Repository navigation
Replies: 2 comments 1 reply
|
@trypsynth Maybe it's easier to start by copying Client/TeamTalkAndroid to Client/TeamTalkAndroid2 (or other name) and start working on it there. That way it can be developed on master-branch and we can still maintain TeamTalkAndroid until the new client is ready. |
1 reply
|
The best path for now is to keep this client as java, until later converted to qt. I onnistly don't like kotlin and I always choos java for android apps. |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
@bear101 I would like to ask if you would accept a gradual move of TeamTalkAndroid to Jetpack Compose. I am willing to do the work and to maintain the client afterward. I am asking before I write any code.
This follows two earlier discussions that have both gone quiet: #3357 on Flutter clients, where you said the Android client needs a makeover, and #3436, where I asked if Kotlin is welcome in this module. Neither reached a decision, so the Android client still has no agreed direction. Compose requires Kotlin, so an answer here also answers #3436.
Why:
MainActivity.javais about 2,600 lines and mixes UI, state and event handling. That is where most of the review cost and many of the bugs are.How I would do it:
Library/TeamTalkJNIstays as it is, in Java. Only the UI inTeamTalkAndroidchanges.TextMessageActivity, after Fix action bar in text messages activity and add copy all and clear #3379 is settled. If the pilot is not as good as the current screen with TalkBack, we stop and only that PR is lost.Risks:
Viewwidgets. It exposes its own semantics tree, so a regression in Compose can affect every screen. I raised this same concern about Flutter in Teamtalk mobile clients on flutter #3357, and it applies here too. Paperback shows me that Compose can be fully accessible, but TeamTalk has its own needs. I use TalkBack daily and would test every screen with it before opening a PR, and I still want a pilot first.If the answer is no, that is fine, and it would still help to write it down so that contributors know the direction. If you prefer Kotlin with the existing XML layouts, or the Flutter client, I would like to know that too.
All reactions