| 14 Mar 2026 |
| amadaluzia -> 4d2.org changed their display name from amadaluzia[uorg] to amadaluzia -> 4d2.org. | 21:23:08 |
| 21 Mar 2026 |
Benedikt Ritter (britter) | There's something off with jdk11: https://github.com/NixOS/nixpkgs/issues/502023 I've looked into openjdk/generic.nix but gave up pretty quickly 😅 | 20:25:55 |
Benedikt Ritter (britter) | This makes the tests for Gradle fail as well, because they expect the package version to equal the version reported by the Java binary. | 20:26:31 |
| 22 Mar 2026 |
Benedikt Ritter (britter) | It looks like 11.0.31+0 incorrectly reports 11.0.30. I've created a PR to bump to 11.0.31+1 which correctly reports 11.0.31 🤷♂️ reviews appreciated! https://github.com/NixOS/nixpkgs/pull/502168 | 20:31:02 |
emily | 11.0.31 isn't released yet | 20:38:10 |
emily | the -ga tags are final releases https://github.com/openjdk/jdk11u/tags | 20:38:18 |
emily | the update script should only update to those | 20:38:25 |
emily | may be better to backport whatever was needed for the new glibc than to keep following a pre-release | 20:38:43 |
Benedikt Ritter (britter) | Now that sheds some light! I agree we should not be following pre-releases. I wasn't aware of the details of the numbering scheme and assumed +x would mark some kind of patch version. | 20:49:40 |
Benedikt Ritter (britter) | Maybe we should add a check that makes sure the tag name ends with -ga | 20:50:14 |
Benedikt Ritter (britter) | @emilazy:matrix.org I've closed my PR. At the same time I don't see myself backporting a patch that fixes the glibc issue. If anybody else wants to give that a try, feel free. Or we'll just wait until ga is released. | 20:57:20 |
emily | yeah, it's the most confusing numbering scheme around… | 20:57:45 |
emily | the JSON files deliberately recall the full tag name to have the canonical version available | 20:57:53 |
emily | but the update script only updates to final releases | 20:58:00 |
emily | basically it counts up to +n and then they tag -ga at the same commit as +n | 20:58:38 |
| 1 Apr 2026 |
| @onur-ozkan:matrix.org left the room. | 08:18:36 |
| 5 Apr 2026 |
Infinidoge 🏳️⚧️ | Is there anything stopping us from bumping the default JDK package to 25? | 23:03:57 |
msgilligan | It's been months since I've looked at this, but there were issues when I last checked.
See: https://github.com/NixOS/nixpkgs/issues/430389 | 23:06:36 |
msgilligan | Related stale PRs:
- https://github.com/NixOS/nixpkgs/pull/443523
- https://github.com/NixOS/nixpkgs/pull/462041
| 23:07:57 |
| 6 Apr 2026 |
Infinidoge 🏳️⚧️ | Hmm | 05:11:00 |
Infinidoge 🏳️⚧️ | At some point will try running nixpkgs-review against a JDK 25 default bump and let the ~1600 packages build | 05:11:28 |
| 7 Apr 2026 |
Benedikt Ritter (britter) | Hey folks, I think we need to form an opinion on our readiness for 26.05. @NixOS/java has been pinged on the release issue asking for feedback. If everything is good fo 26.05 it's fine to just state that :) | 13:34:07 |
| 8 Apr 2026 |
Benedikt Ritter (britter) | Okay, so let's do it like this: if I don't hear anything about 26.05 readiness by the end of this week, I will respond that java is ready, ok? Please react with 👍️ or 👎️ | 17:30:26 |
msgilligan | I'm assuming thumbs-up means I agree to your proposed process, not that I'm saying Java is ready. | 17:44:10 |
msgilligan | I'd like to do some minor version bumps on some of the JDK 25 packages (e.g. Zulu) that need it, but I don't think that is a gate for readiness. | 18:24:44 |
msgilligan | I'd also like to bump jextract to the JDK 25 version for 26.05. | 18:25:19 |
Infinidoge 🏳️⚧️ | In reply to @britter:yatrix.org Hey folks, I think we need to form an opinion on our readiness for 26.05. @NixOS/java has been pinged on the release issue asking for feedback. If everything is good fo 26.05 it's fine to just state that :) I'm currently running nixpkgs review on an attempt to bump the default JDK, to see how reasonable it would be to do before the 26.05 cutoff. If there aren't issues with that build, would that be reasonable to get into 26.05? | 19:03:46 |
Benedikt Ritter (britter) | yes | 19:49:29 |
Benedikt Ritter (britter) | Sounds good. I had some luck running a Gradle nixpkgs-review with some tweaks to the build-args: https://matrix.to/#/!aRKdLCkUeIFjRPZuJT:nixos.org/$d_QXGM7Q6ZuAAGth7phzxwq7_5fqV-2nIuLi2HEs0V4?via=nixos.org&via=matrix.org&via=nixos.dev I did this to work around the issue that my laptop was becoming unusable due to the amount of builds being started in parallel. It still took more than 8 hours to complete. | 19:51:41 |
Infinidoge 🏳️⚧️ | Conveniently I have a pretty strong server I use for builds, Daedalus :) | 20:05:57 |