Exposes the interface being added in D59214 for ASTMatchers.
Details
Diff Detail
- Repository
- rC Clang
Event Timeline
| include/clang/ASTMatchers/ASTMatchers.h | ||
|---|---|---|
| 6425 | From https://reviews.llvm.org/D59214#change-5GOe7CiMwWxl const Stmt *OMPExecutableDirective::getStructuredBlock() const {
assert(!isStandaloneDirective() &&
"Standalone Executable Directives don't have Structured Blocks.");It really doesn't make sense to return false/nullptr, | |
| include/clang/ASTMatchers/ASTMatchers.h | ||
|---|---|---|
| 6425 | Hmm, actually, i guess it would be best to return false here instead then. Will change. | |
| include/clang/ASTMatchers/ASTMatchers.h | ||
|---|---|---|
| 6393 | "Matches standalone OpenMP directives, i.e., directives that can't have a structured block." | |
| 6404 | Please follow the existing comment style. Either: Given \code <code snippet> \endcode <matcher expression> matches "<code snippet>". or Example: <matcher expression> matches "<code snippet>" in \code <code snippet> \endcode For example: Given \code #pragma omp parallel default(none) #pragma omp parallel default(shared) #pragma omp parallel \endcode ``ompDefaultClause()`` matches ``default(none)` and ``default(shared)``. Similarly for other comments in this patch. | |
LGTM aside from a NFC change
| include/clang/ASTMatchers/ASTMatchers.h | ||
|---|---|---|
| 6440–6441 | Rather than call the matcher (which is a heavy-handed solution), I'd prefer to just check Node.isOMPStructuredBlock() directly. | |
"Matches standalone OpenMP directives, i.e., directives that can't have a structured block."