dnspy: init at 6.5.1

[js6pak]

https://github.com/dnSpyEx/dnSpy

dnSpy(Ex) is a decompiler frontend and debugger for .NET assemblies. One small issue is that it's windows-only which makes this package a little cursed, but hey, it works. :)

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/)
• 24.11 Release Notes (or backporting 23.11 and 24.05 Release notes)
  • (Package updates) Added a release notes entry if the change is major or breaking
  • (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.

---

Add a 👍 reaction to pull requests you find important.

[GGG-KILLER]

Changing this will make so that other .NET packages support platforms that they don't.

The correct thing would be changing the .NET SDK and/or runtime to include windows.

[js6pak]

Actually the first issue with this code is that it should take the meta.platforms of the dotnet-runtime not dotnet-sdk right? (@corngood)

Also this is an artifact of when this PR was using self-contained, with which you can make a derivation for platforms that nixpkgs doesn't have a runtime package for (because it can download the runtime from nuget).

[GGG-KILLER]

Yes, the correct would be for it to get the runtime's platforms. In this case it doesn't really matter but I guess there could be cases where a runtime exists for a platform which the SDK doesn't in the future.

[corngood]

I'm not sure we should even be intersecting the platforms here, or providing a default. Are there examples of this in other languages?

It seems like we should just let the dependencies fail to eval.

[GGG-KILLER]

I don't know if there are examples of this in other languages, however I don't see the point in allowing platforms where the packages will not be able to run on.
If we can provide this safeguard without too much trouble I don't see why not.

[js6pak]

however I don't see the point in allowing platforms where the packages will not be able to run on

As I mentioned before, it's useful for self-contained builds. But that could be solved by checking if dotnet-runtime is null first.

If we can provide this safeguard without too much trouble I don't see why not

The problem with intersecting like this is that you are just silently dropping platforms the user provided. A better option could be using assert?

[GGG-KILLER]

If we can provide this safeguard without too much trouble I don't see why not

The problem with intersecting like this is that you are just silently dropping platforms the user provided. A better option could be using assert?

We could do that, but it'd break pretty much every .NET package in existence currently and a breaking change we'd need to document.

And it'd make the process of packaging more annoying because instead of lib.platforms.linux, one would instead need to type [ "x64_86-linux" "aarch64-linux" ], which isn't a lot of work so I wouldn't consider this a blocker, just a "I'd prefer not to have to".

[GGG-KILLER]

Can't this be in a separate shell file with the nix-shell shebang instead of writeShellScript?
It'd enable analysis with shell check.

[js6pak]

I don't think it's worth it for 2 lines. I think the real solution is making nix-update run fetch-deps? Other packages would benefit from it aswell

[GGG-KILLER]

True, given that we now have nix-update it'd be good to make it automagically run fetch-deps.
I'll see if I can't get a PR running for it eventually.

[corngood]

take a look at tone.updateScript:

#!/usr/bin/env nix-shell
#!nix-shell --pure -i bash -p bash nix nix-update git cacert
set -eo pipefail

prev_version=$(nix eval --raw -f. tone.version)
nix-update tone
[[ $(nix eval --raw -f. tone.version) == "$prev_version" ]] ||
  "$(nix-build . -A tone.fetch-deps --no-out-link)"

I think it's important not to run fetch-deps unconditionally.

[GGG-KILLER]

With recent changes to nix-update, fetch-deps is automagically called by it if it exists, so it doesn't need to be done here.

[js6pak]

I did notice that, but it's not in a nix-update release yet.