Mail Archives: cygwin-apps/2001/11/03/22:02:18
> -----Original Message-----
> From: Christopher Faylor [mailto:cgf AT redhat DOT com]
> Sent: Sunday, November 04, 2001 1:59 PM
> To: cygwin-apps AT cygwin DOT com
> Subject: Re: [Fwd: Re: cygwin-apps Digest 2 Nov 2001 18:46:34 -0000
> Issue 223]
>
>
> On Sun, Nov 04, 2001 at 01:57:48PM +1100, Robert Collins wrote:
> >
> >> -----Original Message-----
> >> From: Christopher Faylor [mailto:cgf AT redhat DOT com]
> >>Then the package should have a diff dependency, shouldn't it?
> >
> >Well, if the config stuff gets automated, the diff dependency becomes
> >one of setup's, not of the individual packages.
>
> Same argument, though. I am probably beating a dead horse because I
> like diff and use it scores of times a day, but if diff is a
> dependency
> of "something" then "something" should include it as a dependency.
Yes, I agree. What I was indicating was that setup.exe's dependencys are
not explicitly listed outside our discussions here ... as you've covered
below.
> If we don't do this then someone will say "huh, diff. I
> don't know what
> that is." and unselect it with less than satisfactory results.
They can only deselect it because we aren't advanced enough to pop up a
dialog saying "foo bar fly depend on diff, do you really want to
deselect it". And for stuff we require that popup could say something
like "system won't work right if you do this.."
> At some point, I think an unselect of a package that is the
> target of a
> dependency should have an "are you sure?" prompt. Is this there
Hmm, that might be easy enough, the problem is that we then need to
transactionally manage the whole statemachine - and it's knock on
effects.
I'm comfortable that the number of users who _deselect_ something are
going to be << the number that don't know they should have selected it
in the first place.
Rob
- Raw text -