I was thinking recently about layering HIDs. I'm using two HIDs for EPS export; one just gathers data and the second does the actual drawing. I'll probably use the same trick for gerber, since we need to enumerate the apertures. Anyhoo, I was also fixing a bug in the lesstif layer that I noticed while using thindraw, and a thought occurred to me. Previously, I had decided that "thindraw" was a HID-specific thing; the core didn't need to have the logic to implement thindraw (hence the separate "trace" and "pad" end caps :). The new thought was that we could have a "thindraw HID" that did nothing but convert the desired drawing commands (line, circle, etc) to thindraw commands, and pass them along to the underlying HID.
Not only does this mean all the GUIs get thindraw with little effort, but we can also do thindraw Postscript or EPS plots. Since the EPS plots stack all the layers, and can't support the masking we need for polygon clearances, I think thindraw would be useful there. Comments?
