Men-cotton wrote: Thanks for the feedback. I agree that these test-only changes do not add enough value to justify a standalone PR.
For context, @koparasy encouraged me to get involved in OpenCL support for CIR, which led me to start working on [these PRs](https://github.com/llvm/llvm-project/pulls?q=is%3Apr+author%3AMen-cotton+in%3Atitle+CIR+OpenCL). I intended this PR as preliminary cleanup before working on CIR's representation of OpenCL built-in object types, including pipes. The decision to do this cleanup first and to make it a separate PR was entirely mine. On reflection, this preparatory cleanup is not a prerequisite for that work. I had seen multi-line `RUN` commands elsewhere in `clang/test/SemaOpenCL`, such as [this test](https://github.com/llvm/llvm-project/blob/8b6145d2aa547768cede4bf934db963890124550/clang/test/SemaOpenCL/features.cl), and applied that formatting too broadly here. That example separates a substantially longer command and feature list, whereas splitting the `RUN` lines in the files changed by this PR does not improve readability. I also agree that the diagnostic annotations should stay at their original locations in these files. I mostly contribute to MLIR and CIR, but I should have paid closer attention to the local style of these Clang tests. Sorry for the unnecessary changes. I will close this PR. https://github.com/llvm/llvm-project/pull/220269 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
