eddyz87 wrote: @peilin-ye ,
Sorry for the delayed reply. > > But there should be a way to tweak existing `fail` function to stop after > > errors are reported. > > Can we `exit(1)` ? :-) Let's not hard-code the `exit(1)` in the backend. In case if LLVM is used as a library, such exit might be quite unexpected. I double checked and clang's diagnostic handler is installed in `clang::BackendConsumer::HandleTranslationUnit()` (clang/lib/CodeGen/CodeGenAction.cpp:294) it redirects all handling to `BackendConsumer::DiagnosticHandlerImpl()`. It does not seem customizable. I'd say let's go with what you suggest, but avoid `exit(1)` call. We need to figure out how to do `fail()` w/o backtrace, but that is outside of the scope for this pull request. https://github.com/llvm/llvm-project/pull/108636 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
