On Sun, Nov 06, 2005 at 07:50:37AM -0500, Dan McMahill wrote: > > > Long story short: I am still trying to decide if there is a bug > >in gnetlist (really libgeda) that causes two nets to be created instead of > >one when you draw two nets that cross each other (but not connected > >together) and a pin connecting to the intersection of the two crossing > >nets. Should the two nets be connect together with the pin even though > >the nets are not connect to each other? Libgeda/gnetlist doesn't seem > >to think so. Right now, I don't think so either, since you did not > >explicitly draw the nets to connect together. However, I am interested > >to hear what other developers think. > > Here are my thoughts > > 1) gnetlist should never produce a netlist that claims a certain pin > connects to two different nets. To me this is the difference between a > netlist which may have not been what you wanted and a netlist which is > flat out broken. > > 2) never, ever, draw schematics the way shown in the example. I > _never_ use 4-way connections. By this I mean, don't draw something > that looks like a + with a solder dot in the middle. Instead draw a > single vertical line and offset the 2 horizontal ones vertically from > each other. This recommendation goes all the way back to pre-computer > era drafting. Visually, it is easy to miss seeing the solder dot or see > one which isn't there (make a few photocopies or fax a schematic). > Also, gnetlist isn't the only netlister I've seen doing somethng funny > in such a case.
This is perfectly valid way how to draw a connection, period. > So to summarize, GIGO (garbage in garbage out), but I'd like to see the In this case the GI is gnetlist and GO is the .net file. CL< > output garbage be self consistent. > > -Dan
