kicad-developers team mailing list archive

[Nick Østergaard <oe.nick@xxxxxxxxx>]

2016-03-02 23:23 GMT+01:00 Daniel Silverstone <dsilvers@xxxxxxxxxxxxxxxxx>:
> On Wed, Mar 02, 2016 at 22:54:54 +0100, Nick Østergaard wrote:
>> > If you consider the current situation regarding global labels and heirarchical
>> > labels as "normal" then nothing I have done in my branch alters normal
>> > behaviour.
>> >
>> > The intra-sheet link elements exist only to allow design engineers to mark
>> > otherwise dangling wires as being meant to join (via net-name) to other wires
>> > on the same sheet of the schematic.  If that net *also* has global or
>> > heirarchical labels on it, then they shall behave exactly as before.
>> >
>> > Nothing I have done alters how the wires and pins end up grouped into nets, nor
>> > alters how those nets then derive their names.  It's all for ERC and style checking.
>> >
>> > I hope that makes things a little more clear.
>>
>> Yes, I understand that, but what I don't quite understand is what the
>> result of your ERC will be in the case you have a local label on sheet
>> A, then use the diamond marker on a dangling, but also connect a new
>> global label to that net. Move to sheet B, use the global label and
>> set the diamond or not on sheet B. Will it still report dangling nets?
>
> Phew that took a while to follow, here's what I understand the scenario as:
>
> There are two sheets in this design.  Sheet A has a wire labelled 'WIRE' one
> end of which is a global label (let's say FOO) and the other end of which has
> an intra-sheet link element on it.
>
> Sheet B has a wire, one end of which has the global label FOO and the other of
> which is either dangling or has an intra-sheet link element on it.
>
> What I would expect and desire would be that if sheet A had no other
> FOO-labelled net with an intra-sheet link on it, there'd be an ERC error; ditto
> sheet B.  And that the intra-sheet link on A would not be matched by an
> intra-sheet link on B.
>
> So I tried it...
>
> You had found a glorious corner case I had not considered.  Thank you.
>
> I have pushed a small fix to my series which ensures that the expected
> behaviour (that intra-sheet links do not satisfy one another across sheet
> boundaries) is restored.  (I had failed to check m_SheetPath in my ERC :-)
>
> Thank you.

Ahh, ok. I am glad that I could help in this regard. :)