Run FFTW in GC-safe regions

[JamesWrigley]

Should help prevent GC pressure. In a larger application I observed memory not being reclaimed when doing multithreaded FFTs, which I suspect to be because we were calling FFTW without entering a GC-safe, which blocks the GC when it could be running.

Would it be possible to make a new release with this?

[codecov]

Codecov Report

❌ Patch coverage is 80.00000% with 3 lines in your changes missing coverage. Please review.
✅ Project coverage is 73.58%. Comparing base (290d161) to head (5e185e7).

Files with missing lines	Patch %	Lines
src/fft.jl	72.72%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #340      +/-   ##
==========================================
+ Coverage   73.45%   73.58%   +0.13%     
==========================================
  Files           5        5              
  Lines         535      549      +14     
==========================================
+ Hits          393      404      +11     
- Misses        142      145       +3     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[stevengj]

Does this seem to fix the problem in your application?

[JamesWrigley]

Not sure, I wasn't able to test it yet. Claude did come up with a smaller local reproducer that did seem to be fixed though. In the real application there's lots of threads doing a lot of allocations so memory usage grew quite quickly with multithreaded plans.