EEVblog Electronics Community Forum

Electronics => PCB/EDA/CAD => KiCad => Topic started by: c4757p on January 02, 2014, 03:54:17 pm

Title: Recent KiCad testing - paths for footprints/3D models?
Post by: c4757p on January 02, 2014, 03:54:17 pm

So... KiCad's PCB library system was just changed a few revisions ago, and seems to be completely undocumented... Has anybody else gotten a chance to play with this? I figured out how to import my footprint library. They've apparently done away with the traditional search path and moved to substitution variables. (And it took me a while to figure out to select Legacy, because they've changed the format. I'll get around to porting my library sooner or later...)

My problem is that I can't figure out how to bring in my 3D models. The name of each model is specified in the footprint, and before, I could just add the model directory to the search path and they would show up. Now there are no search paths. :-//

By the way, there are some interesting choices under "Plugin Type": Github, Eagle, Geda. Haven't tried the Eagle one yet, I'm not sure if it actually works or if it's just a placeholder.

Title: Re: Recent KiCad testing - paths for footprints/3D models?
Post by: liquibyte on January 02, 2014, 09:09:21 pm

I can't really answer your question but I can help you look into it.  What branch are you using?  bzr branch lp:~cern-kicad/kicad/ratsnest perhaps?  You've reminded me to get this and try out the new features.  I had to compile wxWidgets myself because the package on my distro is out of date.  I'm trying to build now but keep getting errors so it may take a while to get right.

Title: Re: Recent KiCad testing - paths for footprints/3D models?
Post by: c4757p on January 02, 2014, 09:17:04 pm

I'm using the official testing branch (https://code.launchpad.net/~kicad-testing-committers/kicad/testing), on Debian Wheezy if that matters to you, presently at revision 4598.

Title: Re: Recent KiCad testing - paths for footprints/3D models?
Post by: liquibyte on January 02, 2014, 09:27:08 pm

I run Archlinux so I have to get kicad-bzr 4190 from the AUR or build from source manually.  The stable branch is a 2013/05/18 build so I don't think it has any of the neat stuff in it yet.  Neither is building for me at the moment and I'm starting to think the toolchain is broken somehow for whatever reason.  The cern-kicad/kicad/ratsnest branch is the stuff I think that's merged from cern to be included in stable at a later date but I'm not entirely sure on that.  I was directed there by the mailing list.  When I try to build the cern branch, my compiler segfaults and I can't seem to get past that.

Title: Re: Recent KiCad testing - paths for footprints/3D models?
Post by: c4757p on January 02, 2014, 09:31:33 pm

I build from source manually as well, because 1) it's easy and 2) I've had problems with all the .deb builds.

The official kicad/testing branch has the CERN stuff now. It looks like cern-kicad/kicad/ratsnest is defunct now, probably abandoned when the changes were merged upstream. Try the official branch.

Title: Re: Recent KiCad testing - paths for footprints/3D models?
Post by: liquibyte on January 02, 2014, 09:57:09 pm

Just tried stable and I get:

Code: [Select]

In file included from /usr/local/include/wx-3.0/wx/stdpaths.h:185:0,
                 from /home/liquibyte/kicad/src/stable/common/edaappl.cpp:39:
/usr/local/include/wx-3.0/wx/unix/stdpaths.h: In member function 'bool EDA_APP::SetBinDir()':
/usr/local/include/wx-3.0/wx/unix/stdpaths.h:56:5: error: 'wxStandardPaths::wxStandardPaths()' is protected
     wxStandardPaths() { }
     ^
/home/liquibyte/kicad/src/stable/common/edaappl.cpp:454:32: error: within this context
     m_BinDir = wxStandardPaths().GetExecutablePath();
                                ^
common/CMakeFiles/common.dir/build.make:1034: recipe for target 'common/CMakeFiles/common.dir/edaappl.cpp.o' failed
make[2]: *** [common/CMakeFiles/common.dir/edaappl.cpp.o] Error 1
CMakeFiles/Makefile2:1061: recipe for target 'common/CMakeFiles/common.dir/all' failed
make[1]: *** [common/CMakeFiles/common.dir/all] Error 2
Makefile:126: recipe for target 'all' failed
make: *** [all] Error 2

I think something is not quite right with my toolchain.  I'll keep trying but it may take awhile.

Edit: On line 454 of edaappl.cpp I had to change the following:

Code: [Select]

//m_BinDir = wxStandardPaths().GetExecutablePath();
m_BinDir = wxStandardPaths::Get().GetExecutablePath();

Got passed that error and now it's:

Code: [Select]

In file included from /home/liquibyte/kicad/src/stable/pcbnew/io_mgr.cpp:30:0:
/home/liquibyte/kicad/src/stable/common/../pcbnew/eagle_plugin.h:95:5: internal compiler error: Segmentation fault
     }
     ^
Please submit a full bug report,
with preprocessed source if appropriate.
See <https://bugs.archlinux.org/> for instructions.
common/CMakeFiles/pcbcommon.dir/build.make:849: recipe for target 'common/CMakeFiles/pcbcommon.dir/__/pcbnew/io_mgr.cpp.o' failed
make[2]: *** [common/CMakeFiles/pcbcommon.dir/__/pcbnew/io_mgr.cpp.o] Error 1
CMakeFiles/Makefile2:1127: recipe for target 'common/CMakeFiles/pcbcommon.dir/all' failed
make[1]: *** [common/CMakeFiles/pcbcommon.dir/all] Error 2
Makefile:126: recipe for target 'all' failed
make: *** [all] Error 2

I usually quit when code is this poorly implemented but I'll keep plugging away at it.

Title: Re: Recent KiCad testing - paths for footprints/3D models?
Post by: c4757p on January 02, 2014, 10:03:44 pm

Your wx might be too recent - i.e., they may have made non-backwards-compatible changes between 2 and 3. I suspect this is correct because the Deb repo offers both 2.8 and 3.0. I'm compiling KiCad with 2.8, I believe.

Title: Re: Recent KiCad testing - paths for footprints/3D models?
Post by: liquibyte on January 02, 2014, 10:09:47 pm

I'll try that.  I know I just installed 3.0.0, so that may be the issue.

Title: Re: Recent KiCad testing - paths for footprints/3D models?
Post by: liquibyte on January 02, 2014, 10:58:47 pm

I give up for now.  This POS just won't compile for me.  I've hated it since I first tried to use it and thought that maybe they finally fixed a few of the usability issues now that cern is involved.  I guess there's a reason that the main Kicad is as old as it is in Arch, I bet they couldn't get it to compile either.  Wxwidgets, seriously?  Folks, pick a modern tookit, this thing's a joke.

Nothing against you though.  I'll probably try again later but this is making my head hurt.