On 9/26/18 9:03 PM, Thomas De Schampheleire wrote:
El mié., 26 sept. 2018 a las 13:00, Mads Kiilerich
(<m...@kiilerich.com>) escribió:
One concern could be that this command "isn't Kallithea". Kallithea
"itself" must be run from gearbox or another web/wsgi server. That can
be confusing. The command is more about Kallithea
management/install/setup/control/something.
Yes, I have had the same consideration. I had a look at 'Flask', a
micro web framework (which is also using click), and they are also
using 'flask' for the command-line tool although it is not 'Flask',
the project.
We could use 'kallithea-cli' as name, or kallithea-setup, whatever you prefer.
I also don't know the right answer. If you already had the same concerns
and thought about it, your answer is probably as good as any other.
(But I guess, 'serve' could perhaps also be built in ...)
You mean wrapping 'gearbox serve' within the new CLI command?
One reason against it would be that 'gearbox serve' is just one way of
starting the serving. It does not apply when using Apache, uwsgi, or
other approaches.
Yes. Just like the gearbox serve command also only is one of many ways
to serve wsgi.
But it could perhaps be placed under a "development" or "source" sub
command. A command that also monitor file system and rebuild front-end
code automatically. And perhaps also take care of building/updating
localization.
/Mads
_______________________________________________
kallithea-general mailing list
kallithea-general@sfconservancy.org
https://lists.sfconservancy.org/mailman/listinfo/kallithea-general
