| 16 Sep 2025 |
 msgilligan | I submitted a separate PR to make graalvm-oracle_25 the default for graalvm-oracle:
https://github.com/NixOS/nixpkgs/pull/443476 | 16:53:38 |
| msgilligan | And this PR add Zulu 25: https://github.com/NixOS/nixpkgs/pull/443517 | 18:37:11 |
| msgilligan | The Zulu PR is failing nixpkgs-vet with the following message:
Attribute `pkgs.zulu25` is a new top-level package using `pkgs.callPackage ./pkgs/development/compilers/zulu/25.nix { /* ... */ }`.
Please define it in pkgs/by-name/zu/zulu25/package.nix instead.
See `pkgs/by-name/README.md` for more details.
Since the second `callPackage` argument is `{ }`, no manual `callPackage` in pkgs/top-level/all-packages.nix is needed anymore.
This PR introduces additional instances of discouraged patterns as listed above. Merging is discouraged but would not break the base branch.
| 18:38:24 |
| msgilligan | Can this be merged as-is or is there a fix we can apply to the PR? | 18:38:54 |
| msgilligan | * Can this be merged as-is or is there a fix we can/must apply to the PR? | 20:44:12 |
 Tomodachi94 (they/them) | If someone could run nixpkgs-review on https://github.com/NixOS/nixpkgs/pull/364094 I would appreciate it. Apparently it fixes a build failure. I just rebased it onto master because it no longer needs to be built in staging (I'm guessing something switched away from Ant?) | 22:25:25 |
| msgilligan | The PR is still tagged with 101-500 rebuilds on both Linux and Darwin. I'm trying to run nixpkgs-review on both aarch64-linux and aarch64-darwin and seeing large numbers of rebuilds on each (Darwin: 10/32/440 built, Linux: 2/67/1937 built) ljhazhiv91kphy8mzqga8w9pshfj99fk-sweethome3d-7.5.drv is failing on Darwin. | 22:59:36 |
| msgilligan | I'm up to 3 failures on Darwin and 14 on Linux... | 23:10:01 |
| Tomodachi94 (they/them) | Iirc there are a few failures for some dependencies of Ant that have been there forever, I should probably clean those up | 23:28:41 |
| Tomodachi94 (they/them) | Once you've posted the log I'll take a look at the Hydra builds for the failing builds and file a new PR to mark them broken | 23:32:22 |
| Tomodachi94 (they/them) | * | 23:32:45 |
| Tomodachi94 (they/them) | * | 23:33:04 |
| Tomodachi94 (they/them) | Maybe we move it back to staging again? | 23:33:52 |
| Tomodachi94 (they/them) | Thank you for running that massive build btw! | 23:36:09 |
 emily | might be because of unfree or insecure packages being enabled? | 23:56:44 |
| 17 Sep 2025 |
| msgilligan | I posted the results from aarch64-linux, Darwin is still running and may run for some time. | 00:53:51 |
| Tomodachi94 (they/them) | Feel free to cancel that, something is very wrong (500 failures) | 01:14:13 |
| Tomodachi94 (they/them) | I suspect the packages are trying to build using the wrong Java (the JRE that is now included with Ant) | 01:14:49 |
| msgilligan | Cancelled. (It was in the middle of building swift-5.8) | 01:15:38 |
| Tomodachi94 (they/them) | 
Download :blobcatwithitshandsonitshead: | 01:16:30 |
| Tomodachi94 (they/them) | ... apparently the Java used in builds must also be the same as the Java used to run Ant, which is annoying | 01:26:46 |
| Tomodachi94 (they/them) | Maybe? Ant is weird | 01:28:32 |
| msgilligan | I haven't used Ant for 15 years, but that could be true. | 01:29:13 |
| Tomodachi94 (they/them) | Side note: had no idea that Swift was in the dependency graph for Ant | 01:29:58 |
| msgilligan | The wrapper scripts use the following environment variables (if set):
JAVACMD—full path of the Java executable. Use this to invoke a different JVM than JAVA_HOME/bin/java(.exe).
ANT_OPTS—command-line arguments that should be passed to the JVM. For example, you can define system properties or set the maximum Java heap size here.
ANT_ARGS—Ant command-line arguments. For example, set ANT_ARGS to point to a different logger, include a listener, and to include the -find flag.
Note: If you include -find in ANT_ARGS, you should include the name of the build file to find, even if the file is called build.xml.
| 01:31:31 |
| msgilligan | Looks like it defaults to JAVA_HOME unless you set JAVACMD | 01:32:37 |
| Tomodachi94 (they/them) | Hmm, maybe just setting JAVACMD not JAVA_HOME will work? I'll try it | 01:33:14 |
| msgilligan | Maybe the ant Nix package should work similar to Gradle where it runs with and builds with a default JDK but both can be overridden? (And maybe the defaults should be on the older side?) | 01:35:46 |
| Tomodachi94 (they/them) | Changing it to jdk8 worked out for fixing javacc, haven't tried others | 01:37:25 |
| Tomodachi94 (they/them) | Seems to be working for others when I use jdk8 by default | 01:40:43 |
