Change-Id: I63d67bf6e0b315e2d3360e47e3b62c9517f38987
Details
Diff Detail
- Repository
- rL LLVM
- Build Status
Buildable 15679 Build 15679: arc lint + arc unit
Event Timeline
LGTM. Few minor style questions.
docs/TableGen/LangRef.rst | ||
---|---|---|
98–101 | Should we sort them? | |
lib/TableGen/TGParser.cpp | ||
1251–1252 | Would it make sense to print StartUntyped here, too? I often find TableGen failures baffling, specifically because tablegen will usually (sometimes?) tell you *where* tablegen ran into the problem, but it rarely tells you what exactly it saw after all the folding that made it unhappy. Better error reporting in general should probably be a patch, but I doubt there will be many voices against better error reporting in new code. :-) | |
1251–1258 | You seem to mix fold and !foldl in the error messages throughout the function. Is that intentional? I'd use !foldl everywhere for consistency. |
r326790
docs/TableGen/LangRef.rst | ||
---|---|---|
98–101 | Probably a good idea at some point, but I'd rather not do it in this commit to avoid conflicts with later changes. | |
lib/TableGen/TGParser.cpp | ||
1251–1252 | Yup, makes sense, I've added that, same with ListUntyped below. | |
1251–1258 | You're right, I've changed all the messages to say !foldl consistently. |
Should we sort them?