeSymbol.name.toString.head) | |||
| // TODO: Once we use the 2.12 stdlib, also check the @showAsInfix annotation | |||
| !Character.isUnicodeIdentifierStart(tycon.typeSymbol.name.toString.head) && | |||
There was a problem hiding this comment.
This should be an || not an && as we want infix types to be used either for type constructors which start with a symbol or that have the @showAsInfix annotation.
There was a problem hiding this comment.
There's one more problem. I didnt notice that this annotation takes a boolean argument, which should be checked as well.
|
@smarter Thanks for pointing me to tests, looks like there's a problem - |
|
Ah you're running into a known bug (#2462), you can work around it by using the fully qualified name of the annotation e.g: scala> @scala.annotation.showAsInfix type Mappy[T, U] = Map[T, U] |
CucumisSativus
force-pushed
the
show_as_inflix_used_in_refined_printer
branch
2 times, most recently
from
d5b83d8 to
1a2ad9c
Compare
Refined printer checks if shoAsInflixAnotType annotation is used. Based on "SI-4700 Make infix notation default for symbolic types" from original scala repository
CucumisSativus
force-pushed
the
show_as_inflix_used_in_refined_printer
branch
from
1a2ad9c to
5d70d1d
Compare
|
@smarter I added tests, the only problem I have right now is using type alias |
|
Hmm this is probably the Dotty typechecker being too eager to dealias things, I'll look into it but it might be tricky to fix. We can merge this PR without the type aliases working since it already improves things anyway. |
|
@smarter so my proposition is to merge this pull request and open a new issue where someone (maybe even me, with some guidance :) ) can check how to approach type aliases |
|
LGTM, thanks for your contribution! |
I'm afraid this might be a hard issue to fix :) I suggest you take a look at https://github.com/lampepfl/dotty/issues?q=is%3Aissue+is%3Aopen+label%3Aexp%3Anovice or ask on http://gitter.im/lampepfl/dotty if you're looking for other things to contribute. |
Thanks for your guidance and help :) |
Refined printer checks if showAsInflisAnotType annotation is used (issue #3008).
I would be grateful if you could lend me a hand in testing my solution, couldn't find any tests for RefinedPrinter.
If something is missing please let me know, i'm happy to improve this PR :)