Hi Ales,
> [snip]
> >I recently came across some symbols with strange pin syntax which made
> >gattrib segfault. I have fixed gattrib to be robust against this
> >syntax, but I am now wondering if the syntax is valid or not.
Here's the example I previously used:
P 300 4200 0 4200 1 0 1
{
T 136 4250 5 8 1 1 0 0 1
pinnumber=1
T 136 4250 5 8 0 0 0 0 1
pinseq=1
T 375 4150 3 8 1 0 0 0 1
Vref
}
The person who made the symbol attached the label Vref to the pin as
text. What he wanted to do was attach an attribute "pinlabel=Vref".
By the way, this problem would be avoided if there was a special
object type "attribute", which acted just like regular text, except
that it was treated specially when handling attributes. But we've
talked about that already . . . . . . ;-)
> This is not valid. attributes are text items which are always in
> the form name=value. Anything else is not considered an attribute.
> You cannot attach just plain text items as attributes.
Therefore, only attributes are allowed to be attached to a pin. I
like it.
> [snip]
> >It has text "Vref" attached to the pin. Is this valid?
> >
> >Many other symbols have this construct also.
>
> Many? As in the ones in the symbol library? If so, I better
> add some checks to gsymcheck to catch them.
Yeah, I found a few, but I don't have a list in front of me. As for
modifying gsymcheck, that is where I am going with this. I'd be happy
to take a look at building this check into gsymcheck. After I finish
up a few other projects, that is. . . . .
Stuart
