My Magic file has zero DRC error but when I write GDS, the output GDS has a lot of DRC errors. The reason being, the contacts in my layout are split into smaller contact squares and those have DRC issues. I have two questions 1. Magic does not split all contact regions. I have noticed that it leaves really big contacts as-is. Just overlaying metal layers over diffusion of locali does not create automatic contacts either. What triggers Magic to create/ split the contacts? 2, How can I override Magic to not create/split contacts.

Magic represents contacts by contact areas, which is an abstraction of areas that require arrays of contact cuts and surrounding metal. So magic generates the contact cuts in the GDS when it writes the GDS output. When reading the GDS back in, it may or may not re-create the contact area as it was originally, depending on how the input rules are written. Magic layers are translated to and from GDS layers, and sometimes the translation is not exactly reversible, so you cannot be guaranteed that you can recover the same layout view exactly by writing out GDS and reading back in. That said, the contacts, whether they get merged into one area or remain split up into smaller areas around the cuts, should still be DRC clean. If you can post an example layout or screenshot of where the DRC errors are, I can take a look at it.

The error I get is % drc find Error area #2: mcon.spacing < 0.19um (mcon.2)

I don't see any DRC errors when reading in the GDS, other than errors flagging your missing well and substrate taps when you run with the drc(full) deck, but that's true of the .mag file as well as the .gds file.

Thanks Tim I had tried to use DRC full earlier.I pressed "DRC Update" and I see the issues now in mag file as well. "cif see" gives me a lot of visibility too.