dbartol wrote: @AaronBallman
> are these stable IDs usually displayed to the user in tools consuming SARIF? > If so, are we sure these default IDs are reasonable for those cases? Or do we > have to worry about tools expecting stable IDs to be short (so our long > identifiers might get cut off such that the user only sees warn_foo_w and not > warn_foo_was_bad_because_bar? The SARIF consumer I'm most familiar with is GitHub Advanced Security, which does display the rule ID, but not particularly prominently. Screenshot attached, with rule ID highlighted. <img width="724" height="627" alt="GHAS Alert" src="https://github.com/user-attachments/assets/321fa0de-61c2-44e2-9a36-952ff77622a5"; /> As far as ID length goes, MSVC keeps it short (e.g., `C4101`), but that's mostly because MSVC has had those stable 4-or-5-digit warning IDs since at least the mid-90's. However, CodeQL has significantly longer IDs, like `cpp/local-variable-hides-global-variable` or `cpp/comma-before-misleading-indentation`, which seem comparable to Clang's existing warning enum names. I think we're OK using the IDs we've already got. https://github.com/llvm/llvm-project/pull/168153 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
