I have built a private copy of libgeda implementing my search path:

$system/gafrc
$home/gafrc
./gafrc
$system/gschemrc
$home/gschemrc
./gschemrc

It works fine.  The only problem is that I get a warning when it tries
to find the "required system-gafrc" file:

Did not find required system-gafrc file 
[/usr/local/geda/share/gEDA/system-gafrc]
Did not find optional ~/.gEDA/gafrc file [/home/sdb/.gEDA/gafrc]
Read local gafrc file [./gafrc]
Read system-gschemrc file [/usr/local/geda/share/gEDA/system-gschemrc]
Did not find optional ~/.gEDA/gschemrc file [/home/sdb/.gEDA/gschemrc]
Did not find optional local gschemrc file [./gschemrc]

However, differentiating between optional and required files can be
handled in another way than what we do now.

I like Peter's suggestion that the system-commonrc become the
system-gafrc.  This is a good first step on the road to migrating
towards using gafrc everywhere & deprecating the old g[schem | netlist
| attrib | symcheck]rc system.  It does require some changes to the
*rc.in files -- particularly in the symbols directory -- but that I
don't mind doing.

I don't think the old system should go away; that might cause breakage
of legacy schematics.  However, we can incorporate the new gafrc
system while deprecating the old system for new designs.

One question:  If I make changes to libgeda, I would like to have
folks test my stuff.  Are there any volunteers out there who are
willing to get stuff out of CVS, build it and test it when I post an
update?  Peter?  

In general, it would be a good thing to have a few beta testers who
can verify that the code in CVS builds reliability.

Stuart



> 
> 
> Hi,
> 
> thank you for including my request. 
> 
> I would prefer to read in the gafrc file before the individual rc file. That 
> gives you the chance to make changes for a single program after the common 
> setup is read.
> Ales, what would be the advantage of reading the program rc before the gafrc?
> 
> The order: system-wide, user-wide, local is the best.
> 
> Having two system common rc files is certainly not what we want. Stuarts 
> proposal is very logical. The name gafrc is in all three cases the same. 
> Maybe renaming system-commonrc to system-gafrc would be a compromise. But I 
> don't know how much programming work this is ....
> 
> Peter
> 
> 

Reply via email to