thank you @danielmarjamaki for your comments! > I can see that you test with travis. Did you consider to test it with > appveyor also to get windows testing also?
i don't have access to machine which runs windows, neither have experience to develop on windows. that's one of the reasons i would like this code part of Clang repository, that other people could improve it. > ear.c: > > - not written in C++, doesn't seem to me it conforms to LLVM/CLANG coding > style > - As far as I know, `__attribute__` is not ANSI C.. and can't be compiled by > some compilers. > - `unistd.h` is a posix header isn't it? can't be used by some compilers. - about coding style: - those few C files in LLVM and Clang repository does not really follow the C++ style guide either. (so could not decide what is the style for C.) - by it's function it is better to be written in C. (to keep dependencies minimal.) - i think we can rewrite it smartly in C++, but i would require help from others. - about `__attribute__`: since this code require Clang to be on the machine anyway, maybe it is not a strict requirement to expect that it should be compiled by Clang. - about `unistd.h`: yes, currently the command interception relies on POSIX methods. (same thing applies here what i said about windows port.) > typo reach => rich? fixed in commit 25455. thanks for it! http://reviews.llvm.org/D9600 EMAIL PREFERENCES http://reviews.llvm.org/settings/panel/emailpreferences/ _______________________________________________ cfe-commits mailing list cfe-commits@cs.uiuc.edu http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
