On Wed, May 02, 2018 at 02:41:59PM +0100, Ramy Elkest wrote: Personally I would put focus on outputting the Node code, rather than porting the generator from Python to Node, since the former is more immediately useful to developers. By all means port the generator to Node later at the end once the bindings are succesfully being generated. Even if you try to auto-generate APIs, the libvirt API complexity and variance is such that you'll inevitably end up manually writing a good number of API bindings, as we've done in Python. A key thing to thus understand is what level of coverge you have of libvirt APIs. So I would suggest writing a test that analyses the combination of the manual code and autogenerated code, to validate that together they have 100% API coverage. We did this in the python binding with sanitytest.py or in the Go binding with api_test.go or Perl with t/030-api-coverage.t if you want insipiration. Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|