Add multiple cases to NonEmptyString support #375
Conversation
lbialy
added
kind/improvement
lbialy
changed the title
| @@ -130,6 +130,27 @@ trait BesomSyntax: | |||
| res <- ctx.readOrRegisterResource[A, EmptyArgs](companion.typeToken, name, EmptyArgs(), CustomResourceOptions(importId = id)) | |||
| yield res | |||
|
|
|||
| extension (s: String) | |||
There was a problem hiding this comment.
I love those converters, very useful!
| import besom.util.interpolator.* | ||
| inline def fromStringOutput(inline s: Output[String]): Output[NonEmptyString] = ${ fromStringOutputImpl('s) } | ||
|
|
||
| private def fromStringOutputImpl(expr: Expr[Output[String]])(using quotes: Quotes): Expr[Output[NonEmptyString]] = |
There was a problem hiding this comment.
wow, I didn't know you can poke inside interpolators with macros like that, impressive
| @@ -39,8 +64,34 @@ object NonEmptyString: | |||
| private def fromImpl(expr: Expr[String])(using quotes: Quotes): Expr[NonEmptyString] = | |||
| import quotes.reflect.* | |||
|
|
|||
| // kudos to @Kordyjan for most of this implementation | |||
| def downTheRabbitHole(tree: Term): Either[Term, String] = | |||
There was a problem hiding this comment.
this is quite dense, I would appreciate adding some comments to guide the reader through this code, esp. for people not fluent in macros
I understand that we are recursively examining the AST and trying to figure out all cases where it can be proven the result string will be not empty, but I'm sure I'm missing some important details.
In the interest of maintenance, adding comments would be greatly appreciated.
There was a problem hiding this comment.
Sure, no problem. Give me a minute.
|
Ah. One more thing, please update the Changelog :) |
…till necessary for cross-compile unit inference?
lbialy
force-pushed
the
lbialy/extend-nonemptystring
branch
from
faa74d9 to
8e1686b
Compare
done! |
No description provided.