rocmPackages.migraphx: override oneDNN

[dotlambda]

Things done

• Built on platform(s)
  • x86_64-linux
  • aarch64-linux
  • x86_64-darwin
  • aarch64-darwin
• For non-Linux: Is sandboxing enabled in nix.conf? (See Nix manual)
  • sandbox = relaxed
  • sandbox = true
• Tested, as applicable:
  • NixOS test(s) (look inside nixos/tests)
  • and/or package tests
  • or, for functions and "core" functionality, tests in lib/tests or pkgs/test
  • made sure NixOS tests are linked to the relevant packages
• Tested compilation of all packages that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage
• Tested basic functionality of all binary files (usually in ./result/bin/)
• Nixpkgs 25.11 Release Notes (or backporting 24.11 and 25.05 Nixpkgs Release notes)
  • (Package updates) Added a release notes entry if the change is major or breaking
• NixOS 25.11 Release Notes (or backporting 24.11 and 25.05 NixOS Release notes)
  • (Module updates) Added a release notes entry if the change is significant
  • (Module addition) Added a release notes entry if adding a new NixOS module
• Fits CONTRIBUTING.md, pkgs/README.md, maintainers/README.md and other contributing documentation in corresponding paths.

---

Add a 👍 reaction to pull requests you find important.

[LunNova]

This seems more fragile to randomly breaking the overriden version migtaphx uses on future bumps to the onednn package, or if caught requiring conditioning on finalAttrs.version or making the override more complicated. What's the upside? Usually don't we remove these older versioned packages once there are no users?

…

----- Original message ----- Subject: Re: [NixOS/nixpkgs] rocmPackages.migraphx: override oneDNN (PR #416821) Date: Saturday, 14 June 2025 16:50 @dotlambda <https://github.com/dotlambda> requested your review on: #416821 <#416821> rocmPackages.migraphx: override oneDNN. — Reply to this email directly, view it on GitHub <#416821 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAF7A2EU2TATOZ7UJYBVNZ33DSYLTAVCNFSM6AAAAAB7KRVHGWVHI2DSMVQWIX3LMV45UABCJFZXG5LFIV3GK3TUJZXXI2LGNFRWC5DJN5XDWMJYGE2TGMJUHEYTMMI>. You are receiving this because your review was requested.Message ID: ***@***.***>

[dotlambda]

What's the upside? Usually don't we remove these older versioned packages once there are no users?

This way it's clear who's responsible for making sure oneDNN version 2 still works and we won't forget to remove the package when it's not needed anymore.
I think in general, versioned attributes should be avoided as much as possible.