On Sun, Mar 06, 2005 at 08:16:02PM +1100, Anand Kumria said
> On Sun, Mar 06, 2005 at 04:21:26PM +1100, Rob Weir wrote:
> > Hrm, I thought the baz source regexp was broad enough to include this.
> > But now I look, ^core$ is in the backup regexp. Should it just be
> > removed? Or should "import -a" have magic "set
> > backup/unrecognized/junk/precious regexps to empty" behaviour? Should
> > it record this override in {arch}/=tagging-method?
>
> I prefer magic myself. Sure it'll mean that subsequant commits might be
> an issue but it means that 'baz import -a' Just Works.
>
> And as we all know, as soon as you've put those two words in capitals,
> you can no longer argue with any solution proposed. Longer term you'd
> probably want to move away from regexps for =tagging-method and into
> filespecs or something.
>
> That make subsequant commits easy to fixup without having to learn the
> baroque tagging-method regexps GNU Arch has.
On Sun, Mar 06, 2005 at 10:17:34PM +1100, Robert Collins said
> On Sun, 2005-03-06 at 20:16 +1100, Anand Kumria wrote:
>
> > > Hrm, I thought the baz source regexp was broad enough to include this.
> > > But now I look, ^core$ is in the backup regexp. Should it just be
> > > removed? Or should "import -a" have magic "set
> > > backup/unrecognized/junk/precious regexps to empty" behaviour? Should
> > > it record this override in {arch}/=tagging-method?
>
> I think that import should force the regexps to
> junk ^$
> precious ^$
> backup ^$
> unrecognized ^$
> source .
>
> - this is more in line with the baz goals of 'included everything until
> ignored.
Ok, this is what I was thinking of. Do you mean encoding that in
{arch}/=tagging method?
-rob
--
Words of the day: Kennedy SDI Compsec subversive Firefly quarter Yukon Merlin
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
