Remove dependency on Google Play Services

[png5lyfe]

MonTransit cannot load the map view without google play services installed. This means a large portion of the app's functionality is inaccessible. Can the app be built with a different map view, like OSM?

[joelmaxuel]

Came here to post the same issue; I will +1 this one instead.
FWIW, it can even be a scaled back (e.g. map-free) edition for F-Droid.
This app would stand out for being the only GMS-independent app using real-time GTFS data if this can be achieved.

[mmathieum]

The MonTransit app mtransit-for-android-gradle currently depends on Google Play Services for:

• installing/updating security provider in commons-android => pure-AOSP flavor would not have this feature
• showing POIs on a map in mtransit-for-android => pure-AOSP flavor could have no map OR a different map SDK (OSM...?)
• finding user's device location in mtransit-for-android => pure-AOSP flavor could use the Android SDK Location classes like old times.
• tracking app usage in mtransit-for-android => pure-AOSP flavor could not track app usage(?)
• detect crashes (fatal & non-fatal) in mtransit-for-android => pure-AOSP flavor could not detect bugs in production
• searching for nearby addresses/places in mtransit-for-android => pure-AOSP flavor could not have this nice-to-have feature
• monetize app usage with ads in mtransit-for-android => pure-AOSP flavor could not monetize or use another ad provider
• monetize app usage with subscription in mtransit-for-android => pure-AOSP flavor could not monetize or use PayPal, Patreon...?

[joelmaxuel]

Quite a lot of tie-ins.
My understanding, then, is that you have no interest in a GSF-independent version/edition?

[mmathieum]

It's not that have no interest in a GoogleServiceFramework-independent flavor, it's more about prioritizing this over other features that I think more users want.
But this project being open-source, I would take the time to review any Pull Request containing such feature and merge it once it's ready.

[joelmaxuel]

That is reasonable. I'll do a fork shortly (the android app, commons, and gradle repos) and take a look, play, what-have-you. Given the circumstance of work to be done, it would more likely be better served as it's own branch so I will start one of those.

It will very likely be a slow trudge for me to go through it, however I am that interested in the outcome. When (or even if) I manage something to send back as a PR, we can discuss future logistics of that branch (I am open to continue its maintenance but I don't want to be intrusive enough for me to submit to F-Droid - though I would like to see it there when it comes time). Feel free to close this issue for now, and thank you for the above guide. Cheers.

[joelmaxuel]

Hi @mmathieum ,
Due to RL, I have changed the scope of what is important, to have core functionality without bothersome views that have no impact/response/effect when GMS/Nlp is not present.
What may end up bizarre is that even tough I have maintained half a dozen or so devices for custom firmware, compiling an APK is still a mystery for me.

Based on the Android documentation, the commit in the following link should do what the revised scope became. Please let me know if this is helpful.
joelmaxuel@122fb98

As part of the revised scope, I do not believe F-Droid would be necessary (probably a lot of extra work as well if to consider the related GTFS APKs). Hope this helps.

[TPS]

@mmathieum Due to the GPlay shenanigans you wrote about, are you revisiting this? F-Droid would probably welcome you, if this were accomplished.

[joelmaxuel]

Based on what was written before, there's probably a greater amount of work to rework as FOSS and be reasonable for F-Droid than reinstate on GPlay.
This is unfortunate news because even with using it without GSF as-is, it remains more complete and easier to use than GTFSOffline.

[TPS]

From what I've read in similar situations, even if reinstated, it may be temporary. Why not work both tracks simultaneously? #9 & https://gitlab.com/fdroid/rfp/issues/1081 could serve as meta-issues to track work toward inclusion.

[joelmaxuel]

Considering the F-Droid option is being started without having to make a case, working in parallel makes that easier. Given I don't have much spare time, I try not to expect the time of others. If people are offering help to achieve what they want to see, I don't have a reason to knock that.