Johan Vos
@johanvos.bsky.social
Java Champion, Gluon and LodgON.
co-lead OpenJFX, lead OpenJDK-Mobile . Java for science, Quantum Computing. PhD
co-lead OpenJFX, lead OpenJDK-Mobile . Java for science, Quantum Computing. PhD
Working on OpenJDK for Android: in 43 seconds, I run an app on my phone, make a change in OpenJDK (change dlllookup into dll_lookup), rebuild OpenJDK, link it with my app, push to my phone, run again.
OpenJDK development is extremely efficient.
#java #innovation #mobile #openjdk
OpenJDK development is extremely efficient.
#java #innovation #mobile #openjdk
October 25, 2025 at 9:32 AM
Working on OpenJDK for Android: in 43 seconds, I run an app on my phone, make a change in OpenJDK (change dlllookup into dll_lookup), rebuild OpenJDK, link it with my app, push to my phone, run again.
OpenJDK development is extremely efficient.
#java #innovation #mobile #openjdk
OpenJDK development is extremely efficient.
#java #innovation #mobile #openjdk
The openjdk-mobile project (github.com/openjdk-mobile) provides daily OpenJDK builds for iOS, packaged as XCFrameworks. Recently, there was a failure because I had to update the bootjdk to 25. Apart from that, very smooth!
#java #innovation #openjdk
#java #innovation #openjdk
October 25, 2025 at 9:08 AM
The openjdk-mobile project (github.com/openjdk-mobile) provides daily OpenJDK builds for iOS, packaged as XCFrameworks. Recently, there was a failure because I had to update the bootjdk to 25. Apart from that, very smooth!
#java #innovation #openjdk
#java #innovation #openjdk
Devoxx registration: what they promised versus what you got. (but hey, it was a great conference, really enjoyed talking to so many interesting folks!)
October 11, 2025 at 8:16 AM
Devoxx registration: what they promised versus what you got. (but hey, it was a great conference, really enjoyed talking to so many interesting folks!)
October 6, 2025 at 2:54 PM
If you're on mac, please try JavaFX 26-ea+6 and share feedback. Tip: add `-Dprism.verbose=true` which will confirm you're really using metal, as shown below. Many thanks to Kevin, Ambarish, Ajit, Jayathirth and others for their great work.
September 9, 2025 at 2:00 PM
If you're on mac, please try JavaFX 26-ea+6 and share feedback. Tip: add `-Dprism.verbose=true` which will confirm you're really using metal, as shown below. Many thanks to Kevin, Ambarish, Ajit, Jayathirth and others for their great work.
See screenshot below from gluonhq.com/products/jav...
We have no devrel/marketing/sales, so I'll say it myself: great job, @gluonhq.com (2/2)
We have no devrel/marketing/sales, so I'll say it myself: great job, @gluonhq.com (2/2)
July 2, 2025 at 4:37 PM
See screenshot below from gluonhq.com/products/jav...
We have no devrel/marketing/sales, so I'll say it myself: great job, @gluonhq.com (2/2)
We have no devrel/marketing/sales, so I'll say it myself: great job, @gluonhq.com (2/2)
Headless JavaFX platform passes almost all tests, with remaining failures being understood.
Time to bring this to the next step: mail.openjdk.org/pipermail/op...
Time to bring this to the next step: mail.openjdk.org/pipermail/op...
June 5, 2025 at 6:56 AM
Headless JavaFX platform passes almost all tests, with remaining failures being understood.
Time to bring this to the next step: mail.openjdk.org/pipermail/op...
Time to bring this to the next step: mail.openjdk.org/pipermail/op...
I know that I have ambitious goals in JDK space: OpenJDK/Mobile +tooling, synced OpenJFX building with OpenJDK , and a Headless platform for OpenJFX.
Progress on the latter: 88% of the tests pass.
Progress on the latter: 88% of the tests pass.
May 28, 2025 at 1:41 PM
I know that I have ambitious goals in JDK space: OpenJDK/Mobile +tooling, synced OpenJFX building with OpenJDK , and a Headless platform for OpenJFX.
Progress on the latter: 88% of the tests pass.
Progress on the latter: 88% of the tests pass.
One of my favorite t-shirts. But also a broken promise by a large company.
I keep believing it's not too late though. Technically, we already have Java running on iOS. But we need more support, technical and marketing. And courage.
I keep believing it's not too late though. Technically, we already have Java running on iOS. But we need more support, technical and marketing. And courage.
May 10, 2025 at 6:31 PM
One of my favorite t-shirts. But also a broken promise by a large company.
I keep believing it's not too late though. Technically, we already have Java running on iOS. But we need more support, technical and marketing. And courage.
I keep believing it's not too late though. Technically, we already have Java running on iOS. But we need more support, technical and marketing. And courage.
Still working on major performance improvements in Strange (QC in Java, github.com/redfx-quantu...).
274/286 tests are passing, but I can only merge once I am at 100%.
I'm glad I wrote many tests when starting this. Need to detect all possible edge cases.
274/286 tests are passing, but I can only merge once I am at 100%.
I'm glad I wrote many tests when starting this. Need to detect all possible edge cases.
May 4, 2025 at 9:46 AM
Still working on major performance improvements in Strange (QC in Java, github.com/redfx-quantu...).
274/286 tests are passing, but I can only merge once I am at 100%.
I'm glad I wrote many tests when starting this. Need to detect all possible edge cases.
274/286 tests are passing, but I can only merge once I am at 100%.
I'm glad I wrote many tests when starting this. Need to detect all possible edge cases.
While on it, I added basic support for macos builds of OpenJDK including OpenJFX (missing javafx.media and javafx.web).
All code changes are in github.com/johanvos/jdk...
All code changes are in github.com/johanvos/jdk...
April 2, 2025 at 7:53 PM
While on it, I added basic support for macos builds of OpenJDK including OpenJFX (missing javafx.media and javafx.web).
All code changes are in github.com/johanvos/jdk...
All code changes are in github.com/johanvos/jdk...
Progress with combined OpenJDK/OpenJFX repo (my OpenJDK fork that contains all my changes is at github.com/johanvos/jdk...).
I did a cross-compiled build for Linux-aarch64.
Here is a JavaFX app running on the Pi:
#java #javafx
I did a cross-compiled build for Linux-aarch64.
Here is a JavaFX app running on the Pi:
#java #javafx
March 21, 2025 at 4:15 PM
Progress with combined OpenJDK/OpenJFX repo (my OpenJDK fork that contains all my changes is at github.com/johanvos/jdk...).
I did a cross-compiled build for Linux-aarch64.
Here is a JavaFX app running on the Pi:
#java #javafx
I did a cross-compiled build for Linux-aarch64.
Here is a JavaFX app running on the Pi:
#java #javafx
Prima opinie stuk in @de-standaard.bsky.social maar aub, stop nu toch eens met mensen aan te moedigen om dingen te delen op Facebook/X. Je voedt het algoritme dat dingen mogelijk maakt die in dit stuk net aangeklaagd worden. Extreme content -> meer clicks -> meer ad revenues.
February 28, 2025 at 10:56 AM
Prima opinie stuk in @de-standaard.bsky.social maar aub, stop nu toch eens met mensen aan te moedigen om dingen te delen op Facebook/X. Je voedt het algoritme dat dingen mogelijk maakt die in dit stuk net aangeklaagd worden. Extreme content -> meer clicks -> meer ad revenues.
Een aanrader: van dezelfde auteur, en met veel prentjes :)
February 18, 2025 at 9:56 AM
Een aanrader: van dezelfde auteur, en met veel prentjes :)
I still see them (until december, but they are always a few weeks behind)?
January 11, 2025 at 1:50 PM
I still see them (until december, but they are always a few weeks behind)?
I wanted to check how good ChatGPT is at some basic quantum computing questions. Let's start with simple math. If n = 2, the number of qubits required for Shor's algorithm is 2n+3.
After wrong steps, ChatGPT gives the correct answer. Because, 2 + 6 + 1 =... 🤔🤣
After wrong steps, ChatGPT gives the correct answer. Because, 2 + 6 + 1 =... 🤔🤣
January 3, 2025 at 4:14 PM
I wanted to check how good ChatGPT is at some basic quantum computing questions. Let's start with simple math. If n = 2, the number of qubits required for Shor's algorithm is 2n+3.
After wrong steps, ChatGPT gives the correct answer. Because, 2 + 6 + 1 =... 🤔🤣
After wrong steps, ChatGPT gives the correct answer. Because, 2 + 6 + 1 =... 🤔🤣
I'm working on StrangeFX, and looking for input from JavaFX developers. What would be your preferred root layout container for this concept? StrangeFX currently use a Group for this, but I am thinking about changing this, to improve embedding capabilities in e.g. JTaccuino.
December 21, 2024 at 9:47 AM
I'm working on StrangeFX, and looking for input from JavaFX developers. What would be your preferred root layout container for this concept? StrangeFX currently use a Group for this, but I am thinking about changing this, to improve embedding capabilities in e.g. JTaccuino.
Today, Manning has 50% off all eBooks, including my book on Quantum Computing with Java:
www.manning.com/books/quantu...
www.manning.com/books/quantu...
December 13, 2024 at 2:12 PM
Today, Manning has 50% off all eBooks, including my book on Quantum Computing with Java:
www.manning.com/books/quantu...
www.manning.com/books/quantu...
There's an interesting issue with the URL in your post. It wasn't clickable, and copy-paste it and retrieve with curl failed:
curl: (3) URL using bad/illegal format or missing URL
Manually typing the URL did work. So I checked the bytes, and it starts with {-30, -128, -86} which is U+202A (LTR)
curl: (3) URL using bad/illegal format or missing URL
Manually typing the URL did work. So I checked the bytes, and it starts with {-30, -128, -86} which is U+202A (LTR)
December 13, 2024 at 7:37 AM
There's an interesting issue with the URL in your post. It wasn't clickable, and copy-paste it and retrieve with curl failed:
curl: (3) URL using bad/illegal format or missing URL
Manually typing the URL did work. So I checked the bytes, and it starts with {-30, -128, -86} which is U+202A (LTR)
curl: (3) URL using bad/illegal format or missing URL
Manually typing the URL did work. So I checked the bytes, and it starts with {-30, -128, -86} which is U+202A (LTR)
(probably unimpressive for most)
Progress. This is from the XCode console, output from an iPhone.
HelloWorld.java on iOS using hotspot (zero interpreter mode). Finally back to the point reached by the great Bob Vandette many years ago.
Need to anchor this.
Progress. This is from the XCode console, output from an iPhone.
HelloWorld.java on iOS using hotspot (zero interpreter mode). Finally back to the point reached by the great Bob Vandette many years ago.
Need to anchor this.
December 10, 2024 at 8:18 PM
(probably unimpressive for most)
Progress. This is from the XCode console, output from an iPhone.
HelloWorld.java on iOS using hotspot (zero interpreter mode). Finally back to the point reached by the great Bob Vandette many years ago.
Need to anchor this.
Progress. This is from the XCode console, output from an iPhone.
HelloWorld.java on iOS using hotspot (zero interpreter mode). Finally back to the point reached by the great Bob Vandette many years ago.
Need to anchor this.
Time flies.
This is my typical reminder that another 6 months have passed. Need to update the build JDK for building OpenJDK.
This is my typical reminder that another 6 months have passed. Need to update the build JDK for building OpenJDK.
December 9, 2024 at 1:10 PM
Time flies.
This is my typical reminder that another 6 months have passed. Need to update the build JDK for building OpenJDK.
This is my typical reminder that another 6 months have passed. Need to update the build JDK for building OpenJDK.
VisualVM (visualvm.github.io) can be a great help in performance tuning for JavaFX apps. Screenshot below shows the single-threaded JavaFX AppThread spends ca 2/3 of its time in "userspace" (Platform.runLater) and about 30% time in rendering (QuantumToolkit$$...).
November 29, 2024 at 9:25 AM
VisualVM (visualvm.github.io) can be a great help in performance tuning for JavaFX apps. Screenshot below shows the single-threaded JavaFX AppThread spends ca 2/3 of its time in "userspace" (Platform.runLater) and about 30% time in rendering (QuantumToolkit$$...).
Not impressive to most, but I'm happy! Porting latest hotspot to iOS, I am back to the path Bob Vandette followed years ago, using the Zero interpreter. Current status is a Java stacktrace, which is great as that means the classloading and interpreter loop are working, including native code! Pfew.
November 25, 2024 at 7:57 PM
Not impressive to most, but I'm happy! Porting latest hotspot to iOS, I am back to the path Bob Vandette followed years ago, using the Zero interpreter. Current status is a Java stacktrace, which is great as that means the classloading and interpreter loop are working, including native code! Pfew.