Perturbation solutions getting stuck when rapidly evaluated

[jlperla]

This seems to occur with a @btime as well as with calls to FiniteDifferences, such as when called with test_rrule etc.

Currently going through to see if concurrency issues are at play:

• [x] Caching
• [x] Deep copy of the module/module_wrapper
• [x] Globals and closures
• [x] GeneralizedSylvester
• [x] RecursiveFactorization's lu
• [x] General LAPack threading
• [x] lyap in matrixequations
• [x] Container has Module

[jlperla]

Happens without any actual calculations. Compiler failing in some odd way?

[jlperla]

Cracked. Some sort of compiler bug with globals. Will add MWE later.