gEDA-dev: pcb/cygwin

[Dan McMahill]

Bert Timmerman wrote:
> Hi Dan and all,
> 
> And after being called in for having dinner, and putting the kids to bed, I
> ran ./pcbtest and didn't get the message: "Share installation path is...".

good!  that should only happen with mingw.

> Hi Dan,
> 
> I did a fresh check out of pcb.
> 
> ./configure ran without any problems.
> 
> make did the usual things and I went to make a fresh cup of coffee (or two)
> when I saw it hitting the pcblib-newlib script.
> 
> On a P4 3GHz 1Gb and 80GB available disk space it took about 68 minutes to
> build, andit stumbled on the puller.png (I forgot the maintainer mode again
> ;-).
> 
> Now for people with older hardware this is certain a job to run at night,
> while you're sleeping.

yeah, I'm not happy with how long this takes.  It shouldn't take so long 
but on some systems, calling sh over and over is expensive.


> I noticed a couple of things on which I would like to make some remarks.
> 
> 1) I think now is the time to reconsider "light versus heavy footprints", by
> this I mean: should all values of the 0.25W resistors be available as heavy
> footprints ?

no.  I'm planning on removing a bunch of those.

> 2) Some, but not all, footprints seem to have tabs between parameters in the
> pin/pad definitions, for instance: all amphenol footprints.
> 
> 3) the amp and bourns footprints (and maybe others I didn't check) have a
> lot of comment lines in the beginning of the footprint file, and also
> between the lines.

it is because you're getting the raw output instead of something that 
has been loaded and resaved by pcb.


> 4) I think adding another search path for newlib style footprints is not a
> Good Thing (TM), is there a reason for not adding the generated M4
> footprints to the newlib directory.

I'll think about this.