On Sun, Oct 19, 2008 at 08:11:31PM +0000, Junio C Hamano wrote: > Miklos Vajna <vmiklos@xxxxxxxxxxxxxx> writes: > > > Signed-off-by: Miklos Vajna <vmiklos@xxxxxxxxxxxxxx> > > --- > > > > On Sat, Oct 18, 2008 at 03:17:23AM +0200, Miklos Vajna <vmiklos@xxxxxxxxxxxxxx> wrote: > >> > This adds a new feature to say --no-z from the command line, doesn't > >> > it? > >> > And I suspect the feature is broken ;-). > >> > >> Right, I fixed this in option_parse_z(). --no-z should set > >> line_termination to \n instead of 1. > > > > Originally in option_parse_z() I had > > > > line_termination = unset; > > > > which is in fact right, because (as Pierre pointed out) unset for short > > options are always false, but I changed it to > > > > line_termination = 0; > > > > to make it more readable. > > I think Pierre's comment is short-sighted. Think of what would happen > when somebody adds "--nul" as a longer equivalent to "-z", since it is > extremely easy to do things like that with the use of parse-opt API? Err I was only pointing out that --no-z would no nothing, I actually didn't really read the argument :) I didn't say having --null was a bad idea, and I think we have -z/--null at a couple of places already, and it's probably a good thing to actually _add_ --null. -- ·O· Pierre Habouzit ··O madcoder@xxxxxxxxxx OOO http://www.madism.org
Attachment:
pgpZSjeDtJcz4.pgp
Description: PGP signature
