Stream2 fix

[MarkRivers]

Implements code changes described in #79

[domitto]

Small but important detail: addr 1 is already in use by the Monitor interface in the current driver (see eigerDetector.cpp:L1849 —

ADEiger/eigerApp/src/eigerDetector.cpp

Line 1849 in b586e87

doCallbacksGenericPointer(pImage, NDArrayData, 1);

Reusing it for threshold callbacks would collide with existing monitor consumers. We probably need to keep addr 1 reserved and offset the threshold addresses, or move the monitor stream elsewhere.

[ericonr]

Small but important detail: addr 1 is already in use by the Monitor interface in the current driver (see eigerDetector.cpp:L1849 —

ADEiger/eigerApp/src/eigerDetector.cpp

Line 1849 in b586e87

doCallbacksGenericPointer(pImage, NDArrayData, 1);

Reusing it for threshold callbacks would collide with existing monitor consumers. We probably need to keep addr 1 reserved and offset the threshold addresses, or move the monitor stream elsewhere.

I think the property of 1st threshold in address 1 is simpler to think about than having to offset by 1 every time, so I would suggest moving the monitor interface to a large offset (something like addr=10?). That way, this offset range can also be used if any other kinds of streams are added (addr=11, ...).

[tguruswamy]

I tested the functionality with an Eiger2 Si 500K, firmware version 2024.1. This model has 2 thresholds.

Basic operation works fine, I could see 2*NumImages frames on array address 0, and NumImages frames on array address 1 & 2.
Conceptually I think this is the right approach, as it more closely mirrors the structure of the source stream, which is basically interleaved frames.

The collision of address 1 with the existing Monitor interface (if both are enabled, frames from both sources are received on 1) is the only thing I see still to be resolved.

Note that in contrast, the FileWriter mode's new HDF format "v2024.2" is using a 4-D array, always -- I've opened #83 to track progress implementing support for reading that.

While testing I also found the other new feature StreamAsTSSource doesn't seem to behave correctly, but that appears unrelated to this PR (reopened #77).