focusTel issues after #23

[EastEriq]

After 093f632, about #23:

1. focusTelInSlave generates plots without fitting parabolas. Original feature or new issue? AFAIR parabolas used not to be drawn, or drawn multiple times, in corner cases with outliers, but it doesn't seem the case here
   
2. it is reported that FWHM is no more updated after the focusLoop has been run, and the values shown are the last ones of the loop, even after having observed several targets with obsByPriority2
   
   Here the new code was in effect only on Unit 10.

[EastEriq]

S.queryCallback("Unit.Slave(1).Responder.query('Unit.Camera{1}.ComputeFWHM')",10) reports 'omit' instead of the expected default 'last'

[PolishookDavid]

Please also add units (arcsec, focus position) to the axes labels

[EastEriq]

For the run of yesterday, to which the images refer, pngs were also not created in ~/log/focus_plots. I suspect that focusTelInSlave did not run to the end because of some error, and didn't restore .ComputeFWHM. I have only the master log which says:

Sent mount to focus coordinates
Focusing telescope 1 FAILED!
Focusing telescope 2 FAILED!
Focusing telescope 3 FAILED!
Focusing telescope 4 FAILED!
{17:13:48.750|obs.unitCS[10]} Checking definitions and connections of unit 10:
{17:13:58.771|obs.util.Messenger[10_slave_1.Messenger]} 10_slave_1.Messenger timed out waiting for a reply to "true"
{17:13:59.038|obs.unitCS[10]} Slave 1 status: "notresponding"
{17:14:09.053|obs.util.Messenger[10_slave_2.Messenger]} 10_slave_2.Messenger timed out waiting for a reply to "true"
{17:14:09.335|obs.unitCS[10]} Slave 2 status: "notresponding"
{17:14:19.357|obs.util.Messenger[10_slave_3.Messenger]} 10_slave_3.Messenger timed out waiting for a reply to "true"
{17:14:19.621|obs.unitCS[10]} Slave 3 status: "notresponding"
{17:14:29.639|obs.util.Messenger[10_slave_4.Messenger]} 10_slave_4.Messenger timed out waiting for a reply to "true"
{17:14:29.902|obs.unitCS[10]} Slave 4 status: "notresponding"
{17:14:29.903|obs.unitCS[10]} check failed!
{17:15:11.899|obs.util.Messenger[10_slave_3.Messenger]} 10_slave_3.Messenger receiver reports: The logical indices in position 1 contain a true value outside of the array bounds.
{17:15:18.475|obs.util.Messenger[10_slave_1.Messenger]} 10_slave_1.Messenger receiver reports: The logical indices in position 1 contain a true value outside of the array bounds.
{17:15:30.953|obs.util.Messenger[10_slave_4.Messenger]} 10_slave_4.Messenger receiver reports: The logical indices in position 1 contain a true value outside of the array bounds.
{17:15:42.515|obs.util.Messenger[10_slave_2.Messenger]} 10_slave_2.Messenger receiver reports: The logical indices in position 1 contain a true value outside of the array bounds.

[EastEriq]

I thought that Slaves were spawned by default with a listener MasterMessenger, but they weren't. .focusTel was therefore not blocking, causing .operateUnit premature continuation commands while .focusTelInSlave was running, and mixup. Most likely, a checkWholeUnit which detected no response from Slaves and thus respawned them.
Formerly operateUnit was using .checkFocusTelSuccess which was sitting there, polling the focus log, but removed it. Let's see if the corrections are enough.

[EastEriq]

What would remain to do here are embellishments to the plots, either those generated on the fly by .focusTelInSlave, which I want to suppress (because the X channel prevents disowning, see EastEriq/LAST_Messaging#4), or those generated a posteriori by obs.util.observation.plotFocusData. The former include parabolic fits without or with outliers depending on the case, with arbitrary threshold parameters, whose concept may change if the focus algorithm is changed, as it is badly called for. I think therefore that at this stage is sufficient only to highlight the best value estimated, without graphical hints of how.

Btw re: the original

1. focusTelInSlave generates plots without fitting parabolas.

that was just because due to what said in #36 (comment), the focus loop running was killed before completion.