It also has “upstream source is not entirely free” anti-tag.
For now, I want to applaud developers based on the criteria stated above: Metadata in F-Droid; what sites users would use to file issue reports and look at source for the app from F-Droid. Maybe later we consider second-order connections in giving a Score.
For now, IMO a note could be added to the listing, but Mull should not be removed.
Gmaps WV and Privacy Browser depend on Webview, which may or not be developed on free systems, depending which Webview you have. Should Privacy Browser be delisted or penalized because of that? All run on devices with non-free firmware. How far do we go with second-order considerations outside the control of the app developers?
IIUC, gitlab enterprise edition is non-free, but self-hosted gitlab community and gitlab for open source are free. Currently.
Personally, I tend to agree with excluding apps on main gitlab dot com because of difficulty using gitlab dot com over Tor; intentionally broken search when not logged in; not easily knowing what kind of account (EE or Community) a developer uses, AFAIK; similarity with other sites like MetaBook in terms of privacy and data collection (or too complex Terms and Privacy policy). It would also reduce the number of apps to list here by another approximately 350, and get it down to a reasonably small number for a list. I was trying to not bias this list too much by my personal feelings, but I can easily go along excluding main gitlab dot com apps due to overwhelming anti-reasons over-riding the technical free license applying to selected circumstances… How’s that for semi-arbitrary.
Don’t forget site searches, on whatever search works for you: https://html.duckduckgo.com/html?q=site%3Agitlab.com fdroid%2Fadmin%2F-%2Fissues "leave gitlab"
Each word carries thousands of hours of use, of scripts making, of tweaks, of ideas. Again, do treat each of my answer as an attack, it will lead to a good place.
I think F-Droid Client doesn’t need that much CI time, but especially F-Droid Data does. Hosting this would probably cost a lot and take a lot of time to maintain.
As seen in this growing list, dozens or hundreds of apps are independently developed on F/LOSS network systems, by developers with limited resources. F-Droid has enough resources to pay some of their developers, and gets deals and support from CI runners, cloudflare, etc. But does not have the resources to set a good example by using F/LOSS net services, with just F-Droid app, or more?
Moving only one part of F-Droid would really complicate group membership and access rights and stuff and increase maintenance burden, so that sounds pretty awful. I’d personally rather see us move completely or not at all.
With the couple issues linked by you and LK, it looks like F-Droid is being pushed along by whims of others (slowly), more than being proactive. It’s good to see a little inclusion of values in the discussion; would be better to see that be the driver.
All or nothing
Perfect is the enemy of good enough, they say. One step, in the right direction, is a good start.
