kicad-developers team mailing list archive

[Dick Hollenbeck <dick@xxxxxxxxxxx>]

On 03/29/2014 12:31 PM, Lorenzo Marcantonio wrote:
> Did the merge. Really less painful than I tought...
> 
> Taking for example pcbnew (which IIRC is the only complely kifaced yet)
> it builds in pcbnew:
> 
> - pcbnew 
> - _pcbnew.kiface
> - _pcbnew.so
> 
> (also this is a non-scripting build; it says it can't find swig... do
> I need to upgrate do 3.0?)
> 
> So the first file is the executable (smallish), the second one (bigger)
> is the DSO containing the actual pcbnew program (I think)... what's the
> last .so for?

The *.so is the swigged pcbnew DSO.  So evidently you ARE building for scripting,
otherwise it should not be built.

The swig DSO is a superset of the *.kiface, adding the python entry-points to what is
otherwise identical to *.kiface.  I'd like to eliminate that duplication, thinking about
it for milestone C), in anycase the duplicate compilation, but probably also the binary
combination can be simplified also.


> 
> Also what are the rules it uses for finding the kiface DSO? some libexec
> path, looks around argv[0] or what?

argv[0], grep for "POLICY", twice in common/single_top.c


> 
> Seems to work fine at the moment, anyway (only a few assertions about
> absolute paths to investigate)
>