Reflectometry: Disable mode save

[John-Holt-Tessella]

As an instrument scientist if I restart my reflectometry server I want it to recover to disable mode if that was the mode I was in, including the incoming beam positions.

Acceptance Tests

1. On enter disable mode all component incoming beam are saved to disk in autosave area
2. On restart if in disable mode and file exists it is loaded
   1. All components are disabled
3. On restart if in disable mode but no file exists
   1. error
   2. all beam paths set to incoming beam path
4. Test values by:
   1. setting super mirror and theta angles in Polarised NR mode
   2. change to disable mode
   3. move super mirror, theta and one offset
   4. record motor positions
   5. Restart IOC
   6. Move back to position given
   7. motors should be at same place

[aaron-long]

• Can you please clarify more explicitly the test steps. Specifically what you mean by "Move back to position given".

We should wait for feedback on what the exact desired behaviour is for theta in disabled mode.

[ThomasLohnert]

Sorry, something is wrong when exiting disabled mode after making changes - I think the beam path needs to be re-evaluated at this point.

Steps to reproduce (assuming SURF-like config):

1. Set beamline positions to 0
2. Enter PNR mode
3. Move sm_angle to 1
4. Enter Disabled mode
5. Move sm_angle to 0
6. Go back to PNR mode
   • What I would expect: Motors stay where they are, parameter readback values of components further along the beam update to show they are offset from the beam (equivalent to 1 deg)
   • What happens: Motors stay where they are, parameter readback values continue to show 0. They can be forced to update to the correct value by triggering a monitor update (e.g. move any motor axis used by the reflectometry IOC)

[John-Holt-Tessella]

Pull requests:

• Add super mirror to test and run polarised test EPICS-IOC_Test_Framework#302
• Ticket3897 disable mode save EPICS-inst_servers#265

[ThomasLohnert]

The unittest test_GIVEN_beam_line_with_disable_autosave_position_WHEN_init_THEN_incoming_beams_set_correctly_on_start is now failing.

The failure message is a bit cryptic (TypeError: coercing to Unicode: need string or buffer, PositionAndAngle found) but I think it is just because the actual PositionAndAngle from the result is different from the expected (y=0, z=0, angle=0) when it should be (0, 1, 0). Could you please fix it ?