[clang] [OpenACC] Implement compound construct parsing (PR #72692)
Erich Keane via cfe-commits
cfe-commits at lists.llvm.org
Mon Nov 20 07:42:18 PST 2023
erichkeane wrote:
> > Need to add tests
>
> I mentioned in the commit message, that unfortunately this doesn't change behavior in a way that we can discover in our testing (unless you know how to check which character a diagnostic was issued on with `VerifyDiagnosticsConsumer`?). I'd left the tests for these in place already, but the `loop` was being interpreted as the beginning of a `clause`, so it ends up being diagnosed as `unknown clause`. However, now that we're consuming 'loop' as well, we still get the same diagnostic for the NEXT token.
Actually, just realized there would be a difference in behavior for `pragma acc parallel loop` (with no clause). So I'll add tests for those.
https://github.com/llvm/llvm-project/pull/72692
More information about the cfe-commits
mailing list