kadircet added a comment. > Emitting fixes for a diagnostic in another file seems dangerous, what's the > intended use case for this?
one particular example is fixing layering violations in more strict build systems like bazel/blaze. for example if you `#include "a.h"` in a file, but build target of that particular file doesn't depend on a target exporting `a.h` clangd can generate an edit to update the relevant build target to have a dependency on `a.h`. ================ Comment at: clang-tools-extra/clangd/Diagnostics.h:113 + const ClangdDiagnosticOptions &Opts, + llvm::function_ref<void(clangd::Diagnostic, llvm::ArrayRef<Fix>, + llvm::ArrayRef<CodeAction>)> ---------------- we keep broadcasting fixes and actions separately to keep embedders that can't handle codeactions happy. we can start emitting just codeactions here by converting fixes to codeactions in this layer (all we need is fileuri, and we have it). then embedders can still iterate over all code actions and pick only the edits for mainfile while ignoring the rest. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D96245/new/ https://reviews.llvm.org/D96245 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
