cmd/readall: clean up todos

[apainintheneck]

• Have you followed the guidelines in our Contributing document?
• Have you checked to ensure there aren't other open Pull Requests for the same change?
• Have you added an explanation of what your changes do and why you'd like us to include them?
• Have you written new tests for your changes? Here's an example.
• Have you successfully run brew style with your changes locally?
• Have you successfully run brew typecheck with your changes locally?
• Have you successfully run brew tests with your changes locally?

---

Edit: We decided to keep the current behavior of running all os/arch combinations and remove the todos and deprecations.

Removes temporary default to readall for every os/arch combination
after updating brew-test-bot and brew/cask to pass the appropriate
arguments in on CI.

• Add --os and --arch flags to readall. #15470
• Cleanup various TODOs #15937 (comment)

Also, add more constants for os/arch combinations. This allows us
to make validating all os/arch combinations the default in
Readall.valid_tap? which is needed to keep the same behavior
in brew tap that we had before. I also updated a few other
spots around the codebase to use those new constants.

One more thing was updating the integration test. In local testing,
this didn't change the runtime so it seemed like a no-brainer.

Todo

1. The CI used by homebrew/core and homebrew/cask has been updated.
   • Add os/arch options to readall command homebrew-test-bot#969
   • Add os/arch options to readall command homebrew-cask#155397
2. The readall performance improvements PR has been merged in.
   • Improve performance of brew readall #16007
     • I don't actually think there's much overlap between these PRs.

[Bo98]

Just to check: does this change the default beahviour of brew readall to be no simulation?

[apainintheneck]

Yes, the idea here is to follow-up on the todo to match the behavior described in the original issue and the deprecation notice for the --no-simulate switch. If we want to just keep the default behavior the same as it is now, we could just remove the deprecation notice and tweak things so that all os/arch combinations are passed by default. I haven't really looked into why we chose one default over the other.

[MikeMcQuaid]

Great work @apainintheneck, loving this so far.

Yes, the idea here is to follow-up on the todo to match the behavior described in the original issue and the deprecation notice for the --no-simulate switch. If we want to just keep the default behavior the same as it is now, we could just remove the deprecation notice and tweak things so that all os/arch combinations are passed by default. I haven't really looked into why we chose one default over the other.

This seems better!

[apainintheneck]

@MikeMcQuaid By "This seems better!" do you mean we should continue deprecating the --no-simulate switch or make all os/arch combinations the default instead? I'm not really sure which one you meant.

[apainintheneck]

@Bo98 Do you think we should just keep the current behavior as the default and undeprecate (precate?) the --no-simulate switch?

[MikeMcQuaid]

@MikeMcQuaid By "This seems better!" do you mean we should continue deprecating the --no-simulate switch or make all os/arch combinations the default instead? I'm not really sure which one you meant.

Sorry, the latter but I don't have strong feelings either way!

[apainintheneck]

I'm fine with re-enabling the --no-simulate switch in that case. Is there anything special we should do from a user facing perspective for that? I don't think it's really a big deal. Also, I guess this means that the previous two cleanup PRs were unnecessary but I guess it's just clearer to include the options explicitly. We can revert them too but it doesn't really matter either way.

[MikeMcQuaid]

Thanks again @apainintheneck!