Keyword sanitization for Lens code generation (#2924) #2925
Conversation
|
Thanks for your contribution, @hrafnthor! May I suggest to use the entire set of keywords instead of a few? You can copy the list from here. |
* Set moved out of function to avoid its creation on each invocation. * Mock test package header now has keywords from each category.
| fun String.sanitizeDelimited(delimiter: String = ".", separator: String = delimiter) = | ||
| if (isNotBlank() && contains(delimiter)) { | ||
| splitToSequence(delimiter).joinToString(separator) { if (kotlinKeywords.contains(it)) "`$it`" else it } | ||
| } else this |
There was a problem hiding this comment.
I don't think you need to consider empty and not containing delimiter as special cases, splitToSequence + joinToString should do the right thing
| if (isMarkedNullable) "$n?" else n | ||
| } | ||
| else -> when (val qname = declaration.qualifiedName?.asString()) { | ||
| else -> when (val qname = declaration.qualifiedName?.asString()?.sanitizeDelimited()) { |
There was a problem hiding this comment.
It might read better if asStringSanitized was an extension of Name (but if you prefer this way, it's also fine)
There was a problem hiding this comment.
I agree, it would be more readable as well as limit the scope of the function to relevant parts i.e KSName rather than any String. Will change the extension signature and apply it where applicable.
There was a problem hiding this comment.
Thank you @hrafnthor for being so quick to fix your own bug report 🙏 👏 Really awesome!
And congrats on your first contribution to Arrow 🙌
* Extension function applied to KSName rather than String * Function signature modified to match its function better * Unnecessary function checks removed based on feedback
Kotlin keyword sanitization added to class path fields used during Lens code generation.
Only keywords matching potential TLDs are observed.