Re: gEDA-user: gsch2pcb problem: caused by guile/Mandriva?

[Peter Clifton <pcjc2@xxxxxxxxx>]

On Sun, 2007-11-11 at 23:17 +0100, Richard Rasker wrote:
> Op zondag 11-11-2007 om 20:55 uur [tijdzone +0000], schreef Peter
> Clifton:
> > On Sun, 2007-11-11 at 21:38 +0100, Richard Rasker wrote:
> > > Hello,
> > > 
> > > Since I installed Mandriva 2008 (which offers guile 1.8.1), gsch2pcb
> > > doesn't seem to work any more. After drawing a schematic with gschem,
> > > and checking for (and of course correcting) errors with gnetlist, this
> > > is what happens when gsch2pcb is invoked (Filter.sch contains a few
> > > dozen components):
> > > 
> > > $ gsch2pcb Filter.sch
> > > 
> > > ---------------------------------
> > > gEDA/gnetlist pcbpins Backend
> > > This backend is EXPERIMENTAL
> > > Use at your own risk!
> > > ---------------------------------
> > > 
> > > ERROR: Unbound variable: nil
> > 
> > It sounds like you might not be using a recent enough version of gEDA to
> > be compatible with guile 1.8.
> 
> I downloaded and installed the most recent ISO version I could find,
> geda-install-20070221a.iso.

Heh.. we were just having a discussion on IRC as to whether many people
installed from that CD, and whether it warranted making a new one.

What distribution do you use? There may be an efficient place to grab
and install a later version.

[snip]

> > If you have others, open up the .scm file and replace the "nil" with
> > "#f".
> 
> OK, I got half a dozen more of these. Applying your suggested changes
> got rid of the first error message; however, the other problem persists:
> 
> $ gsc2pcb Filter.sch
> ...
> Using the m4 processor for pcb footprints
> /usr/bin/m4:stdin:37: bad expression in eval: / 2

That might be a bug in the PCB footprint library, or an illegal
character somewhere. m4 does not like "-" characters appearing in
footprint names - do you happen to have one? I'm sure there have been
various bugs fixed which caused similar errors, bug I can't remember
whether they were before / after the CD version you have.

> >From here on, m4 is claiming 100% CPU, with memory usage skyrocketing,
> and the size of the resulting .pcb file just keeps growing and growing,
> until I kill the job. After this, I'm left with a seemingly correct
> netlist file, and a correct .cmd file.
> The .pcb file, however, is not OK at all. When opened with PCB, an error
> message about a syntax error shows up:
> 
>   ERROR parsing file '/home/rr/Filter.pcb'
>       line:        54
>       description: 'syntax error'
> 
> Line 54 contains a reference to a footprint definition I created:
> 
>   PKG_CHOKE_SU9V(CHOKE_SU9V,TX1,SU9V)

m4 is used to expand lines like that into "native" or "newlib" format
data. (This is the format PCB has always worked from).

> This is followed by a dozen or so empty lines. There are more of these
> PKG_ lines, most of which seem to be connected to symbols and/or
> footprints I created.
> For all the rest, the .pcb file seems to contain mostly valid items, but
> at a certain point, this shows up:
> 
>   Element(0x00 "" "" "" 142 50 3 100 0x00)
>   (
> 	Pad(0 25 50 25 20 "1" 0x00)
> 	Pad(0 75 50 75 20 "2" 0x100)
> 	Pad(0 125 50 125 20 "3" 0x100)
> 	Pad(0 175 50 175 20 "4" 0x100)
> 	Pad(0 225 50 225 20 "5" 0x100)
> 
> The list with pad definitions goes on for thousands of lines, with ever
> growing numbers:
> 
> 	Pad(0 10975 50 10975 20 "220" 0x100)
> 	Pad(0 11025 50 11025 20 "221" 0x100)
> 	Pad(0 11075 50 11075 20 "222" 0x100)
> 	Pad(0 11125 50 11125 20 "223" 0x100)
> ...
> 	Pad(0 1742375 50 1742375 20 "34848" 0x100)
> 	Pad(0 1742425 50 1742425 20 "34849" 0x100)
> 	Pad(0 1742475 50 1742475 20 "34850" 0x100)

> 
> No doubt, this is the result of m4 being caught in some endless loop. I
> don't quite understand what is wrong here -- both the symbol and
> footprint definitions I created in PCB and gschem respectively seem just
> fine (with all the right names, pinseq, pintype, other attributes
> etcetera), and I get no complaints from either application when using
> these definitions. Neither does gnetlist find anything wrong.

Your diagnosis seems correct, but I don't know what's caused it.

If you're able to, send a copy of the schematic / custom footprints and
someone (perhaps me) will take a look.



Peter C




_______________________________________________
geda-user mailing list
geda-user@xxxxxxxxxxxxxx
http://www.seul.org/cgi-bin/mailman/listinfo/geda-user