Comment by KacperHarasim (JIRA)

Hi @akyrtzi,
I would like to attempt to implement this.
I see this as a range based refactoring with the following behavior:
User chooses a range of text and every string value that is within the range is converted. This way, if in example provided range begins on "h" and ends on name variable result would be the following:

let text = "hello \(name)" + "!"

Since "!" was not selected it was not converted.
Does it make any sense? What do you think?

Thanks for the feedback,

Yes, I think it makes sense as range refactoring.

Note that if the range includes other expressions (or statements) unrelated to string concatenation then it should not be available.

Comment by KacperHarasim (JIRA)

That seems reasonable. Thanks!

Comment by KacperHarasim (JIRA)

Hi @akyrtzi
I've started to look around the source code in order to implement it and I have question:
Should I rather look at TokensInRange property of ResolvedRangeInfo and check for sequence of binary operators and literals, or maybe use SourceEntityWalker to walk through ResolvedInfo.ContainedNodes property? What's the clearer way to do things in this case?

Hi kacperh (JIRA User),

I prefer we use SourceEntityWalker to walk through ContainedNode (in this case, a single expression) to analyze its content instead of scanning tokens. The reason is token contains much less semantic information that AST node does.

Comment by KacperHarasim (JIRA)

Thanks @nkcsgexi!
One last question(I hope :-) ): I’ve tried to go this approach and my logic was: scan expression with SourceEntityWalker and check whether it contains literals, Binary + expressions, interpolations and string objects. If given any other type of Expression I intended to report refactor as unavailable. Problem is that running it on expression like:

„a” + „b” + „c”

Results in going through DotSyntaxExpr. I’m not really sure why is it here, because I don’t see any expression like that in the example. It’s probably result of my limited knowledge about a compiler. Could you please point me to some information on why is it included?

Thanks in advance!

Hi kacperh (JIRA User), great progress! You approach makes sense to me. I think the DotSyntaxExpr you see is implicit node synthesized by compiler to help later code generation; i think you can ignore any `implicit()` expressions.

Comment by KacperHarasim (JIRA)

Opened a PR to implement this: #11944

Comment by KacperHarasim (JIRA)

Committed in: 80698a0