kicad-developers team mailing list archive

[Strontium <strntydog@xxxxxxxxx>]

On 13/10/16 22:21, Wayne Stambaugh wrote:
> I'm going to keep this brief, please respond in kind.  I'm going to
> create a new entry in the version 5 road map over the weekend with the
> pertinent requirements so this will be my last comments on this subject.
>   The horse is dead.
Noted.
> Why the bypass?  Just change the connectivity algorithm to keep the
> currently assigned nets.  Your patch creates a maintenance issue and is
> confusing.
The connectivity algorithm uses the net of an entity set as Unassigned 
as an internal flag to test if the entity is reconnected or not yet.  
Its global data.  To change that would require changing the global data 
structure that holds entities.  I tired to keep the change minimal and 
self contained. It could be done, as you suggest, with a change to the 
global data structure of entities, if that's the preference.
> The via net defines to which entity it is connected.
No, it doesn't, not currently.
> There are no
> assumptions here nor would I ever be comfortable with making any such
> assumptions.
KiCad currently makes the assumption that the entities net can be traced 
back to a pad and the pads net is the real net of the entity, and if it 
can not, then the Net of the entity is unimportant and therefore 
"Unassigned".  The net assigned to the via or track is currently 
irrelevant to DRC.

Steven