add test-dds-adapter-depth test

[maloel]

Adds basic streaming test for rs-dds-adapter.

Moved the librs library from dds-specific code to rspy, so other tests can use it.

Tracked on [LRS-1217]

[OhadMeir]

Moved the librs library from dds-specific code to rspy, so other tests can use it.

I am not sure librs is the appropriate name for this in rspy scope. Was named librs because it was a helper to test librealsense DDS capabilities (also testing realdds in same tests folder). Maybe incorporate it into devices.py?
What do you think?

[OhadMeir]

Thought about it, we have a live/tools unit test folder. I think the adapter test should be there. Maybe under an internal dds folder

[maloel]

Moved the librs library from dds-specific code to rspy, so other tests can use it.

I am not sure librs is the appropriate name for this in rspy scope. Was named librs because it was a helper to test librealsense DDS capabilities (also testing realdds in same tests folder). Maybe incorporate it into devices.py? What do you think?

It's librealsense, with everything that it has, plus some additions for unit-testing (but not for dds). Has nothing to do with dds.

[maloel]

Thought about it, we have a live/tools unit test folder. I think the adapter test should be there. Maybe under an internal dds folder

I don't agree: everything for dds should be under dds. The 'live' folder is not always right, and using it to group things does not always make sense. That's why we have --live and --not-live as switches for run-unit-tests. It was a good grouping at the time, but it's no longer needed.

If you insist, I'll change.

[OhadMeir]

Thought about it, we have a live/tools unit test folder. I think the adapter test should be there. Maybe under an internal dds folder

I don't agree: everything for dds should be under dds. The 'live' folder is not always right, and using it to group things does not always make sense. That's why we have --live and --not-live as switches for run-unit-tests. It was a good grouping at the time, but it's no longer needed.

If you insist, I'll change.

This test should fail if no USB camera is attached. How is it marked to run only with live cameras now?

[maloel]

Thought about it, we have a live/tools unit test folder. I think the adapter test should be there. Maybe under an internal dds folder

I don't agree: everything for dds should be under dds. The 'live' folder is not always right, and using it to group things does not always make sense. That's why we have --live and --not-live as switches for run-unit-tests. It was a good grouping at the time, but it's no longer needed.
If you insist, I'll change.

This test should fail if no USB camera is attached. How is it marked to run only with live cameras now?

--live does not check the tag or folder - it checks whether it has a test:device directive in it.

[maloel]

--live does not check the tag or folder - it checks whether it has a test:device directive in it.

Not sure you saw this... I got no email...

[OhadMeir]

OK if --live is looking for device directive, but I don't see that your test is running. It is skipped both for GHA and LibCI.

[maloel]

I don't see that your test is running

Good catch.
The reason is that this test is live and DDS at the same time, and we have no setup that runs both together...
We have to make LibCI compile DDS (possibly only on nightly?) @Nir-Az any preference?

[Nir-Az]

I don't see that your test is running

Good catch. The reason is that this test is live and DDS at the same time, and we have no setup that runs both together... We have to make LibCI compile DDS (possibly only on nightly?) @Nir-Az any preference?

Did a test run, build time increase from 36 [m] to 45[m]
For nightly only it's OK

1 thing I am not sure from the code is if dds on will run only non live test or both live and not live?

[maloel]

1 thing I am not sure from the code is if dds on will run only non live test or both live and not live?

I will make it so that, on nightly only, LibCI will compile DDS. This will be a configuration in Jenkins.
This will also cause the nightly to run all tests -- not just live -- i.e., including DDS -- and this test would then run.

[maloel]

I had a lot of trouble - the adapter test failed in LibCI under Jenkins, and Dima could not help. It turns out that the std::cin.ignore( ... ) code doesn't work under jenkins: cin reports it's at EOF and the adapter quits right away.

To get around this, I implemented proper Control-C handling in Windows/Linux and added it to rsutils.

You can see the run here:
https://rsjenkins.iil.intel.com/job/LRS_libci_pipeline/9859/

[OhadMeir]

I will make it so that, on nightly only, LibCI will compile DDS. This will be a configuration in Jenkins. This will also cause the nightly to run all tests -- not just live -- i.e., including DDS -- and this test would then run.

Is this handled?

[maloel]

I will make it so that, on nightly only, LibCI will compile DDS. This will be a configuration in Jenkins. This will also cause the nightly to run all tests -- not just live -- i.e., including DDS -- and this test would then run.

Is this handled?

As soon as the PR is in. It's a change in Jenkins.