jetbrains.jdk: fix fallback to default X11 cursor

[jys1670]

The cursor theme in all JetBrains products on X11 has been broken since the __structuredAttrs changes were introduced (see #425529 (comment)):

The root cause seems to be a complete env override of the original JDK package. These changes fix the cursor theme in both PyCharm and CLion for me.

Things done

• Built on platform:
  • x86_64-linux
  • aarch64-linux
  • x86_64-darwin
  • aarch64-darwin
• Tested, as applicable:
  • NixOS tests in nixos/tests.
  • Package tests at passthru.tests.
  • Tests in lib/tests or pkgs/test for functions and "core" functionality.
• Ran nixpkgs-review on this PR. See nixpkgs-review usage.
• Tested basic functionality of all binary files, usually in ./result/bin/.
• Nixpkgs Release Notes
  • Package update: when the change is major or breaking.
• NixOS Release Notes
  • Module addition: when adding a new NixOS module.
  • Module update: when the change is significant.
• Fits CONTRIBUTING.md, pkgs/README.md, maintainers/README.md and other READMEs.

---

Add a 👍 reaction to pull requests you find important.

[tom-ricci]

Just tried it with nix run github:jys1670/nixpkgs/jetbrains-jdk-env-fix#jetbrains.idea, and it solves all my issues from #467636 (comment)

[theCapypara]

I can't test locally at the moment, but note that this PR can't be built with nixpkgs-review-gha.

The build takes multiple hours and then fails with the runner losing communication, probably going OOM.

I don't know if this is a GitHub issue or something caused by this PR, I haven't experienced it yet with building the JetBrains IDEs

https://github.com/theCapypara/nixpkgs-review-gha/actions/runs/20972181982

[aoli-al]

Tried the patch with idea and confirm it fixes the scale and cursor. It is unclear why build time env affects this.

Is it because these flags are missing? https://github.com/NixOS/nixpkgs/blob/54671be46b6c40328559e3c06b7b31a919455b00/pkgs/development/compilers/openjdk/generic.nix#L451C1-L462C8

    NIX_LDFLAGS = lib.concatStringsSep " " (
      lib.optionals (!headless) [
        "-lfontconfig"
        "-lcups"
        "-lXinerama"
        "-lXrandr"
        "-lmagic"
      ]
      ++ lib.optionals (!headless && enableGtk) [
        (if atLeast11 then "-lgtk-3" else "-lgtk-x11-2.0")
        "-lgio-2.0"
      ]
    );

[theCapypara]

Can we please confirm why nixpkgs-review fails before merging? I also wasn't able to build this PR locally with 32 GB RAM, my machine ran out of memory

I think it's animeko, so unrelated to this PR probably, but I'd like to be sure. Can anybody else build this?

I hope I'll have some time later today to investigate this.

My current attempt via GHA is still running. 5h already now, this is very unusual
https://github.com/theCapypara/nixpkgs-review-gha/actions/runs/21028220108/job/60457743798

[theCapypara]

nixpkgs-review result

Generated using nixpkgs-review.

Command: nixpkgs-review pr 479767 --skip-package animeko
Commit: cfd5a5e7e8a28c835f46d6e227176f9d0489845c

---

x86_64-linux

✅ 21 packages built:

• jetbrains.clion
• jetbrains.datagrip
• jetbrains.dataspell
• jetbrains.gateway
• jetbrains.goland
• jetbrains.idea
• jetbrains.idea-oss
• jetbrains.jdk
• jetbrains.jdk-no-jcef
• jetbrains.jdk-no-jcef-17
• jetbrains.jdk-no-jcef-17.debug
• jetbrains.jdk-no-jcef.debug
• jetbrains.jdk.debug
• jetbrains.mps
• jetbrains.phpstorm
• jetbrains.pycharm
• jetbrains.pycharm-oss
• jetbrains.rider
• jetbrains.ruby-mine
• jetbrains.rust-rover
• jetbrains.webstorm

[theCapypara]

I was able to build now locally when skipping animeko, but I notice that pycharm-oss and idea-oss are using quite a significant amount of memory to build. I'm not sure this was the case before, I don't think so?

The oss packages also run (and show the correct cursor).

I personally don't think animeko should block this, if the issue is even caused by this PR, so LGTM!

I started a nixpkgs-review-gha run without animeko, to see if aarch64 builds.

Thanks for figuring this out!

pinging @pokon548 if you want to look into animeko.

[theCapypara]

nixpkgs-review result

Generated using nixpkgs-review-gha

Command: nixpkgs-review pr 479767 -P animeko
Commit: cfd5a5e7e8a28c835f46d6e227176f9d0489845c (subsequent changes)
Merge: 0aaecd18751207f31fa3cef030a7e133c5fd1e97

Logs: https://github.com/theCapypara/nixpkgs-review-gha/actions/runs/21039435876

---

x86_64-linux

✅ 21 packages built:

• jetbrains.clion
• jetbrains.datagrip
• jetbrains.dataspell
• jetbrains.gateway
• jetbrains.goland
• jetbrains.idea
• jetbrains.idea-oss
• jetbrains.jdk
• jetbrains.jdk-no-jcef
• jetbrains.jdk-no-jcef-17
• jetbrains.jdk-no-jcef-17.debug
• jetbrains.jdk-no-jcef.debug
• jetbrains.jdk.debug
• jetbrains.mps
• jetbrains.phpstorm
• jetbrains.pycharm
• jetbrains.pycharm-oss
• jetbrains.rider
• jetbrains.ruby-mine
• jetbrains.rust-rover
• jetbrains.webstorm

---

aarch64-linux

✅ 21 packages built:

• jetbrains.clion
• jetbrains.datagrip
• jetbrains.dataspell
• jetbrains.gateway
• jetbrains.goland
• jetbrains.idea
• jetbrains.idea-oss
• jetbrains.jdk
• jetbrains.jdk-no-jcef
• jetbrains.jdk-no-jcef-17
• jetbrains.jdk-no-jcef-17.debug
• jetbrains.jdk-no-jcef.debug
• jetbrains.jdk.debug
• jetbrains.mps
• jetbrains.phpstorm
• jetbrains.pycharm
• jetbrains.pycharm-oss
• jetbrains.rider
• jetbrains.ruby-mine
• jetbrains.rust-rover
• jetbrains.webstorm

---

x86_64-darwin

✅ No rebuilds

---

aarch64-darwin

✅ No rebuilds

[MattSturgeon]

nixpkgs-review-gha still fails, so the memory requirements really seem to be higher. I don't have the time at the moment to directly compare before and after.

It would be great if we could fix this, since this will probably also prevent the update scripts of the oss packages from running and make reviewing IDE updates harder, but I'd still rather merge this to fix the cursor and scaling issues.

If there is an issue caused by some env attr that gets passed into the JDK derivation, that's worth investigating.

But this PR's actual diff is very small and seems trivially correct to me; when overriding env or any other derivation attr, you almost certainly want to preserve prev attrs rather than overriding them all.

I notice you were able to run nixpkgs-review-gha after all, so maybe it was a transient GHA runner issue rather than an actual package issue?

[nixpkgs-ci]

Successfully created backport PR for release-25.11:

• [Backport release-25.11] jetbrains.jdk: fix fallback to default X11 cursor #480793