On 12/06/2014 08:50, Justin Bogner wrote:
Alp Toker <[email protected]> writes:--- cfe/trunk/include/clang/Basic/DiagnosticIDs.h (original) +++ cfe/trunk/include/clang/Basic/DiagnosticIDs.h Tue Jun 10 04:31:37 2014 @@ -60,28 +60,28 @@ namespace clang { /// (emit a warning), MAP_ERROR (emit as an error). It allows clients to /// map errors to MAP_ERROR/MAP_DEFAULT or MAP_FATAL (stop emitting /// diagnostics after this one). - enum Mapping { + enum Severity { // NOTE: 0 means "uncomputed". - MAP_IGNORE = 1, ///< Map this diagnostic to nothing, ignore it. - MAP_REMARK = 2, ///< Map this diagnostic to a remark. - MAP_WARNING = 3, ///< Map this diagnostic to a warning. - MAP_ERROR = 4, ///< Map this diagnostic to an error. - MAP_FATAL = 5 ///< Map this diagnostic to a fatal error. + MAP_IGNORE = 1, ///< Map this diagnostic to nothing, ignore it. + MAP_REMARK = 2, ///< Map this diagnostic to a remark. + MAP_WARNING = 3, ///< Map this diagnostic to a warning. + MAP_ERROR = 4, ///< Map this diagnostic to an error. + MAP_FATAL = 5 ///< Map this diagnostic to a fatal error. }; }Seems a bit strange to have an enum called "Severity" with values tagged "MAP_", doesn't it?
Yes, it's a little strange. I didn't change them because there are out-of-tree users (clang-tools-extra at least) and it seems preferable to keep code compiling with a typedef for now while the diagnostic code gets cleaned up.
I'm hoping it'll be possible to reuse one of the several other diagnostic level enums in clang for mapping and get rid of this one completely.
Alp. -- http://www.nuanti.com the browser experts _______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
