X-Authentication-Warning: delorie.com: mail set sender to geda-user-bounces using -f
X-Recipient: geda-user AT delorie DOT com
Message-ID: <20230331011920.15319.qmail@stuge.se>
Date: Fri, 31 Mar 2023 01:19:20 +0000
From: "Peter Stuge (peter AT stuge DOT se) [via geda-user AT delorie DOT com]" <geda-user AT delorie DOT com>
To: geda-user AT delorie DOT com
Subject: Re: [geda-user] lepton-sch2pcb: wrong ${prefix} for pcb
References: <20230330032511 DOT 12329 DOT qmail AT stuge DOT se>
 <20230330064204 DOT 16FA585F72C3 AT turkos DOT aspodata DOT se>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20230330064204.16FA585F72C3@turkos.aspodata.se>
Reply-To: geda-user AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: geda-user AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

karl AT aspodata DOT se [via geda-user AT delorie DOT com] wrote:
> > > 1, the dirname trick:
> > >  $(dirname $(dirname $(which pcb)))/share/pcb
> > >  followed by m4, pcblib-newlib alt. newlib depending of what
> > >  we are looking for
> > 
> > This isn't a good choice at all since it makes assumptions about how
> > pcb was compiled - specifically that there is a fixed relationship
> > between the pcb binary path and the library path.
> 
>  Well, there seems there is one, at least per default:
> 
> $ cd <git>/pcb
> $ grep -B1 BINDIR_TO_PCBLIBDIR config.h
> /* Relative path from bindir to pcblibdir */
> #define BINDIR_TO_PCBLIBDIR "../share/pcb"
> $
> 
>  which can be changed with
>   --datarootdir=DIR       read-only arch.-independent data root [PREFIX/share]
> 
>  ./configure --datarootdir=/opt
>  gives:
> /* Relative path from bindir to pcblibdir */
> #define BINDIR_TO_PCBLIBDIR "../../../opt/pcb"

Exactly; because it can be changed by pcb configure it's
unfortunately not possible to assume anything.


> > > 2, convince pcb devs. to add some cmdline option like
> > >  $ pcb --libpath
> > >  /usr/local/share/pcb
> > 
> > This is by far the best!
> 
>  Attached patch gives me (on stdout, without X required):
> $ pcb --show-libdir
> /home/local/bin/../share/pcb
> $

It's a short and sweet patch but I guess we don't want to wait for it to
be included and then a release before using this functionality so for
now we can make do with what unpatched pcb offers.


> > karl AT aspodata DOT se [via geda-user AT delorie DOT com] wrote:
> > > pcb already have:
> > >   boxA:
> > > $ pcb --show-defaults 2>&1 | grep lib-command-dir
> > > lib-command-dir "/home/local/bin/../share/pcb"
> > > 
> > >   boxB:
> > > $ pcb --show-defaults 2>&1 | grep lib-command-dir
> > > lib-command-dir "/usr/bin/../share/pcb"
> > 
> > Perfect! Great find.

This is a good substitute for the explicit option added by your patch.


> > karl AT aspodata DOT se [via geda-user AT delorie DOT com] wrote:
> > > >  But unfortunately, pcb depends on X beeing available:
> > > > $ pcb --show-defaults
> > > > Error: Can't open display: 
> > > 
> > > "pcb -x ps --show-defaults" solves that.
> > 
> > Okay - attached is a C snippet to grab the string.
> 
>  The problem is that -x ps might be disabled with
> ./configure --with-exporters= --with-printer=
>  so we cannot rely on any exporter to be available at all
>  and we have to accept that X has to be up and running to
>  probe the value for unpatched pcb.

It's even possible to configure --with-gui= and have no hid at all.

How about we first try pcb -p --show-defaults, then pcb -x ps, then
with default hid, then fail?

If there's no X then gschem/lepton might anyway not start
(interactively) - right?


//Peter