Guillaume Smet
@gsmet.bsky.social
Happy camper at Red Hat. Improving Quarkus one commit at a time.
I'll add more info in the announcement on Monday once I have all the details about which typologies of issues can be caused by this particular regression.
August 30, 2025 at 8:10 AM
I'll add more info in the announcement on Monday once I have all the details about which typologies of issues can be caused by this particular regression.
As mentioned, the issue wasn't in Quarkus itself but in Vert.x and it's this one: github.com/eclipse-vert... . Now, it's not exactly easy to understand which problems it could have caused.
Ack the HTTP client response window from the correct event-loop thread. by vietj · Pull Request #5683 · eclipse-vertx/vert.x
Motivation:
The recent change of the HTTP client response flow control to avoid releasing a connection to the pool not fully received introduced a bug creating a data race when acking the response ...
github.com
August 30, 2025 at 7:59 AM
As mentioned, the issue wasn't in Quarkus itself but in Vert.x and it's this one: github.com/eclipse-vert... . Now, it's not exactly easy to understand which problems it could have caused.
We will add more information soon but the priority was to get the release out (and timing was less than ideal). As to the problems you encountered, happy to start a conversation about what we can improve. I understand the urge to vent on social media but it won't make the project better.
August 30, 2025 at 7:54 AM
We will add more information soon but the priority was to get the release out (and timing was less than ideal). As to the problems you encountered, happy to start a conversation about what we can improve. I understand the urge to vent on social media but it won't make the project better.
Thanks to me and a lot of others! That's the plan!
February 7, 2025 at 12:06 PM
Thanks to me and a lot of others! That's the plan!
I decided try to reproduce it based on the stacktrace we got. I managed to do it and posted the reproducer. One hour later, my colleague created a PR with a fix.
So please, pretty please, provide reproducers, it makes a huge difference and is the best guarantee to get your issues fixed promptly.
So please, pretty please, provide reproducers, it makes a huge difference and is the best guarantee to get your issues fixed promptly.
January 30, 2025 at 2:17 PM
I decided try to reproduce it based on the stacktrace we got. I managed to do it and posted the reproducer. One hour later, my colleague created a PR with a fix.
So please, pretty please, provide reproducers, it makes a huge difference and is the best guarantee to get your issues fixed promptly.
So please, pretty please, provide reproducers, it makes a huge difference and is the best guarantee to get your issues fixed promptly.
Also building a simple reproducer quite often give us a better idea of where to look at because you will have to isolate the issue. github.com/quarkusio/qu... is a good example of this. We didn't have a proper way to reproduce it. My colleague decided to close the issue - which is our process. 🧵
IllegalArgumentException when parsing Accept Header · Issue #45385 · quarkusio/quarkus
Describe the bug I've seen 34858, but it seems that the Issue has resurfaced, because I can reproduce it with version 3.15.1 When sending a request with a header "Accept: Any", the following Stack-...
github.com
January 30, 2025 at 2:14 PM
Also building a simple reproducer quite often give us a better idea of where to look at because you will have to isolate the issue. github.com/quarkusio/qu... is a good example of this. We didn't have a proper way to reproduce it. My colleague decided to close the issue - which is our process. 🧵
Note that I'm all for finding a way to escape from this vicious circle and ideally the best way to do it is to contribute fixes that help make this thing more solid.
December 21, 2024 at 1:54 PM
Note that I'm all for finding a way to escape from this vicious circle and ideally the best way to do it is to contribute fixes that help make this thing more solid.
Yeah, unfortunately, it's a bit of a vicious circle. Given it's brittle, we are always afraid to break something and as we don't have experts to review changes, things can be slow. I'll try to have a look at the PR during the holiday.
December 21, 2024 at 1:53 PM
Yeah, unfortunately, it's a bit of a vicious circle. Given it's brittle, we are always afraid to break something and as we don't have experts to review changes, things can be slow. I'll try to have a look at the PR during the holiday.
Gradle is a very flexible and complex beast and that makes integration with it a lot more difficult... and unfortunately quite brittle. The fact that we don't have Gradle expertise in house doesn't help. Issues with test cases are welcome (and contributions even more so!).
December 20, 2024 at 4:18 PM
Gradle is a very flexible and complex beast and that makes integration with it a lot more difficult... and unfortunately quite brittle. The fact that we don't have Gradle expertise in house doesn't help. Issues with test cases are welcome (and contributions even more so!).
FWIW, we have seen very weird behaviors/bugs with GraalVM for Java 17, we are not recommending it at all. Things are moving so fast, it's hard to provide stable information about the versions you should use...
November 27, 2024 at 5:09 PM
FWIW, we have seen very weird behaviors/bugs with GraalVM for Java 17, we are not recommending it at all. Things are moving so fast, it's hard to provide stable information about the versions you should use...