INSTRUMENT_Parameters.xml no longer loaded by default

[Anders Markvardsen]

1. E.g. open any HRPD data say the recent HRP58468.nxs from cycle_13_4

2. plot spectrum 100 and try to fit a BackToBackExponential to any peak in this spectrum. The parameters of this fitting functions (A, B and S) have not be updated according to information in HRPD_Parameters.xml (i.e. there values are 1.0, 0.05 and 1.0 respective)

3. However if LoadParameterFile is used to load HRPD_Parameters.xml explicitely then the information is loaded in this file, which can be verified by repeating step 2 above a note that A, B and S are different from 1.0, 0.05 and 1.0 respectively.

I tested the following downloaded version

Most recent build as of this ticket DOES NOT WORK

Nightly build date 2nd Dec 2013 DOES NOT WORK

Stable release 3.0 DOES NOT WORK

Stable release 2.6 DOES WORK

Hence this defect has been introduced somewhere between release 2.6 and release 3.0.

Further tested that the same behaviour is observed for GEM.

As part of this ticket or as a separate ticket please add systemtest which tests that INSTRUMENT_Parameters.xml gets loaded. For example:

1. Load HRPD dataset

2. Adds a BackToBackExponential fitting function to fit spectrum 100

3. Check that the A, B and/or S fitting parameters of this function are different from the hardcoded values, which are 1.0, 0.05 and 1.0 respectively

[Owen Arnold]

The following issue must be fixed before release

run1 = Load('POLREF00008481.nxs')
print run1.getInstrument().getNumberParameter('LambdaMin') #Prints an empty list even though LambdaMin is definied in the POLREF_Parameters.xml file!

'''
Explicitly calling LoadParameters with the target file fixes the problem. My guess is that either there is no attempt to load any parameter file, or that the parameter
file selection is not working correctly.
'''

[Anders Markvardsen]

Note the problem is not widespread, in more detail the problem may only be when ISIS histogram data are loaded, at least just observed that parameters are loaded for ISIS Event data

[Anders Markvardsen]

Changed from blocker to critical since now identified that code has been broken just for the case of ISIS histogram nexus files, for all other cases it works, however this is still a critical bug

[Anders Markvardsen]

Make isis hist nexus read params again. re #8608

Changeset: f3e33dedd4ffac8a7b7b9f55099bdbe329c3644a

[Owen Arnold]

Replying to Anders Markvardsen:

Changed from blocker to critical since now identified that code has been broken just for the case of ISIS histogram nexus files, for all other cases it works, however this is still a critical bug

I think this should be a blocker. I agree that it might not affect as many instruments as we initially thought, but we wouldn't want to release with it in this state.

[Anders Markvardsen]

Revert "Make isis hist nexus read params again. re #8608"

This reverts commit f3e33dedd4ffac8a7b7b9f55099bdbe329c3644a.

Changeset: b2d15cf681e398a1ede3509f4e27e0c67222ee91

[Anders Markvardsen]

Ensure sample and log loading as before. re #8608

Changeset: d01b1fa9b327ab5e8614eeb580efac15b4f0aec0

[Anders Markvardsen]

To test run something like:

run1 = Load('POLREF00008481.nxs') print run1.getInstrument().getNumberParameter('LambdaMin') #Prints an empty list even though LambdaMin is definied in the POLREF_Parameters.xml file!

I.e. load any ISIS HISTOGRAM .NXS file which have an associated INST_Parameters.xml file and check that parameters are loaded.

Separate tested added for systemtest: #8704

[Keith Brown]

Parameters are loaded. Has been checked with other parameters and instruments

[Keith Brown]

This has been merged, but the macro didn't post here it seems

[Stuart Campbell]

This ticket has been transferred to github issue ​9452