add a command to enable atmos devices on arbitrary maps

[Alkheemist]

About the PR

Adds a command to re-enable atmos on maps.

Why / Balance

Admins can have a little atmos, as a treat

Technical details

Adds a hashset containing maps that atmos should be enabled on to the atmos system and checks if the map is in that hashset when checking whether or not to process atmos devices.

Note that the way AtmosInputCanRunOnMap works means that this is more computationally expensive the more atmos devices you have on maps that are not the main one, so hopefully the additional check isn't going to bog down the server. But before/after profiling should probably be done to see the real impact.

And the hashset seems cleaner to check compared to iterating through a list. The easiest would be a field on the component, but that's in the engine and I don't want to try and mess with that right now.

How to test

• Go to your ATAG (or on exped)
• Build an atmos setup, lament the devices not doing anything
• Use the command
• Rejoice

Media

dotnet_01_11-24-15.mp4

Requirements

• I have read CONTRIBUTING.md and and am following the Pull Request and Changelog Guidelines.
• I have added media to this PR or it does not require an ingame showcase.
• I have reviewed the Ship Submission Guidelines if relevant.
• I confirm that the content in this PR is my own work, and/or is properly attributed to the original author(s).

Breaking changes

Changelog

N/A, admin only

[Alkheemist]

reject ECS, return to hashset
(I forgot I could've just checked for the presence of a component)

[JojyAsmir]

Realistically, this is going to be used for two things mainly: Cryo pods on the Rimworld event, and making the scrubbers work on the minicentcomm POI. Both of which are great to me! Though, how would I go about testing that before/after profiling stuff?

[NGGJamie]

Since it's unlikely mamy maps will be set like this, and we're not using special properties of HashSets, it would probably be more optimal to use a normal list. HashSets have to hash the value being checked before knowing if it's in the list, which does have some measure of cost.

But since we're unlikely to get more than 10 of these maps on any shift even in generous terms, the O(n) check of a list would probably beat the O(1) check with hash cost. Lists will also play nicer with caching and branch prediction. Granted, we are checking an integer, so that cost will be especially low and probably measured in terms of ns.

A SortedSet could also work if we'd want to skip hashing and get O(log(n)) lookups.

[Alkheemist]

rewrote to use ECS since we'd need to do so for persistence reasons later.

[arimah]

I'm sorry, I'm a nitpicky bastard 😢 Looks good, just one TINY change.

[blackknight954]

tested and works as intended, now all we need is working thermal scanners..