Update README.md to trigger CI

[ViralBShah]

@amontoison - Running CI here to see if we can tag MKL 0.6. We can tag MKL 0.6 if this basically works on all the platforms. There are some tests failing because of numerical precision mismatch in MKL 2021/2022, e.g. gsvd - but that shouldn't hold up the version bump.

This should help jump-dev/Ipopt.jl#327

[ViralBShah]

I think this looks pretty good. I feel we can go ahead and tag 0.6. Win64 tests did pass locally for me.

Anyone know what this is:

      From worker 27:	┌ Error: The `LinearAlgebra/addmul` test set mutated ENV and did not restore the original values

@KristofferC @DilumAluthge @chriselrod

[amontoison]

Thanks @ViralBShah! I also use LBT for HSL.jl and It's great to easily switch between OpenBLAS32 and LP64 MKL with just using MKL.

[DilumAluthge]

I think this looks pretty good. I feel we can go ahead and tag 0.6. Win64 tests did pass locally for me.

Anyone know what this is:

      From worker 27:	┌ Error: The `LinearAlgebra/addmul` test set mutated ENV and did not restore the original values

@KristofferC @DilumAluthge @chriselrod

This is a test to make sure that if a test sets mutates ENV, it restores the original values at the end of the test set. This is to ensure that modifications to ENV do not persist between test sets.

Does MKL.jl mutate ENV?

[DilumAluthge]

That check is important in Base CI, where we want to keep earlier test sets from causing problems for later test sets.

It's less useful here. We can add an escape hatch so that you can disable that check in this repo while keeping the check for Base CI.

[DilumAluthge]

We can add an escape hatch so that you can disable that check in this repo while keeping the check for Base CI.

JuliaLang/julia#47148