Fix corrupt analog data for non-float values

[taligentx]

Hi,

This fixes an issue that causes invalid data in SmuView for non-float analog data. I ran into this when writing a driver for serial-dmm where the value is stored as a double. feed_in_analog() is getting the original value converted to float, but then push_interleaved_samples() is sending the original unit size (8 for double) instead of the float size and results in corrupt data.

Let me know if there are any issues or if there's another way to tackle this - thanks!

-Nikhil

[knarfS]

It took some time, but I was finally able to merge your commits:

• "Fix analog data handling for non-float values" with commit 114a934
• "Fix segfault when adding signal to data table with no signals loaded" was merged with commit 744a561

I changed the commit messages slightly.

Thank you for contributing