Use surveil in move tutorial

[xsebek]

This should make the move tutorial run faster by making most of the robots surveil the door to their section.

[xsebek]

@byorgey this is how instant has to be used to not pay for tangible run. The result is the rest of the computation which waits and places the wall.

[byorgey]

For me, it's super laggy up until you move to the first flower, then very snappy after that. By "super laggy" I mean e.g. it takes several seconds to respond at all when I start typing move; move; even after that it still takes up to half a second to respond to each keypress.

I actually don't quite understand what's going on here. Sometimes, when I start the tutorial it pauses for several seconds before showing the goals dialog, and then is reasonably fast after that. But sometimes, it shows the goal dialog right away, and then once I exit that it sits there not responding for several seconds. The first feels annoying but acceptable; the second feels pretty bad because it looks like it is ready for you to start typing but it is not.

[xsebek]

@byorgey I removed a few busy waiting robots and put the check robot to sleep if you are far from the door.

I don't think I can easily optimize this further; most of the time is spent parsing the same file 60 times:

I could make a robot that runs the file and instantly builds new robots, which would teleport to their place.
That would only parse once, but have new problems:

• timing problems - build takes another tick, so the first corridor would have to be longer
• it would use manually written position lists with no declarative map
• I don't want to write that; I would rather work on import

[byorgey]

Hmm, it definitely feels more responsive now (@DanRyba253 also posted some timing data in Discord suggesting the same thing). However, it seems slightly broken: when I execute move; move to go to the first flower, the corridor to the second flower opens up, but there is no goal popup dialog. After I move to the second flower, two goal dialogs pop up in succession.

[byorgey]

And yes, as for import, definitely don't spend effort trying to optimize the repeated file reading, we'll work on that separately.

[xsebek]

@byorgey, thanks for pointing it out; I made a typo while changing the conditions to use one robot.

[byorgey]

Looks good now, this seems a strict improvement over the previous version.

I still don't understand why sometimes I get a long pause before the initial goal modal is shown, and sometimes I get a long unresponsive pause after dismissing the goal modal. I have not been able to find any particular way to consistently get one behavior or the other. But whatever the reason I don't think it's the fault of this scenario.