Experienced the problem described here: http://lists.openstreetmap.org/pipermail/tilesathome/2008-February/001835.html http://lists.openstreetmap.org/pipermail/tilesathome/2008-April/002106.html https://bugs.launchpad.net/inkscape/+bug/202103/+viewstatus It seems, as the links describe, there is a way to use a switch for compiling the garbage-collector to allow working with larger data. Could that possibly be used to compile Inkscape/GC for Fedora, if not in general maybe at least on x86_64 (which might theoretically have a bit more memory) or so?
Thanks for the report Stefan. I've CCed GC's maintainer, Rex -- His opinion is definitely needed here. Also, to me it seems like the launchpad bug you linked to is another issue.
building gc with --enable-large-config seems perfectly reasonable, I'll look into it.
yes, sorry - you might be right about the launchpad-bug ...
gc built with --enable-large-config: rawhide: http://koji.fedoraproject.org/koji/taskinfo?taskID=712359 F-9 (scratch build): http://koji.fedoraproject.org/koji/taskinfo?taskID=712365
OK, a legit and working F-9 (scrach) build: http://koji.fedoraproject.org/koji/taskinfo?taskID=712400
ping, anyone have a chance to try out the test gc build?
Ping?
sorry, can't (shortly) test this myself
So the new gc is in effect; and noone complained so far, so I am closing this.