Maybe deprecating libgweather



Hello all,

If you are a GNOME Weather user, you might have seen bug 770321, which caused any libgweather-using app to be completely not working for the last week, as the NOAA shut down their METAR provided.

While the bug is now fixed, thanks to contributors that promptly provided patches, I think this is a good moment to reflect on the future of libgweather, and my take is that it should be deprecated, and you should move away from it.

At the moment, libgweather provides two things: a location chooser entry and associated location model/db, and a client to a couple weather forecast APIs.

If you're using the first part, as I understand is the case of gnome-initial-setup, maybe gnome-control-center and gnome-clocks, I think you should really move to a pure geocode-glib implementation. You will have better data (as the location db does not have good coverage of the world), less code and fewer bugs (like libgweather moving Polish cities to Germany because the closest airport is across the border).

If you're using the second part, the main advantage you gain is a METAR parser... and that's about it. Call the underlying service yourself, be it OpenWeatherMap or met.no, and you'll have better data, not filtered by libgweather's data model, and not tied to airports.

This said, I do not have time to port GNOME Weather away from libgweather, so I will make sure libgweather stays working (more or less), but all apps that are not GNOME Weather should really move away from it.

Any comments / opinions?

Cheers,

Giovanni


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]