From: martin r. <li...@ru...> - 2004-02-18 13:08:47
|
dear sam, i just encountered another problem: in our program, we used a scm-file tree which contains upper/lowercase filenames mixed. i now have problems to load them, because elk now lowers all the filenames. in r5rs, i didn't find a clear answer. it states that case doesn't matter in scheme for identifiers, but i don't find a statement regarding filenames. if case doesn't matter in scheme, then it's perhaps better to make no distinction there for filenames, but that leads to some platform-dependent problems. perhaps we should try to match for upper/lowercase filenames then, which is quite ineffective... another way would be to preserve case for file operations, which is useful for unix systems. on the other hand, mixing upper/lowercase names with the sames letters leads to incompatabilities with case-insensitive filenames. what are your thoughts about that? perhaps we should lower all of our filenames, but it don't know yet what implications this has regarding existing code. thanks, all the best, martin |