Re: [gentoo-dev] Re: Apply patch depending on USE flag

[Drake Wyrm]

On Thu, 2004-04-29, 20:34:12 +0000, in
<200404292034.i3TKYCgi064017@×××××××××××××××××××××××.net>,
brettholcomb@×××××××.net top-posted:
> >
> > From: "Michael Sterrett -Mr. Bones.-" <msterret@××××.com> Date:
> > 2004/04/29 Thu PM 07:33:31 GMT To: Lars Strojny <lars@×××××××.net>
> > CC: gentoo-dev@l.g.o Subject: [gentoo-dev] Re: Apply
> > patch depending on USE flag
> >
> > On Thu, 29 Apr 2004, Lars Strojny wrote:
> >
> > > On Thu, 29 Apr 2004 18:20:48 +0100 Tom Wesley
> > > <tom.wesley@××××××××.com> wrote:
> > >
> > > > On Thu, 2004-04-29 at 18:18, Ajai Khattri wrote:
> > > > > I would like to have an ebuild apply a patch based on whether
> > > > > a USE flag is set or not - what is the standard way to do this
> > > > > in an ebuild? (Or is there some function already available to
> > > > > do it?).
> > > >
> > > > You can
> > > >
> > > >         if use nntp; then
> > > >             epatch ${DISTDIR}/${nntp_patch}
> > > >         fi
> > >
> > > I think
> > > if [ "`use nntp`" ] ; then
> > > is more elegant ;)
> >
> > Between the two, the first example is preferred.
> >
> Out of curiosity and because I'm doing some ebuilds and would like to
> do them correctly why is the first preferred?

I think that, eventually, they would like to phase out the requirement
that use() have any kind of output. Also, decide for yourself which
would be faster to determine and/or the more elegant solution:

1) the exit status of use()

2) the exit status of the `[` built-in after performing command
   substitution on the stdout from use()

-- 
Batou: Hey, Major... You ever hear of "human rights"?
Kusanagi: I understand the concept, but I've never seen it in action.
  --Ghost in the Shell