kicad-developers team mailing list archive

[Lorenzo Marcantonio <l.marcantonio@xxxxxxxxxxxx>]

On Mon, Apr 29, 2013 at 01:32:40AM -0500, Dick Hollenbeck wrote:
> Bullshit. just call the interface with the right argument.

Strange for someone that just a few days ago lamented for 'too many
bugs' in kicad...

Sure. *Call* the interface with the *right* argument. The keyword is
*call*.  It's not academic. The DRC checker works with angles in int.
The display and plotting works with angles in int (for display I suppose
is fine, altough easy to fix (and anyway they use it as a double
inside...), for gerbers and DXF not necessarily). The gerber viewer
(mostly) use int angles. The global pad setting dialog uses the int
orientation (angle) both for filtering and for setting.

Oh right, the last one is not necessarily a bug since you have to
specify the angle as tenths of degrees and it doesn't accept decimal
entries. Wasn't that desiderable?

Would someone notice that? Unlikely. Unless they actually need more than
one decidegree of precision (which is desiderable, as you said,
otherwise the angles would be still ints).

And unless you that want to be paid for fixing bugs I can fix these
myself, don't worry. There are 'only' 68 suspicious symbols to be
checked in my queue, for this angle type issue. This is one of the
reasons I like typedef (I am *NOT* proposing an angle typedef, mind
you:D). And this assuming that eeschema is keeping its 90 degree
innards.

Please, note that I am not saying that double angles are not desiderable
or an error. I'm saying that *just* saying that these function take a
double is not enough until you have checked that *everyone* that call
them is calling them with a *real* double (and not something coming from
a double truncated as an int).

Oh, by the way, are module orientation too supposed to be double?
I counted them too since it seemed logical.

-- 
Lorenzo Marcantonio
Logos Srl