fix: validate TSP loadscript write spans

[AjTheSpidey]

This fixes #500 by checking the full loadscript command before it is sent to the instrument. The existing single-write behavior is kept, but the code now rejects any newline-delimited span longer than the TSP input limit, so a script line cannot run past the required write termination point. The check happens before the old script is deleted, so an invalid script body does not leave the device half-updated.

I added coverage in the SMU test for the 1000-character boundary and for rejecting the over-limit case before a write is sent. I also updated the unreleased changelog.

Tested locally on Windows with Python 3.14:

• poetry run pytest tests/test_smu.py -q
• poetry run pytest -q -k "not test_docs"
• poetry run ruff format --check .
• poetry run ruff check .
• poetry run pyright

I also tried pre-commit run --all-files, but it did not finish locally even after a longer timeout, so I did not count that as passing.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[nfelt14]

This seems to be a duplicate of #581

[AjTheSpidey]

Closing this since #581 already covers it.