> I wish Bison were in C++… Funny, I always have been thinking about the same.
What are showstoppers? - people won't accept cpp tool in their kernel/OS/distribution bootstrap process? - people have access to cpp compiler during bootstrap, but don't like cpp in principle? - rare architectures may not have cpp toolchain? - possible enough Bison is not so complex program, so maintaining cpp code might has more profits. Also, old bison releases are always available for those who has no cpp toolchains... FYI: I ported flex to native (VC++) toolchain on windows couple years ago and moved it to C++. Changes were big enough (in philosophy too), that PR was not accepted. (test suite passed fine) https://github.com/westes/flex/pull/77 I'm also thinking of porting of bison in such way. I already did some experiments during last several months. The biggest issue is same as for flex - custom or portable m4. -- Egor Pugin
