18x increase in compile time on v1.8

[baggepinnen]

The following snippet takes

7.833256 seconds (36.60 M allocations: 1.727 GiB, 7.05% gc time, 100.00% compilation time

on julia v1.7.3
but

146.158104 seconds (474.09 M allocations: 79.036 GiB, 4.20% gc time, 100.00% compilation time)

on julia v1.8

using ControlSystems
a_2 = [-5 -3; 2 -9]
C_212 = ss(a_2, [1; 2], [1 0; 0 1], [0; 0])
@time norm(C_212)

an @profile is available in html format here.

[KristofferC]

There are at least two regressions I think. Bisecting I see at one time it goes from ~10 seconds to

 39.839046 seconds (102.91 M allocations: 9.583 GiB, 9.93% gc time, 100.00% compilation time)

[ianatol]

There are at least two regressions I think. Bisecting I see at one time it goes from ~10 seconds to

 39.839046 seconds (102.91 M allocations: 9.583 GiB, 9.93% gc time, 100.00% compilation time)

Any particularly guilty commits? @KristofferC

[KristofferC]

Few more steps left :)

[KristofferC]

First 5x regression is 2c0e1bc, cc @JeffBezanson

[KristofferC]

The second big regression (30s -> 200s) comes from 26b0b6e, cc @aviatesk