fix!: make strict preparation the default

[gdalle]

Warning

Breaking change for DifferentiationInterface

This PR makes strict=Val(true) the default setting in preparation functions DI.prepare_something. From now on, when the signature at execution differs from the signature at preparation (e.g. a gradient prepared on Vector{Float64} and executed on Vector{Float32}), an informative error will be thrown. Indeed, only the case where the types match is actually supported.

Tip

If this change causes your code to error, a temporary fix is to manually set strict=Val(false) in every call to DI.prepare_something. The right fix however is to figure out why types differ, and resolve that.

[codecov]

Codecov Report

Attention: Patch coverage is 81.25000% with 12 lines in your changes missing coverage. Please review.

Project coverage is 97.93%. Comparing base (4f0e496) to head (36f5a4a).

Files with missing lines	Patch %	Lines
...tiationInterfaceTest/src/tests/correctness_eval.jl	81.25%	12 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #799   +/-   ##
=======================================
  Coverage   97.93%   97.93%           
=======================================
  Files         128      128           
  Lines        7693     7693           
=======================================
  Hits         7534     7534           
  Misses        159      159           

Flag	Coverage Δ
DI	98.81% <ø> (ø)
DIT	95.61% <81.25%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 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.