Received-SPF: pass (google.com: domain of linux-bluetooth+bounces-2094-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Precedence: bulk
MIME-Version: 1.0
References: <20240221111901.2807853-1-joakim.tjernlund@infinera.com>
 <20240221111901.2807853-2-joakim.tjernlund@infinera.com> <CACvgo52M-Efhp92u=fd8qDjV39SWvuoakLgt2JtFftDmZ2pn4g@mail.gmail.com>
 <4fdde6807ed152d0d24e086fd3357b7c4e410391.camel@infinera.com>
In-Reply-To: <4fdde6807ed152d0d24e086fd3357b7c4e410391.camel@infinera.com>
From: Emil Velikov <emil.l.velikov@gmail.com>
Date: Fri, 23 Feb 2024 11:09:27 +0000
Message-ID: <CACvgo51GX0e5hHzVSD1AEMOLU8hiaDpE0R9su+jjN0mw5HaSBw@mail.gmail.com>
Subject: Re: [PATCH] CUPS lives in libexecdir
To: Joakim Tjernlund <Joakim.Tjernlund@infinera.com>
Cc: "linux-bluetooth@vger.kernel.org" <linux-bluetooth@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"

On Thu, 22 Feb 2024 at 13:22, Joakim Tjernlund
<Joakim.Tjernlund@infinera.com> wrote:
>
> On Thu, 2024-02-22 at 13:06 +0000, Emil Velikov wrote:
> > On Wed, 21 Feb 2024 at 11:19, Joakim Tjernlund
> > <joakim.tjernlund@infinera.com> wrote:
> > >
> > > From: Joakim Tjernlund <Joakim.Tjernlund@infinera.com>
> > >
> > > Signed-off-by: Joakim Tjernlund <Joakim.Tjernlund@infinera.com>
> > > ---
> > >  Makefile.tools | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/Makefile.tools b/Makefile.tools
> > > index 044342e29..b785ebea8 100644
> > > --- a/Makefile.tools
> > > +++ b/Makefile.tools
> > > @@ -538,7 +538,7 @@ endif
> > >  endif
> > >
> > >  if CUPS
> > > -cupsdir = $(libdir)/cups/backend
> > > +cupsdir = $(libexecdir)/cups/backend
> > >
> >
> > Is this true, can you provide some references?
> >
> > Looking at Debian [1] (which contrary to Arch has libexec) - backends
> > are in /usr/lib. Glancing at the cups backends makefile it's using
> > SERVERBIN. Casual grep shows that libexec can be applicable for
> > BSD/Darwin platforms [3], which are out of scope for bluez IIRC.
>
> on Gentoo you have:
>
> qlist net-print/cups | grep backend
> ...
> /usr/libexec/cups/backend/implicitclass
> /usr/libexec/cups/backend/beh
> /usr/libexec/cups/backend/serial
> /usr/libexec/cups/backend/parallel
> /usr/libexec/cups/backend/driverless-fax
> /usr/libexec/cups/backend/driverless
> /usr/libexec/cups/backend/cups-pdf
>
> but now I see this in Gentoo ebuild:
>         # Install in /usr/libexec always, instead of using /usr/lib/cups, as that
>         # makes more sense when facing multilib support.
>         sed -i -e 's:CUPS_SERVERBIN="$exec_prefix/lib/cups":CUPS_SERVERBIN="$exec_prefix/libexec/cups":g' configure ||die
>
>         econf "${myeconfargs[@]}"
>
>         sed -i -e "s:SERVERBIN.*:SERVERBIN = \"\$\(BUILDROOT\)${EPREFIX}/usr/libexec/cups\":" Makedefs || die
>         sed -i -e "s:#define CUPS_SERVERBIN.*:#define CUPS_SERVERBIN \"${EPREFIX}/usr/libexec/cups\":" config.h || die
>         sed -i -e "s:cups_serverbin=.*:cups_serverbin=\"${EPREFIX}/usr/libexec/cups\":" cups-config || die
>
> so it appears to be a Gentoo addition.
> Would perhaps
>   pkg-config cups --variable=cups_serverbin
> or
>   cups-config --serverbin
> be acceptable? If no we can just drop this and Gentoo will continue to patch bluez for cups
>

Using the foo-config tools sounds better than the current hard-coded
behaviour. Patch that does that gets a big thumbs-up from me.
Personally I have slight inclination towards the pkg-config - although
it'll be up-to the maintainers.

Aside: It'll be great if the Gentoo team probably propose their cups
change upstream - either as the default or opt-in behaviour.

HTH
Emil