Also don’t rename libglass on Windows to avoid PDB name collision.

…6879)

Removed from API in #1868, sim definition was left dangling.

Also adds extern "C" to the sim file.

This makes it easier to define schemas when the type name is non-trivial (e.g., templated structs).

This is breaking for a) custom struct implementations and b) anything calling `wpi::Struct<T>::GetTypeString(info...)` in C++ directly. In both cases, it's a simple translation: For A, rename `GetTypeString()` to `GetTypeName()` and remove the struct: at the beginning, and for B, use `wpi::GetStructTypeString<T>(info...)` instead.

It checked out the main branch instead of the PR branch. Now it checks out the PR branch.

#6884)

That behavior has not been present since PR #4158 was merged more than 2 years ago and imo should not be added back because it was surprising and not consistent with the most common use case of registering a callback permanently.

If the interrupt edge tests are running while under heavy CPU load (like building wpilib) they are prone to failure since the interrupt thread doesn't have enough time to set up callbacks. The interrupt edge tests now copy the original AsynchronousInterrupt test, which has a 0.5s delay after the interrupt is enabled. Running the new interrupt tests while building allwpilib causes far less failures than the old tests.

…ly construct Triggers (#6738)

This has been a common footgun for teams, due to calling the factory functions in periodic loops.