Mail Archives: geda-user/2012/07/12/09:46:57
On 12/07/12 05:18, Andrew Poelstra wrote:
> On Thu, Jul 12, 2012 at 12:13:52AM +0100, Ed Simmons wrote:
>> Sorry for the noise, I've figured it out. Dividing from the base
>> unit and printing an integer works fine for my purposes.
>>
>> pcb_fprintf(fp, "3 : 230: 1: 3: 0:F 1X %iY %iA %iR %s %s %s\n",
>> x/10000 , y/10000 , theta * 10, name, value, FRONT
>> (element) == 1 ? "top" : "bot");
>>
> Please don't use base units -- %i won't work on 64-bit systems, and
> isn't guaranteed to work on any systems in the future. Instead use
> the %mn specifier to output coordinates as nanometers. (Then you can
> use /10000 as you have.)
>
>
Thanks, that's looking much better now. I did think it was a bit ugly to
use the base unit, considering there's obviously been plenty of work to
avoid using the base units directly.
I'm now finding another issue... I've made a fiducial symbol and
footprint, and I'm successfully identifying fiducial 1 in the exporter,
then trying to get it's location to use as the bias offset for all the
location data that is subsequently output.
Using the same element loop and pin/pad loops to find the location of
the fiducial seems to not work properly for me yet. I always get X axis
location as 0, but Y looks as though it is working.
This is what I have:
ELEMENT_LOOP (PCB->Data);
{
if (NSTRCMP ( NAMEONPCB_NAME (element), "FID1") == 0)
{ // hooray! fiducial 1 has been found
pcb_fprintf (fp, "# Found fiducial 1\n");
int count = 0;
PAD_LOOP (element);
{
sumx += (pad->Point1.X + pad->Point2.X) / 2.0;
sumy += (pad->Point1.Y + pad->Point2.Y) / 2.0;
pcb_fprintf (fp, "# Sum X = %.2mm Sum y = %.2mm\n",
sumx,sumy); //seems here that X is never getting a value other than 0
count++;
}
END_LOOP;
pcb_fprintf (fp, "# Found %i pads\n", count);
if(count > 0 && count <= 2){ // the fiducial can have pads
top and bottom!!
offsetX = sumx;
offsetY = sumy;
}
}
}
END_LOOP;
Any bright ideas would be greatly appreciated.
Thanks again.
Ed
- Raw text -