> > Absolutely not. Why should I interrupt a nice sensible command-line
> > workflow to edit a boring file? You're talking not only about
> > expanding a 1-second task to *at* *least* 10 seconds or more just to
> > fire up the editor and fool around,
>
> Who said anything about an editor?
>
> >>_darcs/prefs/boring echo '^(./)?yow$'
Not so easy. If I'm working on GHC, it is almost certainly
../../_darcs or sometimes ../_darcs. On other projects _darcs might
work. Do I want to keep track of the location of _darcs by hand?
I do not! That's what tools are for. And prefs/boring is not
reliable either because somebody might have set the boringfile
preference.
<rant>
I want something that *works*, not something that probably works most
of the time. That means if I'm sitting in compiler/cmm and try to add
CmmParse.hs as a boring file, I don't have to worry about the pathname
relative to _darcs---the tool should figure it out.
> > I can't write a reliable script because darcs has no 'getpref' so I
> > can't easily discover the identity of the boringfile my script should
> > be editing.
>
> x=awk '/^boringfile/ { print $2; }' _darcs/prefs/prefs
> x=${x:-_darcs/prefs/boring}
>
> Now $x contains the location of the boring file.
But can I rely on it? What if the format of the prefs file changes?
If darcs setpref exists, so should darcs getpref. Otherwise, what's
abstraction for?
</rant>
Norman
_______________________________________________
darcs-users mailing list
[email protected]
http://lists.osuosl.org/mailman/listinfo/darcs-users
