2M & 4M undulators

[prjemian]

• close 2M insertion device (only 01DS) #975
• close 4M insertion device (only 11US and 11DS) #976

NOTE: 2M & 4M -- No PV for Done Moving, rely on Busy and set done_value=0 instead.

@cooleyv
@cpchuang

[prjemian]

Random and unrelated EPICS CA connection timeout problems in testing.

[prjemian]

On 3rd (or 4th?) re-run of all tests, they passed. Might adjust master timeout if failures continue:

apstools/apstools/tests/__init__.py

Line 12 in ded5219

MASTER_TIMEOUT = 3

Applied here:

apstools/apstools/tests/__init__.py

Lines 27 to 32 in ded5219

	EpicsSignalBase.set_defaults(
	auto_monitor=True,
	timeout=MASTER_TIMEOUT,
	write_timeout=MASTER_TIMEOUT,
	connection_timeout=MASTER_TIMEOUT,
	)

[prjemian]

Testing is more stable with the longer master timeout. 3s was borderline to encounter EPICS CA connection timeouts.

[prjemian]

@MDecarabas Now, this is ready for review.

[MDecarabas]

Small question about if we can just reuse the 2m undulator class for the 4m as well. Might have missed something but it does not look like any of the code is different

[MDecarabas]

Isn't this the same content with just a different class name?

[prjemian]

Made 4M as a subclass of 2M

[prjemian]

Thought about this, too. Similar to the STI_Undulator? Once (or if) we identify differences, then we can break the dependency.

[MDecarabas]

Thought about this, too. Similar to the STI_Undulator? Once (or if) we identify differences, then we can break the dependency.

Are you saying as the class will grow you expect to see differences?

[prjemian]

I'm saying that as the class is used at the beam lines, we might discover signals needed by one but not the others.