Ok wafrn people, it's story time. A tale of the tech challenges I have been going through in the last weeks.
TL:DR I messed up when submitting my app to IzzyOnDroid. Now I have removed all google / firebase dependencies from the apks, built and published npm package to use Unified Push in modern Expo / React Native and found a way to reduce the app size to be 3x smaller by creating an apk for each architecture
So… A few weeks ago @IzzyOnDroid told a few tips about how to make my app (the wafrn mobile app) better for publishing on android alternative stores
Main pain point were the use of propietary services like Google Services and Firebase Cloud Messaging and also the apk size.
I was told about Unified Push, a library, a FOSS alternative to Google and Firebase services for sending notifications to android devices. After scraping the bottom of the barrel on github and npm found that there was not a person in earth who had tried and succeeded to integrate the Unified Push library with modern versions of Expo (using React Native new architecture), so that was my first challenge.
I had to create and publish a library so people could call the Unified Push Android library from their JS code in React Native. I had to learn modern android development with Kotlin. I had to set up a whole android development environment with all the Gradle stuff, the Java stuff and such.
I tried various VS Code plugins and Fleet from Jetbrians but in the end, the only one that worked well for me and could run my project was Android Studio.
After reading all the documentation for using unified push with Kotlin, I designed how the API surface of the library was going to look like. I wanted to make this abstraction the easiest possible one.
But there were more things to discover. To receive events from unified push in my JS code, I had to learn how to use the internal expo native APIs for android, so I read a lot of kotlin / java code written by the expo team.
After publishing my first version of the expo-unified-push integration I had to rollback to a previous attempt of bridging UP messages to JS because it broke hard with more complex JS code and later publish a third version because I did not get what an Android Intent was in the other previous two
Later came adapting the backend, which was not that hard, but I ran into some rocks there too.
With all this completed, and google / firebase dependencies removed my app was still heavy, at 150 MB.
So, with my newly found knowledge of being a native android developer, I start to tinker with graddle settings and graddle-expo integrations and I found that I was building an app for 4 architectures inside of one. When I split the apks, each to its own architecture, they now weighted between 40mb and 50mb
and I should add it to my resource list as well as soon as i'm back at my computer, so I can point others to it. Thanks for having accepted that challenge! 🤩
This is a non-exhaustive list of the first few end-user applications that use UnifiedPush.
Android Name Description Date Since version Docs¹ FluffyChat Matrix chat January 2021 v0.26.
@unifiedpush@fosstodon.orgCan you please give me a ping with the link once it's up, so I can point developers-in-need to it? Thanks in advance! @javascript
Android Official libraries 3rd parties libraries Linux Official libraries If want to support UnifiedPush to receive notifications using an external distributor.
Sure, thanks! I think that each time I visit it, but then there's so much to do… I also think to maybe move that to our wiki as well, now that comments seem to be possible there as well. But then, that's one of the last repos left on GitLab and (hopefully soon) to be moved to Codeberg, and there are no comments… 🤷♂️ @javascript
## Push<br><br>There are 3 options to be notified of new events:<br>- Keeping a constant connection to the server<br>- Periodically fetching notifications<br>- Supporting push notifications (UnifiedPush + optionally fallback to Google notifications)<br><br>Depending on your application, it may makes sense to support several options. For example, for a social application, some users may want to be informed at most every hour, and others "in real time".<br><br>#### Keeping a connection to the server<br><br>[Quoting Jeroen](https://github.com/Adamant-im/adamant-android/issues/3#issuecomment-553585062):<br><br>> There's multiple options for getting push\*, there's xmpp which works for push\*, an MQTT system works, Server Sent Events is an Http standard for this, and websockets is also often used for this.<br>Tutanota and mastodon seem to like SSE, while telegram and signal use websockets in their FOSS environment (but use google whenever possible)<br><br>> Conversations / Kontalk just uses the xmpp protocol it uses for chat for push notifications\* as it's a realtime protocol anyway<br><br>\* Here, _push_ is used to refer to solutions that allow you to maintain a connection with the server to find out when a new event occurs. This is generally not what push notifications are about.<br><br>Keeping a connection requires implementation of the service, a means of restarting it in the event of failure, network callbacks, etc. This will also increase your application's battery and data consumption. This solution is suitable for applications that require “real-time” events, such as instant messaging applications, for everyday users.<br><br>#### Periodically fetch new messages<br><br>If it's acceptable to have a certain delay before being notified, a periodic task works well. For example, this is how F-Droid checks whether updates are available.<br><br>This solution may also be a good solution for users that rarely use an application and are OK to get non-instant notifications.<br><br>#### Push notifications<br><br>[UnifiedPush](https://unifiedpush.org) is an open standard for push notifications that rely on web push. It allows having efficient "real-time" notifications:<br>- The application will use the service the user has installed<br>- One service keeps a connection for all other applications<br>- You don't have to implement a constant connection to the server<br><br>Web push is another open standard, defined by 3 RFC, commonly used for web applications. Your server may already support it.<br><br>Furthermore, as explained in [the blog post about push for decentralized apps](https://unifiedpush.org/news/20250131_push_for_decentralized/), it is possible to get a fallback to Google Services, that works:<br>- with a FOSS library<br>- without any gateway (the server sends web push requests to Google)<br><br>For some Background on UnifiedPush, see:<br>- [UnifiedPush: a decentralized, open-source push notification protocol](https://unifiedpush.org/news/20221218_unifiedpush/) (12/2022 edited on 04/2025) originally posted [on F-Droid](https://f-droid.org/2022/12/18/unifiedpush.html) (12/2022)<br>- [Push notifications, security and privacy ](https://unifiedpush.org/news/20250513_push_security_privacy/) (05/2025)<br>- [Dear Privacy Aware Android App Devs - Please Use UnifiedPush](https://www.andrewjvpowell.com/articles/dear-privacy-aware-android-app-devs-please-use-unifiedpush/) (10/2021)<br><br>This solution works well for everyday users, and for occasional users<br>
I've dropped support for Gotify-UP (the fork we started UnifiedPush with) and NoProvider2Push a while ago. I think it is better to remove them from the page. Maybe you can list Sunup instead, which is intended to be the goto solution.
PS: Is it possible to correct the heading from Unified Push to UnifiedPush ?
Sunup is listed there, and the Gotify Distributor for UP (Gotify-UP) was removed from the list long ago (Gotify itself is still maintained, so their non-UP client is listed in the "others" section – with "others" meaning "non-UP").
I've seen that NoProvider2Push was not only marked experimental, but also has not received any commit for years. The Last-Update date reflects that. I'd rather leave it in as long as it's available, maybe someone would "take over" development…
Thanks, updated! Maybe make that clear in the Readme as well? It only shows in the latest commit, and could be interpreted as "that specific version is deprecated", not that library 😉
Rock band REM releasing a remix of its first single, “Radio Free Europe,” to raise money for RFE/RL after the Trump administration froze its funding. “This was a little thing that we could do to support freedom of expression and particularly great journalism," says Michael Stipe. instagram.com/reel/DJJluTOO9Fj…
23K likes, 546 comments - cbsmornings on May 2, 2025: "The iconic rock band @rem is releasing a remix of their first single, “Radio Free Europe,” to raise money for @rfe.rl after the Trump administration froze its funding.
Our trip from #Brno to #Prague and back (420 km) cost 300 CZK (€12) with #EV. A family (2+2) train ticket would cost 2000 CZK (€80). Sadly a no-brainer from a family budget perspective.
IMO individual transport with all its negative externalities shouldn't be so cheap compared to public transport.
@vadim yes, if you don't have a car and are deciding whether to purchase one or rely on the public transport, it makes sense to count total cost of ownership of the car. But if you already have one and you're only deciding if you want to leave it home and go by train or not, it makes sense to only count costs incurred by that particular ride, everything else is a sunk cost. IMO public transport should be attractive for families with cars, too.
I'm always astounded by this exact fact. How can it be cheaper for me to get to where I want to go in a metal box with 4 people compared to a string of metal boxes with hundreds of people? Can someone with better understanding if economics and maybe the railroad system explain this?
@miroslavbucek TCO is definitely much more than 300 CZK, but once you already have the car it's a sunk cost which shouldn't be considered when you're deciding whether you go by car or leave it home. And if you count it, it actually motivates you to use the car even more because the more you drive it the lower the cost of ownership per km is.
If the car cost one million CZK and you don’t plan to drive one million kilometers—more likely around 250,000—then you need to add at least 4 CZK per kilometer to your costs.
@miroslavbucek Sure, but what I had in mind is a situation where a family already has a car (because most of them do) and they're deciding whether they want to go by car or use public transport. For that decision making it isn't longer relevant what they paid for the car beucase it's already been paid, a sunk cost. The decision won't increase or decrease the cost. I know it isn't a fair comparison because the family can omit it as a sunk cost while the tickets have to cover amortization, but it's a decision families face all the time and if the public transport costs 7x more it will push the families into relying on their cars even more and increase individual transport with all its negative externalities.
You have a massive train that is fully utilised only a few hours a day at best. It cannot be cheaper than fully occupied car designed for efficient ride.
Btw: I will ask you six months from now how many times you sit into the other car voluntarily 🙂
From the city. this may not be so obvious, but if im on a village, with close to zero public transport connection, having a car or two is just a must and calculating amortization to each ride does not seem to even the costs. Car makes it possible to get there AND get back.
With family of 5 + dog, any car is always cheaper then any public transport and usually it is the only option anyway
beyond the comparison of operational, ammortization and subsudy impact, there is tje not so direct overhead of R&D between rail & EV R&D: eg BYD has currently ~48.000 EV patents & 100.000 R&D FTEs
this hidden R&D cost + the positive impact of EVs usi.g existing electricity grid propably does not allow for like for like comparison between eventual costs we pay as users
if we're gonna be stuck with vibe coding, might as well try to have some fun with it
("anti-vibe coding" would be my standard schtick: stuff like invariant-based programming, assertions, testing, fuzzing, CI -- all the stuff we do to try to make software actually work)
So, looks like Mr Dutton called the misfortune upon himself! :) He wanted to be Australia's PM. But a 'Trump effect' thwarted Peter Dutton - BBC News apple.news/ATyBbj4iJTOG6Bk2Jlx…
Technisch: Das ist ein generischer Fehler. HTTP Status code 403 ist hier definiert: datatracker.ietf.org/doc/html/… Woher der Fehler kommt ist damit nicht gesagt. Ich vermute(!), daß eine erklärende Fehlerbeschreibung gar nicht so einfach zu realiseren ist.
I decided to try the new Suno 4.5 model which is new at the time of recording.Feeding it one of my own creations in the form of a piece of music called '23A-...
I'll take a look at this, but have you heard what Google has been doing with AI music recently? They have a new product called Music AI Sandbox, and have been testing it over the last year with different artists. Here's Jacob Collier playing with a slightly different take on Sandbox called MusicFX DJ. youtube.com/watch?v=GmtPA1jZCR… Unlike Suno or Udio, these don't create hole songs for you, but instead allow you to get little ideas that you can then sample into your own tracks. Sandbox is also really interesting because you can import your own audio, kinda like Suno, then transform it in a number of ways. Here's a link to learn more about Sandbox, you can sign up to give it a try, will be doing this as well and report back. deepmind.google/discover/blog/…
Google has long collaborated with musicians, producers, and artists in the research and development of music AI tools. Ever since launching the Magenta project, in 2016, we’ve been exploring how...
Ah strange, I just saw they added a sign up page a few days ago, so mabey try again with the link I posted. Most people who wanted to try it had to contact Google before IIRC.
Ok so just reading through the article I sent, looks like it's US only for now, but I'll sign up and will post lots of demos. If you have anything you want me to try it with just let me know. I'm sure access will expand to more countrys sooner rather than later.
Yeah I'm really not sure why they do things like this, makes no sense to me. Something else you can try though, Google has a cool project called Magenta that uses machine learning to generate Midi. You can input a melody for example, then get Magenta to continue it for you. It runs offline on your machine, and all the apps are fully accessible.
Here's the main getting started page. Give Magenta Studio a try first. They have plug-ins designed to work in Ableton, however the version you want is standalone. Just note that they don't seem to support the standalone versions anymore, but you can still download them. magenta.withgoogle.com/get-sta…
you sons of a bitch you try to report me and delte my server. you should be ashamed of yourself. azure employee will be reported, you are logged as creating an issue.
You know my little self was not going to comment on this, but since a human with 0 compassion just spoiled the fun conversation about AI music, I must say, REPORT! Report is all in caps, by the way. Get a life, kindly asking you to please do so. Your unwanted, unnecessary reply had nothing to do with the posts I was reading, now you spoiled it, and the two of me are not very pleased. Also, these dudes are cool, they are acquaintances, and no one talks nasty to my friends, especially not humans lacking manners and compassion. If you have nothing nice to say, keep your silence, thank you.
@ZBennoui I tried it (google sandbox)and found it pretty bad. Suno has generated two versions of every prompt I've tried in 20 seconds with full lyrics that rhyme and are coherent. -I would NOT do this, but I think for $8, you generate the max (500 tracks in one month), put 'em out in 50 albums with non-AI graphic covers and probably make money.
@randulo Ai Sandbox is not for people that like Suno, it's meant to be used to get ideas for your own music that you produce in a DAW. I've signed up and am waiting to be invited to try it, but am excited to see what it can do.
@ZBennoui Gosh, Jacob makes it sound so easy, but I do wonder what the return on time investment really is given that the final product was something that Jacob could probably put together without all the AI involved. No doubt, interesting stuff though.
You must, try, try, try. LOL. See? I@m such a good encouragement giver. I wonder if there's a music AI that would beat or compare to Suno, because I have not really heard others.
@menelion @tardis I haven't tried them for a long, long time but I didn't like their site that much, certainly when I last looked. Also I'm a paid member of Suno so I won't be looking elsewhere for now lol
Okay now this, was pretty cool. You can notice harder it's an AI, but it's noticeable. You also gave it a track of yours, but seriously, it was quite impressive. Especially that piano solo it did, was pretty cool, would imagine quite hard to learn as well. LOL. All in all, good job you, and credits to the AI for experimenting with your good job. Haha.
What happens when one of history’s greatest composers begins to lose the very sense he relies on most? In this episode, we explore how Ludwig van Beethoven continued to create groundbreaking music even as his world fell into silence. Along the way, we uncover the myths, inventions, and raw determination that fueled Beethoven’s defiant creativity, and hear how his lifelong struggles are reflected in his music. Featuring musicologist Laura Tunbridge , author of Beethoven: A Life in Nine Pieces. Enter the “Sound Off” Story Contest at 20k.org/soundoff. Submissions close on May 7th, 2025. Explore the all new Defacto Sound website, and click the Contact Form to get in touch. If you know what this week's mystery sound is, tell us at mystery.20k.org. Follow Dallas on Instagram, TikTok, YouTube and LinkedIn. Join our community on Reddit and follow us on Facebook. Start your free online visit for hair loss treatment at hims.com/20k. Cut your current cloud bill in half with OCI at oracle.com/20k. Episode transcript, music, and credits can be found here: www.20k.org/episodes/the-deaf-composer Learn more about your ad choices. Visit megaphone.fm/adchoices 20k.org/episodes/the-deaf-comp…
Tomorrow Mom and I are headed back to the clinic in Hollywood for my 3 month checkup. Could I please ask people to pass this around again and donate if you're able? Thanks, and thank you so much to those who've already given! #Mutual_Aid
My name is Cristina Kalmanson. Several years ago, my cousin, Kath… Cristina Kalmanson needs your support for We Have a Transplant! Now let’s help the Burkes!
Professor Stephen Whittle OBE, famed UK Trans Legal activist from Press for Change, probably the most reputable trans person in the UK in regard to the law and affecting it, has put out the action plan guide stating that the best chance of influencing things is for trans people and allys to each write physical letters to their MP and 6 copies to specific other members of parliament.
He's provided the relevant facts to be included and laid out what you need to write.
I'm in the UK so shame that's not gonna be available, would save me and my mum the effort of going to Iceland Foods to get the stuff and then carting it back to the car and then getting it up to my apartment lmao.
Before prevalent social media, a static webpage and a few mailing lists could shake a brand.
My 1999 #Swatch #Protest archive shows how amateur radio operators used hand-coded HTML, listserv chains, and Usenet newsgroups to stop Swatch’s #satellite from advertising on their radio frequencies.
In early 1999 Swatch set out to publicize its new “Swatch .beat Internet Time” by launching Beatnik, a grapefruit‑size “mini‑Sputnik” satellite from the Mir space station.
John Calvin, early Calvinist preacher, writing on collections for famine (2 Cor 8), says Jesus “consecrated” the estate of poverty by living in it. Thus: 1) We need not regard this estate with horror. 2) Free from fear of own poverty, may help those in distress.
Today, do we fear the very presence of people in distress, and thus cut people off from any tie of brotherhood?
In this relaxed May edition of At Your Fingertips: Braille Then and Now, we explore fun, low- to mid-tech tools that bring joy and progress to Braille learners.
We all think it can't happen to us, but it can when we least expect it... This reader says his Apple Watch saved his life—make sure yours is set up too 9to5mac.com/2025/05/02/this-re…
Webinar From TPGI: Marking a Milestone: Celebrating 30 Years of JAWS on Global Accessibility Awareness Day; Wednesday, May 14, 4:00 PMEastern Time tpgi.com/webinar-marking-a-mil…
How to block ads in Gmail is getting more complex as new Gmail advertisement is more intrusive than ever, in your mailbox, in your Gmail app, everywhere. But we've got a solution!
Sorry but people who are too lazy to switch from gmail have what they deserve. The whole world now knows that Google is a toxic company but they still don't change.
Uncover the untold story of Apple VoiceOver! Join us as we reveal fascinating insights and genius ideas behind this groundbreaking technology. Despite valid criticisms, we acknowledge Apple's transformative impact on the industry. Discover the real story today! youtube.com/watch?v=5tLSycbey6…
What do the Clicks Keyboard, Ember Mug, Monarch braille display, and AirPods Pro 2 all have in common? They’re all in the inbox this week on Double Tap!Steve...
How do we translate web accessibility standards to command line applications? This is GitHub CLI's journey toward making terminal experiences for all developers.
Updated Blog Post From Hartgen Consultancy: Stay in the Game with Leasey Sports and Games: Your Accessible Sports Companion! groups.io/g/tech-vi/message/91…
Webinar, May 14 at 4pm ET: Marking a Milestone: Celebrating 30 Years of JAWS on Global Accessibility Awareness Day - TPGi tpgi.com/webinar-marking-a-mil…
Mastodon has taken the strategic decision not to accept venture capital investments for growth, but rather restructure to a European non-profit organisation. 👏
![JOHN CALVIN (1509-1564) Calvin's Commentaries 2 Cor 8 "Hence he [Christ] has consecrated poverty in his own person, that believers may no longer regard it with horror. By his poverty he has enriched us all for this purpose -that we may not feel it hard to take from our abundance what we may lay out upon our brethren."](https://fedi.ml/photo/preview/640/664087 "JOHN CALVIN (1509-1564) Calvin's Commentaries 2 Cor 8 \"Hence he [Christ] has consecrated poverty in his own person, that believers may no longer regard it with horror. By his poverty he has enriched us all for this purpose -that we may not feel it hard to take from our abundance what we may lay out upon our brethren.\"")
IzzyOnDroid ✅
in reply to Bolet :neobread_eyes: :yeah_waffy: • • •Bolet :neobread_eyes: :yeah_waffy:
in reply to IzzyOnDroid ✅ • • •IzzyOnDroid ✅
in reply to Bolet :neobread_eyes: :yeah_waffy: • • •UnifiedPush
in reply to Bolet :neobread_eyes: :yeah_waffy: • • •Great ! A page dedicated to 3rd party libraries will be added in the coming week
BTW, you can open a pull request to add wafrn to [the list of applications](unifiedpush.org/users/apps/) supporting UP !
PS: we use to use UnifiedPush in one word, if you want to update the Readme
Apps using UnifiedPush
UnifiedPushIzzyOnDroid ✅
in reply to UnifiedPush • • •UnifiedPush
in reply to UnifiedPush • • •Implementations
UnifiedPushIzzyOnDroid ✅
in reply to UnifiedPush • • •F-Droid ($1896503) · Snippets · GitLab
GitLabS1m
in reply to IzzyOnDroid ✅ • • •IzzyOnDroid ✅
in reply to S1m • • •S1m
in reply to IzzyOnDroid ✅ • • •IzzyOnDroid ✅
in reply to S1m • • •Push Notifications · Wiki · IzzyOnDroid / repo · GitLab
GitLabS1m
in reply to IzzyOnDroid ✅ • • •Thanks :)
About: android.izzysoft.de/applists/c…
I've dropped support for Gotify-UP (the fork we started UnifiedPush with) and NoProvider2Push a while ago. I think it is better to remove them from the page. Maybe you can list Sunup instead, which is intended to be the goto solution.
PS: Is it possible to correct the heading from Unified Push to UnifiedPush ?
Push Clients/Manager - Android App Übersicht
IzzyOnDroidIzzyOnDroid ✅
in reply to S1m • • •Sunup is listed there, and the Gotify Distributor for UP (Gotify-UP) was removed from the list long ago (Gotify itself is still maintained, so their non-UP client is listed in the "others" section – with "others" meaning "non-UP").
I've seen that NoProvider2Push was not only marked experimental, but also has not received any commit for years. The Last-Update date reflects that. I'd rather leave it in as long as it's available, maybe someone would "take over" development…
And I thank you!
S1m
in reply to IzzyOnDroid ✅ • • •IzzyOnDroid ✅
in reply to S1m • • •S1m
in reply to IzzyOnDroid ✅ • • •Oh: I've just seen you linked the FOSS library for the embedded distrib to codeberg.org/UnifiedPush/andro…
But it has been deprecated in favor of the main one, which is fully foss: codeberg.org/UnifiedPush/andro…
android-embedded_fcm_distributor
Codeberg.orgIzzyOnDroid ✅
in reply to S1m • • •S1m
in reply to IzzyOnDroid ✅ • • •