Issue 15163 - Parser bug on double function call

[9rnsr]

https://issues.dlang.org/show_bug.cgi?id=15163

Add allowAltSyntax parameter to isDeclarator, and disable it from parseStatement.

[ibuclaw]

My only request here would be to use self-documenting enums rather than continuing to use an integer for the needId parameter.

[9rnsr]

Walter doesn't like to contain both refactoring and bugfix in a PR. So for the quick bugfix, I usually keep such the magic interger numbers in old code. And after the bugfix, the refactoring change will be ignored usually by Walter, because he doesn't like continuous refactoring changes!

[ibuclaw]

And after the bugfix, the refactoring change will be ignored usually by Walter, because he doesn't like continuous refactoring changes!

I could take that argument as ad-infintum. One day it's the number 3, the next week it's a PR documenting what it means to have the number 42 passed to the function. :-)

OK, I'll give this the 👍 for now, but I'll put strong emphasis on refactoring this in the next months. I'd have to reject anything which involves more magic numbers than I have fingers on my left hand (that excludes the thumb) though.

Regardless of Walter's preferences on PR etiquette, so long as you keep each relevant change as a separate commit, others who review on a commit-by-commit basis will understand what's going on.

[ibuclaw]

Auto-merge toggled on

[yebblies]

Walter doesn't like to contain both refactoring and bugfix in a PR. So for the quick bugfix, I usually keep such the magic interger numbers in old code. And after the bugfix, the refactoring change will be ignored usually by Walter, because he doesn't like continuous refactoring changes!

Yeah, I think it's better to have the refactoring separate to the bugfix PR too. I don't have your knowledge of some parts of the compiler, and the more lines changed the harder it is to review. I will happily merge a followup PR that updates it to use an enum if you want to do that.

[9rnsr]

@ibuclaw Thanks. I opened #5184 while this bugfix is in my brain.

[9rnsr]

Yeah, I think it's better to have the refactoring separate to the bugfix PR too.

@yebblies If a bugfix requires preparation refactorings for clean & understandable change, the refactoring PR will block following bugfix PR. I already have a bugfix PR #5009 as the example.