[Libre-silicon-devel] Magic CIF output support for LibreSilicon

David Lanzendörfer david.lanzendoerfer at o2s.ch
Sat Aug 18 17:46:36 CEST 2018

Hello Tim

> Hello David,
> > Thanks! I'm looking forward to your feedback!
> > The recent CIF output definition I've built spits out the manufacturing
> > masks we require for running our test tape out.
> > There is however maybe a problem with wrong CIF layers or so.
> > I'm looking forward to hearing your opinion on what I've done.

> My first question would be:  Why are you creating a complete copy of the
> magic-8.2 repository just to post a techfile?
So that you can "git fetch" and review, and I might actually help you 
developing in the long run, so I need to push my changes somewhere ;-)

> My first comment would be:  Don't use "cpp" for creating techfiles.  It
> is a preprocessor for C source code and really is very poorly suited to
> anything else.  I tried "m4" for a while on new techfiles before coming
> to the conclusion that it is even worse than cpp.  I made my own
> preprocessor called "preproc.py" in the scripts/ subdirectory of magic,
> which is probably the better preprocessor to use, although I have not
> yet tried to convert the original SCMOS techfile sources to it (well, I
> seem to have started it a year ago but never finished).  I use it at
> efabless to generate techfiles and PDKs for all the tools.  It takes the
> usual -D defines on the command line and parses #ifdef, #ifndef, etc.,
> like cpp, but without all the C-language-specific weirdness.

Okey. I just replicated the way you create tech files and tech includes.
I could just skip the gcc part and just directly include it into your 
scmos.tech generation.
It looked kind of odd anyway that you're using gcc for generating the tech 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 195 bytes
Desc: This is a digitally signed message part.
URL: <http://list.libresilicon.com/pipermail/libre-silicon-devel/attachments/20180818/655d4dcc/attachment-0001.sig>

More information about the Libre-silicon-devel mailing list