Nicolas Dumazet <nicd...@gmail.com> added the comment: Hello,
1) Can you please avoid putting several statements in the same line? 2) wouldnt it be better to compute only once the contents of methods()? I'm not sure that module-initialization time is okay for CPython, but at the very least you can lazily fill a module-level variable, and return it directly from methods()? 3) what happens when a user uses one of the Crypt methods that are referenced from the Module, if this method is not available? Arguably, if I know what I'm doing, I will call mksalt(METHOD_SHA512) without checking that METHOD_SHA512 was in methods(). That's not very intuitive, and it seems that mksalt could break. 4) saltchars should probably be string.ascii_letters+string.digits instead of the hardcoded value 5) you should mention in the documentation that if not salt parameter is given, a different salt will be used for each crypt() call 6) is _MethodClass an old-style class? 7) it seems that the patch duplicates twice the diff of crypt.py, not sure of what happened there? ---------- nosy: +nicdumz _______________________________________ Python tracker <rep...@bugs.python.org> <http://bugs.python.org/issue10924> _______________________________________ _______________________________________________ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
