Re: gEDA-user: Stupid newbie question

[sdb@cloud9.net (Stuart Brorson)]

> Bill Wilson wrote:
> > On Fri, 21 May 2004 13:04:23 -0500
> > Bill Wilson <billw@WT.NET> wrote:
> > Oh, and if snap was off, the symbols themselves might now not be
> > on grid and you might have to start over placing them because once
> > they are off grid they may not just move back on grid when you move them
> > after turning snap on (not sure about that).
> 
> This was the ticket. The symbols LOOKED like they were on the grid but must 
> have been just slightly off. Even with snap off (to connect off-snap-grid) or 
>   back on (always off by a pixel or two) I wasn't able to establish connections.
> 
> Is there any plan to change this sensitivity? It might be nice to have a 
> snap-to-connection-endpoint function. 

FWIW, many commercial schematic capture programs also have this
drawback.  Viewdraw certainly does.  You need to pay attention to your
snap grid when you make the symbols, and when you place the parts.
Otherwise you end up with the problem you just experienced.   

I just tried it in Orcad (which is running in another window).  Orcad
has the same issue:  If you place a part with snap off, and it is off
grid, then you can't wire it with grid on.  Orcad does something
intelligent, however:  If you try moving an off-grid part when you
turn the grid on, it automatically snaps back to the grid.  (Viewdraw
doesn't do this, IIRC.)  Maybe gschem could implement it this way? 

Another idea is: Rather than having a "snap to end of pin" mode, I'd
suggest having a "move pins to grid" button which you could push to
cause all pins to move to the right place.  If your symbols were
broken (i.e. pins were improperly spaced), implementing this button
might be problematic, but gschem could just throw up a warning that
symbols U23, U45, and U67 couldn't move to the grid. 

Stuart