On 2019-02-07, Ricardo Wurmus wrote: >> * gnu/packages/python-xyz.scm (python-daemon): New variable. > > Thanks for the patch! Thanks for the lightning fast review! >> + (add-before 'check 'disable-tests >> + (lambda _ >> + ;; FIXME: Determine why test fails >> + (substitute* "test/test_daemon.py" >> + (("test_detaches_process_context") >> + "skip_test_detaches_process_context"))))))) > > Before pushing this I’ll add a final #T to this phase. (In the past #F > would indicate failure, but that’s on longer the case. “substitute*” > annoyingly returns an unspecified value.) My eyes still have trouble identifying where exactly this would go... and #T/#F vs. #t/#f has me a bit confused for good measure! Obviously, it'd be better to fix the test, or at least understand why it fails... >> + (propagated-inputs >> + `(("python-docutils" ,python-docutils) >> + ("python-lockfile" ,python-lockfile) >> + ("python-setuptools" ,python-setuptools))) >> + (native-inputs >> + `(("python-unittest2" ,python-unittest2) >> + ("python-testtools" ,python-testtools) >> + ("python-testscenarios" ,python-testscenarios) >> + ("python-mock" ,python-mock) >> + ("python-docutils" ,python-docutils))) > > Docutils is both among native inputs as well as regular inputs. Is this > intended? I *think* it only needs to be in native-inputs; I think propagated outputs was generated by the importer... and now that my attention is drawn to it, I'm dubious of python-setuptools needing to be in propagated-inputs as well. I can update it and check again later... >> + (license (list license:asl2.0 license:gpl3+)))) > > What does this list mean? I guess it's best practice in Guix to make a comment when multiple licenses are listed? Apparently only setup.py is gpl3+, everything else is apache 2.0 licensed. That was my quick look at the code itself, and basically confirmed by: https://sources.debian.org/src/python-daemon/2.2.3-1/debian/copyright/ live well, vagrant