Don't Load Any IDFs Until We Know Which Version of the Instrument Should Be Used

[Peter Parker]

The ISIS SANS system tests currently load IDF's without knowing which version of the instrument the data being reduced corresponds to. LOQ has for some reason been ​hard-coded to use the second-to-last (chronologically speaking) of the four available IDF's by default, and since #9361, SANS2D defaults to the *oldest* of two IDF's.

To further complicate matters, the ISIS SANS interface actually waits until some data has been loaded before reading IDFs (and so is able to pick the correct one), except in the case where users go straight to the Masking tab and click "Display Mask" without first loading data. In that case, the defaults mentioned above are also used.

This is a mess.

My preference here would be to get rid of these arbitrary defaults and instead simply enforce a universal rule that no IDF-dependant ISIS SANS command can take place without first knowing which data file (and therefore which exact IDF) is being used. This would be the cleanest (and I'd also argue, easiest) way to guarantee that the correct IDF is being used at any given time.

I need to run this by the scientists, but I think they will be in general agreement.

[Peter Parker]

Note to self:

We have the following line in lot's of our system tests

self.disableChecking.append('Instrument')

This is likely because we're using slightly different instruments at various parts in the reduction (the hard-coded or default IDF at the beginning, the one loaded with the data by the end). I'm pretty sure that by the end of this ticket we'll be able to remove this.

[Peter Parker]

(In #8308) See updated ticket description, and ignore previous comment from Gesner.

[Stuart Campbell]

This ticket has been transferred to github issue ​10210