stdenv: arithmetic fixes

[happysalada]

Motivation for this change

This is a PR in the series of stdenv proposed shell fixes

This PR can be controversial.

This removes the -i flag that makes the shell parser dynamic. (the -i flag does not work with oil shell, but is not needed as well, the drawback is that the intent that the variable is an integer is less clear).
This also uses the ((...)) form for arithmetic computation.

The original proposed commits are 59ec2c7
and e0b7c12

I omitted some changes in the original PR because I'm not 100% sure about those.

Things done

• Tested using sandboxing (nix.useSandbox on NixOS, or option sandbox in nix.conf on non-NixOS linux)
• Built on platform(s)
  • NixOS
  • macOS
  • other Linux distributions
• Tested via one or more NixOS test(s) if existing and applicable for the change (look inside nixos/tests)
• Tested compilation of all pkgs that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review wip"
• Tested execution of all binary files (usually in ./result/bin/)
• 21.11 Release Notes (or backporting 21.05 Relase 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.

[lovesegfault]

This whole set of changes needs to target staging, and not master.

[happysalada]

@lovesegfault thank you for your feedback.
I only mentioned in it in the original PR here #13059
Basically, I'm only targeting master to get faster hydra builds.
This PR will definitely merged (if it can) on staging.

[7c6f434c]

You wanted this -i removed, right?

[happysalada]

Thank you, you're right, I forgot that one.

[happysalada]

I just checked this again. I tried separating the PRs as much as possible and some of the other -i are taken care of in another PR. I'm going to remove that one as well though.
The original intention was to make it easier to test. But since the hydra tests are so long, they have had to be grouped anyways. Now I realise it wasn't the best strategy.

[7c6f434c]

I noticed this declaration because you did update the variable's use.

[happysalada]

Yes, you're right on this one.

[7c6f434c]

Well, for all intention-revealing, in my opinion -i is also a non-local wait-what waiting to happen (although it is true that in this case it should not happen becaue we only assign good integers, but …)

And (( )) was already used, so using it more consistently shouldn't hurt.

Have you checked any cross-compilation?

(In the ideal world we would have a full-cross test where architecture A builds a cross-compiler from B to C, the it is run in Qemu-emulated B, then the output is checked on Qemu-emulated C; but I am not aware of such a test easily available; pkgsCross could be used for some one-step tests, though)

[happysalada]

I haven't checked cross compilation that is a good idea.
I am still trying to figure out how to run massive rebuilds myself with hydra. I'll update when I have made some progress on that.

[happysalada]

Just talked to vladimir about running some cross tests. Hopefully I can do it in the next day or 2

[happysalada]

Ok vladimir was nice enough to run cross compilation on a whole set of PRs
#132490
If you have time on your hands, feel free to have a look at the other PRs in the list
(the tests were run before that last -i removal, but I think we can agree that it won't change much).
Thanks again for taking a look at this!