Corrects wrong insertion point for 'await' at uncovered 'async' site.

[Rajveer100]

Fixes Issue #65913

[Rajveer100]

@YOCKOW @AnthonyLatsis

[AnthonyLatsis]

Please add a test. As is, this change will most likely break some tests, so I’m going to trigger CI for you to inspect the failures.

[AnthonyLatsis]

@swift-ci please smoke test macOS

[Rajveer100]

Please add a test. As is, this change will most likely break some tests, so I’m going to trigger CI for you to inspect the failures.

Well, I suspect that too as we must probably handle the cases separately for each type of expression in an async await call expression (ex. tryExpr, closure, _, etc), let's see!

Regarding tests, I presume there would be cases as mentioned above, let me know if I missed out any!

[Rajveer100]

After some debugging I found that only these two lines by themselves are causing an issue:

auto anchorDecl = dyn_cast<DeclRefExpr>(anchor);
auto anchorVar = dyn_cast<VarDecl>(anchorDecl->getDecl()); (Trigger line)

Regarding the tests itself, before they are even checked, there seems to be internal error (probably memory error/bad access/similar), hence shows 'Please submit a bug report' as stderr.

Test file: ClangImporter/objc_async.swift

[Rajveer100]

@AnthonyLatsis
I am closing this one, since @Jamezzzb has a PR that's pretty close and does not break the tests. Looks like I had force unwrapped the null values which should have instead been checked.

[AnthonyLatsis]

For the future: Don’t be shy to ping someone once every 2 weeks or so in PRs if you want a review. We sometimes do forget to get back to stuff.

[Rajveer100]

For the future: Don’t be shy to ping someone once every 2 weeks or so in PRs if you want a review. We sometimes do forget to get back to stuff.

Indeed!

PS: By the way I am excited for GSoC 2024, hoping to land proposals just like last year and getting accepted this time, larger contributions are always fun!