kicad-developers team mailing list archive

[jp charras <jp.charras@xxxxxxxxxx>]

Le 07/02/2018 à 15:25, Jon Evans a écrit :
> Yes, we can only throw an error if a wire connects to only invisible pins at a given location.

Wires connected to only invisible pins is not necessary a error:
It happens for most of power symbols.

> I will add this to my queue since I am already planning on some reworks to ERC code after the 5.0
> release.
> 
> Thanks,
> Jon
> 
> On Wed, Feb 7, 2018 at 8:36 AM, Maciej Sumiński <maciej.suminski@xxxxxxx
> <mailto:maciej.suminski@xxxxxxx>> wrote:
> 
>     The ERC check sounds sensible. I know that invisible pins are also used
>     to place many pins in a single spot, so you can connect multiple pins
>     with just one wire (think of e.g. ground connections). We may need to
>     take this case into account.
> 
>     Cheers,
>     Orson
> 
> 
>     On 02/07/2018 02:32 PM, Jon Evans wrote:
>     > We could hold this until after 5.0 and either add an ERC check,  or better
>     > yet make it so that you can't actually make connections to invisible pins?
>     >
>     > -Jon
>     >
>     > On Feb 7, 2018 08:29, "Maciej Sumiński" <maciej.suminski@xxxxxxx
>     <mailto:maciej.suminski@xxxxxxx>> wrote:
>     >
>     >> Hi Wayne,
>     >>
>     >> No, I have not reviewed the patch. I had some doubts about potential
>     >> problems caused by invisible pins creating hidden connections. If user
>     >> is neither aware of their presence when selecting a symbol, nor will
>     >> notice them after they are placed on a schematic sheet then he may end
>     >> up accidentally connecting them to some wires. IIRC we do not have an
>     >> ERC test to check against such case, so I was not sure if it is a safe
>     >> change.
>     >>
>     >> Cheers,
>     >> Orson
>     >>
>     >> On 02/07/2018 02:21 PM, Wayne Stambaugh wrote:
>     >>> Orson,
>     >>>
>     >>> Did you ever respond to Jon about this?  I guess the question is whether
>     >>> or not to show invisible pins in the component chooser.
>     >>>
>     >>> Cheers,
>     >>>
>     >>> Wayne
>     >>>
>     >>> On 1/15/2018 9:31 PM, Jon Evans wrote:
>     >>>> Hi Orson, patch is attached again, hopefully it goes through this time.
>     >>>>
>     >>>> Thanks,
>     >>>> Jon
>     >>>>
>     >>>> On Mon, Jan 15, 2018 at 4:24 AM, Rene Pöschl <poeschlr@xxxxxxxxx <mailto:poeschlr@xxxxxxxxx>
>     >>>> <mailto:poeschlr@xxxxxxxxx <mailto:poeschlr@xxxxxxxxx>>> wrote:
>     >>>>
>     >>>>     On 15/01/18 10:00, Maciej Sumiński wrote:
>     >>>>
>     >>>>         Perhaps we should have an ERC rule
>     >>>>         that warns about invisible pins being connected to a wire, any
>     >>>>         thoughts?
>     >>>>
>     >>>>
>     >>>>     Invisible pins are used for three distinct applications.
>     >>>>
>     >>>>     The first one is to remove clutter by hiding pins that should not be
>     >>>>     connected. ERC will complain if you connect such pins if they have
>     >> the
>     >>>>     electrical type "Not connected".
>     >>>>
>     >>>>     The second application is to create "power labels". A invisible
>     >> power
>     >>>>     input pin is handled as a global label. These pins are meant to be
>     >>>>     connected.
>     >>>>
>     >>>>     The third application is again to remove clutter by stacking pins.
>     >> Here
>     >>>>     you have one visible pin and several other invisible pins at the
>     >> same
>     >>>>     location. (Normally all these pins have the same name and electrical
>     >>>>     type. With the exception of power input pins, power output pins and
>     >>>>     output pins.)
>     >>>>     Such pins are again meant to be connected.
>     >>>>
>     >>>>     This means a ERC rule that complains about connecting hidden pins
>     >> will
>     >>>>     create too many false positives. Having a lot of false positives
>     >> means
>     >>>>     users will start to ignore ERC output.
>     >>>>
>     >>>>     It might be a good idea to have a symbol checker that complains if
>     >>>>     invisible pins are used differently than i described above.
>     >>>>     In other words: complain for invisible pins if they are not part of
>     >> a
>     >>>>     stack or of types NC or power input.


-- 
Jean-Pierre CHARRAS