Repository navigation
Teamtalk mobile clients on flutter #3357
Replies: 50 comments 6 replies
|
I heard that Qt can be extended to Android; wouldn't that be a better option? I apologize if I am not, as I haven't been able to look into that aspect yet. |
|
I personally prefer looking to port Qt client to Android instead of use a new language |
|
Frankly speaking, i don't know mutch about qt on mobile platforms and don't have intention to dig into it. I will finish android client and release an evaluation version. I understand that it is risky to go to another language, but this has a benefit to cover both mobile platforms. It is too mutch effort needed to make current android client to be refactored rather than written from ground up. We will see how it will be going on. |
|
Note that Flutter supports not only mobile platforms. It actually covers
all the platforms TeamTalk is now developed for.
|
|
@beqabeqa473 i would also like to try this client, can you provide a way? |
|
Hello all. Here is a first version. https://github.com/gozaltech/TeamTalk5-flutter/releases/download/v1.0/app-arm64-v8a-debug.apk Sources will be available soon. This is for evaluation purposes. |
|
There are two ways of doing things everytime, neither from popup menu, or from actions. I've implemented almost all administration features. it still may lack some features, whjile writing it, i remembered broadcast messages, video capture, desktop-sharing view, those are not available in new client. Note to push-to-talk button, now it can be locked by dragging up, as messengers are doing alongside custom accessibility action. I've implemented microphone selection on android. It is mostly feature complete, i've tried to cover all test scenarios during development, but there still may be bugs. I am still ready to contribute this client to bearware, but in case this will not happen, i will upload this client on behalf of me. I think i was able to complete Hackathon to complete as much as possible in a week of development :) Actually, my goal was to learn dart/flutter, and i chose teamtalk as a goal point to achieve, as nowadays flutter is increasing popularity and there are so many job vacancies to be applied. So, that is what it is at the end. |
changed events setting screen a bit: https://github.com/gozaltech/TeamTalk5-flutter/releases/download/v1.1/app-arm64-v8a-debug.apk |
|
@bear101 can you please look at visual ui when you will have a time? |
|
@beqabeqa473 hi. |
|
I understand your point, but i have strong feeling, that teamtalk-android in the current form is non-maintanable product. it has very large amount of technical debt, which can be fixed only by rewriting. And again, i am not a person who is taking decisions. I am going to finish stuff which is in my backlog for this client and open-source it. You have 2.5k LOC in MainActivity, it is not something anyone will want to deal with. |
|
@beqabeqa473 hi |
|
if there is talking about rewriting teamtalk mobile clients between flutter and qt, I would absolutely choos qt because |
|
First of all, i am not going to do flutter client for desktop untill accessibility of flutter will not be fixed at least to level of QT. |
|
@beqabeqa473 yes. Qt has much better accessibility. So I think qt is only good and pure c++ on windows. and on android even if we are using qt, we are not as direct as java. |
|
but it's only 2 new options to settings that allows user comftability, And it allows all channels expantion and the best way you'd want to reach this. |
|
I think you didn't get properly what I mean. I don't think what I explaned is complicated. i'll explane it clearly. |
|
onnistly I very liked this channel tree desighn here. I wanted to implement it for Current java android client in 5 minutes but All of my pull requests are being merged after years, so no point, even If I wanted to clean up the entire teamtalk android in a week. 😔 |
|
Hi. v1.3 Channel user counts now show x/y for channels that have sub-channels users in the channel over users in the channel and all its sub-channels (so a collapsed parent or the root reveals its whole subtree). Leaf channels show a plain count. download from: |
|
@beqabeqa473 If you didn't like my suggestion regarding multiple channel expantions, You can just remember what channel User has expanded last time instead of collapsing other channels and show the button in the app bar. So If I expand channel audio box, the app bar should show join audio box button, when I also expand gaming channel, it should show join gaming channel button in the app bar as it does now, but don't collaps the audio box channel. |
|
It is not intuitive. and explicyt. We need to find other way. |
|
ok, it's up to you. Even though your client is currently unusable for us here, I still downloaded and providing the feedback that I can, even if I never use it. That's a correct human fealing. Helping is always good. |
|
v1.5 Added noise suppression. Download from you can discuss bugs and new features in telegram group at: |
|
@beqabeqa473 I only have an Samsung Galaxy A10 with Android 11. It looks like I cannot install the app on that device. Can you compile for x86_64, so I can try it in the simulator? |
|
Chiming in here to say that I hope that the flutter client doesn't replace TeamTalk Android, it's fine if it shows up as another option, but as a blind user Flutter apps are horrible, especially on Android. I am willing to help maintain the native TeamTalk android client if necessary, I really prefer native Android apps and would hate to see TeamTalk majorly regress. |
|
@beqabeqa473 Fair ask, so here's the technical version. You're right that Flutter ends up feeding Flutter doesn't use any of those views. It paints to a canvas and exposes a virtual tree through its own
For a chat-and-voice client specifically, the text field and the list are the whole app. Those are exactly the two places where Flutter's reimplementation has the thinnest track record. On the Persian issue in this thread: eSpeak reads Persian channel names fine in the current Java client and in every other native app. If the same synth produces garbage in the Flutter client, that points at Flutter's text and semantics pipeline (which has known RTL gaps), not at eSpeak. Worth checking before writing it off. None of this means your client is unusable; clearly it works for you daily, and the work you've put in shows. My point is narrower: a Flutter client is a fine additional option, but replacing the native Android client (and yes, I call kotlin/java Android native, not...the NDK), with it trades a platform-maintained accessibility stack for one that a single project maintains, with regressions like the heading one as the cost. I'd rather help keep the native client alive alongside yours than see it go. I'm serious about the offer to help maintain it. |
|
@beqabeqa473 You're right on the architecture and I'll say so plainly: framework widgets, AndroidX, and Flutter all end up as AccessibilityNodeInfo, AndroidX ships in the APK just like Flutter does, and Compose uses a virtual tree too. I overstated that. |
|
@beqabeqa473 I didn’t "sound like an AI agent" because I yielded an architectural point; I conceded where you had a valid point because that’s how adults have technical discussions. My initial comment came from my actual, lived experience using Flutter apps with TalkBack every single day. I listed many of them for you, in fact. When an end-user navigates an app, they experience whether focus lands where it should and whether the cursor tracks text accurately. In practice, across dozens of Flutter applications on Android, that experience has repeatedly fallen short compared to platform-native components. Digging into the architecture or telling me that I sound like a chatbot for having a technical discussion doesn't change the reality of those bugs. No one is trying to diminish the massive effort you’ve poured into building your Flutter client, and nobody is calling for your work to be discarded. That said, it's been said that you're the only one who really understands Dart/Flutter in this repository, and replacing the existing native Android client outright removes a proven native client that works well and many people already understand. My offer still stands completely open to help maintain and support the native Android client so that burden doesn't fall on you or anyone else, while users who prefer your Flutter client are free to use it. And, to be clear, I wouldn't care if it shipped in the repo, as I have repeatedly said. I made a statement based on my lived experience with Flutter apps, and you've taken it incredibly personally it seems. This is a technical discussion, no one is attacking you, and I admitted where you had valid points, only to be called an AI agent. Please try to keep this productive and don't resort to baseless personal attacks based on someone's lived experiences that they maybe didn't express in the most constructive way. |
|
@trypsynth hi. |
|
Hi all I finally open-sourced TeamTalkX client. Repo is at @bear101 apk for x86-64 is in releases. Thanks. |
Uh oh!
There was an error while loading. Please reload this page.
Hello all.
As a contributor of teamtalk, I am getting constant complaints of gaps in especially teamtalk android.
As we all know, this codebase is from 2015 and is tried to be maintained till now, but fixed longstanding bugs require too mutch time and effort.
I am in process of writing a new client for mobile platforms in dart using flutter framework, firstly for android and then extending it to iOS.
i am willing to contribute this client to bearware to lift up burden on maintainers to support two independent clients, so we will need to maintain only two of them, qtTeamtalk and teamtalkFlutter.
@bear101 i would like to hear opinion from you.
regardless of decision, i will finish client and release, but i would like to not fragment clients, so if possible i would like to continue it as part of teamtalk5 project.
As one of contributors of teamtalkAndroid, i am tired to fight with this legacy code from time to time.
It is becoming harder and harder to maintain project release by release.
Of course, someone can contribute changes using nowadays popular AI tools, but this doesn't improve situation if not degrades it.
Thanks for understanding.
All reactions