Unnecessary warning: Found a zero length pin:

[KaiMartin]

On start-up, gschem issues a warning if pins with zero length are found. E.g:
    Found a zero length pin: [ P 100 100 100 100 1 0 0]

There is no reason to issue a warning. gnetlist does not care for the length of the pins. They are no problem in the GUI, either. When not connected, they are visible, by the red mark.

Proposed solution:
Remove the check for zero length pins on start-up.

---<)kaimartin(>---

[Peter TB Brett (peter-b)]

 status fixcommitted
 done

commit 9b9080e184607b3cb449a1ccb6082207ed39738f
Author: Peter TB Brett <email address hidden>
Date: Thu Jan 20 16:45:48 2011 +0000

    libgeda: Don't warn about finding zero length pins on load.

    There's no actual reason why zero length pins shouldn't be allowed;
    they work fine in gnetlist, and they show up okay in gschem.

    Closes-bug: lp-705397

[Peter Clifton (pcjc2)]

The only thing you might find is that drawing them is difficult.. if it is zero length, the pin cue won't have any orientation.

[KaiMartin (kmk-familieknaak)]

Currently, gschem prevents creation of zero size pins. It silently refuses to create a pin if the mouse cursor indicates zero length.
( see bug 698614, https://bugs.launchpad.net/geda/+bug/698614 )

There is no meaning to the orientation of the cue anyway. Why is it drawn as a little line? Circular blobs would suit the purpose much better. Should I open a tracker report to request circular cue shape?

---<)kaimartin(>---

[Krzysztof Kościuszkiewicz (k-kosciuszkiewicz)]

Circular cues would look strange in postscript output.
Unless another shape is proposed, maybe unconnected net cue shape/size can be reused?
This would fix two problems at once:
 1) cues for zero length pins are drawn incorrectly
 2) cues are not drawn for angled pins at all

Changes to gschem and libgeda PS output are attached.

[KaiMartin (kmk-familieknaak)]

+1 for the reuse of the unconnected net cue shape and size.
After all, that's what they are: Markers of an unconnected net.

---<)kaimartin(>---

[Peter TB Brett (peter-b)]

I for one quite like the "highlight" style of pin cue -- it's unobtrusive and works well. Big square blobs would IMHO look ugly in printed output. However, it's quite true to point out that the cues don't work properly for zero-length pins, when a larger cue would be helpful for locating the pin.

How about drawing a box cue *only* when a pin is below a certain size?

[KaiMartin (kmk-familieknaak)]

How about not highlighting pin cues in postscript output at all?

Seems to me, this is already the current situation. Pin cues show up on the screen but not in print. This plays nice with the slightly different purpose on screen and on paper. On screen it is nice to have not connected pins stand out and remind me where I potentially have to edit. It is good practice to actively mark those pins that need noc connection. There is a special no-connect symbol for this purpose in the library.
Prints are more for documentation. There is no much point in reminding me about unconnected pins at this stage.

---<)kaimartin(>---

[Peter TB Brett (peter-b)]

On Sunday 23 January 2011 00:35:30 KaiMartin wrote:
> How about not highlighting pin cues in postscript output at all?
>
> Seems to me, this is already the current situation. Pin cues show up
> on the screen but not in print.

Yes, they do show up in print. You need to enable printing in colour.

                       Peter

[KaiMartin (kmk-familieknaak)]

They don't show up in color print. See attachments. Binary is git head, compiled yesterday.

[Peter TB Brett (peter-b)]

Huh, that's weird. That's probably *technically* a bug in the printing code.

Having hashed this over with a user group yesterday, I've changed my mind about the pin endings -- unconnected pin ends probably *should* be rendered like unconnected net ends.

However, it is still useful for pins that have length to continue to have directionality when using e.g. magnetic net mode. I suppose that it would be enough just to ensure that zero-length pins are equally attractive in every direction.