On Tue, Aug 23, 2011 at 10:49:12AM +0200, Tomas Mraz wrote:
The fix for Sourceforge bug 3035919 introduced a regression in
pam_access described here:
https://bugzilla.redhat.com/show_bug.cgi?id=732081
The attached patch tries to hopefully fix it and not introduce another
regression in this fragile code. We need to try to split the user
pattern on this regexp:
(@*[^@]+)((a)+.*)
Why this regexp? Looking at the documentation, it should rather be
(@?[^@]+)((a)+.*), shouldn't it?
--
ldv