Matthew Fosdick
Matthew Fosdick
It already exists in the pattern class and having the group attributes carry around what their group number is, minimizes a lot of the risk of forgetting to pass into...
Sounds good. Maybe @next_pattern? That is always a pattern.
This is due to a lack of support for trailing return types. @jeff-hykin Additionally, keywords should be avoided inside the member access.
A trailing return type is the only valid syntax that looks like that in the function head. More precisely a line that starts with `foo()` can only be a function...
Previous discussion https://github.com/jeff-hykin/cpp-textmate-grammar/issues/371
Care needs to be taken about user-defined template deduction guides. ```c++ template A(C b) -> typename C::type ``` Though I am not sure that there actually is a difference un...
I don't belive that to be valid syntax. > An attribute specifier list may, in future, be permitted to appear after the declarator in a function definition (before any old-style...
Huh, I wonder why clang added support for attributes there.
``` #if DOCTEST_CLANG #ifdef __has_warning #define DOCTEST_CLANG_HAS_WARNING(x) __has_warning(x) #endif // __has_warning #ifdef __has_feature #define DOCTEST_CLANG_HAS_FEATURE(x) __has_feature(x) #endif // __has_feature #define DOCTEST_PRAGMA_TO_STR(x) _Pragma(#x) #define DOCTEST_CLANG_SUPPRESS_WARNING_PUSH _Pragma("clang diagnostic push") #define DOCTEST_MSVC_SUPPRESS_WARNING_PUSH #define...
If non-preprocessor stuff closes on a lookahead for `#endif` but preprocessor stuff consumes the `#endif` wouldn't that prevent double closing?