X-Authentication-Warning: delorie.com: mail set sender to geda-help-bounces using -f X-Recipient: geda-help AT delorie DOT com Subject: [geda-help] PCB segfault - project too big? From: Richard Rasker To: geda-help AT delorie DOT com Content-Type: text/plain; charset=UTF-8 Organization: Linetec Date: Tue, 11 Oct 2011 15:05:56 +0200 Message-Id: <1318338356.29731.2.camel@localhost> Mime-Version: 1.0 X-Mailer: Evolution 2.22.3.1-2.2mdv2008.1 Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by delorie.com id p9BDQoQg014010 Reply-To: geda-help AT delorie DOT com Hello, I'm using gschem + PCB for a largish project (400 components, 2500 pins, 500 nets on a 100 x 160 mm 6-layer board), up until now without problems. Unfortunately, I've run some serious trouble all of a sudden: PCB crashes when importing my latest schematics, and after some experimenting I'm at the point where adding one simple component (e.g. a capacitor, or even a single test pad) makes the difference between crashing and proper functioning. When I run PCB from the CLI and select File -> Import Schematics, everything seems normal at first, but after a few seconds, PCB crashes with no specific error messages, just "Segmentation Fault". Have I stumbled upon the limits of PCB? I don't think that my PC hardware is the problem: its 2GB of memory is never fully used, and there is enough HD space. I also ran gnetlist -g drc2 etcetera on my schematics; no errors were found, apart from lots of unconnected pins and nets which are only connected to one pin -- and hooking up 'nc' elements to these pins and nets doesn't appear to make a difference. So the question is what I can do about this. The irony is that I only need to place another two dozen or so components to finish the design, so it feels a bit as if someone put up a brick wall ten yards in front of the finish line ... Is there a way to obtain more verbose error messages? Or is there a workaround for this problem, other than to eliminate components? Thanks in advance, Best regards, Richard Rasker