Recent changes to path_conv::check -- ouch

[egor duda]

Hi!

Monday, 11 June, 2001 Christopher Faylor cgf@redhat.com wrote:

CF> [following up on myself]
CF> Kazuhiro's patch has forced me to look into this function a little
CF> more.  It looks like we're just calling rootdir repeatedly rather than
CF> invoking the OS.  I guess that is ok.  I'd rather not have even that
CF> perfomance penalty but I don't know if it can be avoided.

yes, i was trying to make update_fs_info() as fast as possible in most
cases. if all symlinks in path stay on the same drive/share, OS is
called only once. As for GetCurrentDirectory(), i wasn't sure if
mount-converted path always has rootdir component present. if we can
guarantee that, GetCurrentDirectory() can be eliminated.

repeated calls to rootdir() is, most of time, just strchr()+strcmp(),
and as i've removed appropriate calls from path_conv::check() i don't
think we'll have any performance hit at all, or it'd be negligible
compared to calls to GetFileAttribute().

Egor.            mailto:deo@logos-m.ru ICQ 5165414 FidoNet 2:5020/496.19