Performance regression in #34789

[amcasey]

The extra work done in #34789 is valuable, but maybe it doesn't need to be so expensive.

Setup:

1. Clone https://github.com/amcasey/material-ui.git
2. Check out Benchmark
3. yarn to restore packages
4. yarn typescript to prebuild and run some TS tests

Repro

1. tsc -p docs

On a random Mac Mini:

• 10-run avg on 7d77ecb: 35,482ms
• 10-run avg on 28050d5: 35,908ms (1% slower)

Note: to build old TS commits, you probably need to change const { default: chalk } to const chalk in scripts/build/utils.js.

[amcasey]

FYI @sandersn

[ahejlsberg]

I would expect some slowdown from this PR since we now do more work. Specifically, when relating two types we take into account whether one is contained in an intersection and separately compute the relationship for that situation. The material-ui test case heavily uses intersection types, so the observed 1% slowdown doesn't seem entirely unreasonable.

[amcasey]

@ahejlsberg I agree that 1% in a pathological case isn't the end of the world and I'm fine won't-fixing this - I brought it up as one more straw on the camel's back.