Re: cygwin1.dll bug in open(O_EXCL)

2007-08-24 Thread Corinna Vinschen
On Aug 24 10:49, Eric Blake wrote: > As a side effect of your change, open("broken_symlink", O_RDWR|O_EXCL) now > fails with EACCES instead of ENOENT, but since POSIX leaves O_EXCL without > O_CREAT as undefined behavior. That's why I didn't add any further testing for other cases with O_EXCL. Isn

Re: cygwin1.dll bug in open(O_EXCL)

2007-08-24 Thread Eric Blake
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 According to Corinna Vinschen on 8/24/2007 8:56 AM: >> According to POSIX, this should have failed with EEXIST, and oops should >> not have been created. > > If I understand this right, it means that O_EXCL implies not following > symlinks. I've appl

Re: cygwin1.dll bug in open(O_EXCL)

2007-08-24 Thread Corinna Vinschen
On Aug 24 07:25, Eric Blake wrote: > -BEGIN PGP SIGNED MESSAGE- > Hash: SHA1 > > I finally found root cause for the tar issue reported here: > http://cygwin.com/ml/cygwin/2007-08/msg00553.html > > $ cat foo.c > #include > #include > #include > #include > #include > #include > > int

cygwin1.dll bug in open(O_EXCL)

2007-08-24 Thread Eric Blake
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 I finally found root cause for the tar issue reported here: http://cygwin.com/ml/cygwin/2007-08/msg00553.html $ cat foo.c #include #include #include #include #include #include int main(int argc, char**argv) { int fd; symlink ("oops", "broke