[WIP] Fix LeCroy WaveSurfer 3k

[qberthet]

Attempt to fix the support WaveSurfer 3k serie of scope.

Initially reported in #996 (and then partially reported in #1026 and #1028).

Most annoying issue is fixed in the current state : averaging is not present per channel but under app.Acquisition.Horizontal.AverageSweeps. For now I have naively updated this path for the WS3k, but I suppose that this might cause issues if ngscopeclient expect an average setting per channel. Seems to work tho, but eager to implement a nicer fix once all the discrepencies with other LeCroy will be identified.

Probe detection seems fixed, it was caused by abscence of probe mux being reported in differently than in other LeCroy scope: no .ActiveProbe property (and a differentt property for probe name).

Sample rate cannot be set dirrectly on these scope, so configuration is set by adjusting the horizontal scale.

Still investingating:

• Some combination of sample rate and memory depth are not working properly: the scope and app report different values

[azonenberg]

What's the status of this PR? We're planning to cut v0.1.1 soon and if you have bug fixes, even if not all-inclusive, for the WS3K I'd like to get them merged then you can send additional fixes in a second PR aimed at v0.2.

[qberthet]

Yes, in its current state, WS3k is usable. There are still a few quirks I’d like to investigate: changing the sample rate and memory size sometimes interact in unexpected ways and can occasionally push the scope into roll mode. I’m happy to address these in a follow-up PR. Having more testers would also be very helpful.

I haven’t been able to test this against other LeCroy models, but based on the code paths involved, it’s unlikely that anything else was affected. That said, please feel free to point out any areas where the implementation could be improved.

One additional note for transparency: I only became aware yesterday of your position regarding AI-generated code (in scopehal-apps code of conduct), which I fully understand and respect. This PR is not fully LLM-generated. However, I did use AI as an autocomplete and spell-checking tool. I can justify and explain every line of this PR, but I prefer to be upfront about this now, as I hope to continue contributing original code to this project in the future.