internal/common: guard SIGUSR2 debug handler on Windows

[AkshatDudeja77]

This change disables the SIGUSR2-based debug signal handler on Windows.

Windows does not support SIGUSR2, which causes build failures due to
syscall.SIGUSR2 being undefined. The handler is now conditionally skipped
at runtime on Windows while preserving existing behavior on Unix-like
platforms.

No functional changes on Linux.

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[klueska]

IMEX is only supported on Linux

[AkshatDudeja77]

IMEX is only supported on Linux

Yes, agreed that IMEX is Linux-only.

This change does not attempt to support Windows. It only guards the debug SIGUSR2 handler to avoid build/runtime issues when the codebase is analyzed or built on Windows (where SIGUSR2 is undefined), while preserving identical behavior on Linux.

There is no functional impact on Linux paths.

[jgehrcke]

We have to think about all of the following:

• https://github.com/nasa/fprime/blob/devel/AI_POLICY.md
• https://openinfra.org/legal/ai-policy
• https://github.com/pola-rs/polars/blob/main/AI_POLICY.md
• https://github.com/beyond-all-reason/Beyond-All-Reason/blob/master/AI_POLICY.md
• https://about.gitlab.com/community/contribute/dco-cla/

[rajatchopra]

Thanks for the initiative @AkshatDudeja77.
Closing this PR at this point. Windows is a non-starter for a lot of code here already. We need a bigger redesign of code to let any functional subset to work.
Just not allowing the sighandlers to be registered for windows is noisy. The right way would be to have an interface implemented for linux and others separately and plugged in appropriately for the platform in question.

Feel free though to re-open if you think otherwise.