delorie.com/archives/browse.cgi | search |
>> Right. I missed the "." in the original message. The change that >> prompted this behavior seems to be >> <http://cygwin.com/ml/cygwin-cvs/2005-q3/msg00224.html>. I'm assuming the >> motivation for this patch was to duplicate Linux's behavior (which doesn't >> allow trailing "." in a path passed to mkdir). > > Indeed. Eric mentioned that the coreutils testsuite tests thsi behaviour > explicitely. Since there's not much impact speedwise, we just added > appropriate checks to be POIX compatible here. I want to state, that cygwin might return the wrong error-code! Instead of "file already exists" it returns "no such file or directory". Cygwin's bahaviour: $ mkdir /tmp/. mkdir: cannot create directory `/tmp/.': No such file or directory Linux' behaviour: # mkdir /tmp/. mkdir: cannot create directory `/tmp/.': File exists Indeed, strace shows me, that on Linux mkdir() returns EEXIST in the case of a mkdir("/tmp/.")-call. Cygwin doesn't seem to do it this way, it seems to return ENOENT which would not be Linux-like ;-) Unfortunatly i don't understand the output of cygwin's strace and cannot check, what the mkdir("/tmp/.") returns there. Does anybody have a clue? Sven -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
webmaster | delorie software privacy |
Copyright © 2019 by DJ Delorie | Updated Jul 2019 |