MUON: save user specified deadtime table filename

[Anders Markvardsen]

When a deadtime table file is selected in Muon interface it is lost when interface opened next time.

I.e. in this ticket save this

See also comment 1 and 2 in #6613 (I accidently pushed to the wrong ticket hence swapped content)

[Nick Draper]

Moved to r2.6 at the end of r2.5

[Nick Draper]

Moved to the Backlog after R2.6

[Arturs Bekasovs]

While investigating this one, found that the whole Dead Time file thing is not working properly. Ticket for that: #7955.

While it's theoretically possible to implement this one, it would be difficult to verify it before the logic is fixed, so doing that first.

[Arturs Bekasovs]

(In #7955) It seems that the last file found is used instead of the current one loaded.

[Arturs Bekasovs]

Trac issues. Returning blocked by ticket.

[Arturs Bekasovs]

Save/load code for the Dead Time filename.

Refs #6691

Changeset: e747eac4431322aa56de8ff41e33903c87220f8b

[Arturs Bekasovs]

Don't reset file to empty when selecting "Dead Time from disk".

Improves the UX in the way that if we have not changed the filename, it will always contain what's saved i.e. previously applied.

Refs #6691

Changeset: 1d8583fb20f689f1b090fe3572d959b091d985a0

[Arturs Bekasovs]

Removed.

[Arturs Bekasovs]

Return the old behavior of clearing input.

Not clearing turned out to introduce more problems. The old behaviour is fixed as well, as before it was just setting the text without updating the file cache.

Refs #6691

Changeset: eb9578213ef5bf973a1b9eda31d70f7590135f19

[Arturs Bekasovs]

Save the file even if it doesn't apply to the current data set.

Refs #6691

Changeset: 510fc73a75621b7c2d2f34ab6cda2971e9fcfb43

[Arturs Bekasovs]

Tester:

Use attached EMU00036854.nxs (data) and dtepar.nxs (Dead Time table). Check that the specified name of Dead Time Table file is saved when the interface is closed and opened again.

Make sure that if the filename is restored it is actually used when data is loaded. One way to do that is to check the history of MuonAnalysis workspace and make sure ApplyDeadTimeCorr was used preceded by LoadNexusProcessed of the appropriate Dead Time file.

Please also check the following scenario:

1. Plot EMU00036854.nxs with Dead Time correction applied from dtepar.nxs.
2. Set correction to "None", plot again.
3. Set correction to "From Disk". Try to plot. Should throw an error, as no file is specified.

[Arturs Bekasovs]

Data set for testing

[Arturs Bekasovs]

Dead Time Table file for testing

[Karl Palmen]

All details of test went OK, but I could not shut down Mantidplot afterwards and needed to use task manager to terminate Mantidplot. The deadfile name survives this. Mantidplot closes OK if I open it and do something other than the test.

[Arturs Bekasovs]

Is not reproducible on my Linux machine, so presumably is a Windows problem.

[Arturs Bekasovs]

Close all MDI windows explicitly when exiting.

Do it at the point when MantidPlot is still alive to avoid segfaults.

Refs #6691

Changeset: ccb5ffeadc6336161643c1867182f0b58596468b

[Arturs Bekasovs]

There was a segfault thrown when we close MantidPlot with MuonAnalysis open, so presumably it was the cause of problems.

[Karl Palmen]

It still does not shut down on Windows. Are all changes pushed?

[Arturs Bekasovs]

That's right. Forgot to git finish. Will do when I get back on Wednesday. Sorry.

[Arturs Bekasovs]

Should be fine now. Peter confirmed that my changes solved the problem on the Develop, so it was just a matter of pushing this branch.

[Karl Palmen]

Merge remote-tracking branch 'origin/feature/6691_muon_save_deadtime_file'

Full changeset: ​988413ffe6d9caf388506e61a7da5ca0453f3e6d

[Stuart Campbell]

This ticket has been transferred to github issue ​7537