Make CheckWorkspacesMatch work with MDWorkspaces

[Michael Reuter]

The algorithm CheckWorkspacesMatch should be able to handle MDWorkspaces.

[Michael Reuter]

Refs #5292. Starting MD checking work.

Changeset: 7a7db90b8fe51e724b54ed92a84624bc2bcd41f5

[Michael Reuter]

Refs #5292. More checks for MDHistoWorkspaces.

Changeset: c29cc3def9e25c126423f204aee8ee821797cdf3

[Michael Reuter]

Refs #5292. Adding checks for MDEventWorkspaces.

I've added checks for MDEventWorkspaces. There is one check that is currently commented out since I need to find out how some autogenerated code was created before being able to use this new function.

Changeset: 7ded05e70002517f87c67852737bd5fba21bb9ff

[Michael Reuter]

Refs #5292. Doing checking in a much easier way and adding final checks.

Thanks to Owen for pointing out the CompareMDWorkspaces algorithm. Makes the code for checking much cleaner.

Changeset: a99ea235f0611eda1ffa964fd1ea40497b363cb0

[Michael Reuter]

Refs #5292. Fix weird success message.

Changeset: c6e8ec5653f7fda97db4e816f037001ea513ef69

[Russell Taylor]

Re #5292. Clear intel compiler warnings.

Changeset: ae874bc83f9aec47afdd7605e6fac38c9fc655c3

[Michael Reuter]

The SNSConvertToMDEventsTest system tests are using this nightly. To test, load an MDEventWorkspace, clone it and run the algorithm.

[Owen Arnold]

I tried this, but I think the default tolerances are not set correctly:

A = CreateMDWorkspace(Dimensions='4',Extents='0,1,0,1,0,1,0,1',Names='A,B,C,T',Units='U,U,U,S')
FakeMDEventData(InputWorkspace='A',UniformParams='100000',RandomizeSignal='1')
B = CloneMDWorkspace(InputWorkspace='A')
result = CheckWorkspacesMatch(Workspace1=A, Workspace2=B)
if not result == "success":
        print result
        raise RuntimeError("Expected result to be 'success'")

[Michael Reuter]

The issue looks like it is a problem with CloneMDWorkspace. The CheckWorkspacesMatch algorithm just uses CompareMDWorkspaces directly. Running the output of the above script generates the same error. I have opened a ticket (#5697) against this. One could change the default tolerance to 1.0e-9 to get it to pass, but this algorithm is heavily used by the system tests, so I'm opting for not doing that.

[Stuart Campbell]

This ticket has been transferred to github issue ​6138