3126: Fix OpenCl and covariance issues

[krzywon]

Description

Any change to a preference was triggering an OpenCL environment reset, which was resulting in an OpenCL error when rerunning fits. This also fixes the tuple out of range issue noted by @wpotrzebowski in #3122 (comment).

Fixes #3126

How Has This Been Tested?

Please describe the tests that you ran to verify your changes. Provide instructions so we can reproduce.

Review Checklist:

Documentation (check at least one)

• There is nothing that needs documenting
• Documentation changes are in this PR
• There is an issue open for the documentation (link?)

Installers

• There is a chance this will affect the installers, if so
  • Windows installer (GH artifact) has been tested (installed and worked)
  • MacOSX installer (GH artifact) has been tested (installed and worked)

Licencing (untick if necessary)

• The introduced changes comply with SasView license (BSD 3-Clause)

[wpotrzebowski]

I've tested the Mac installer. The issue I reported earlier seems to be solved but it hangs on building documentation. (I gave up after 30min and forced it to quit).

[wpotrzebowski]

It is strange but model documentation works, however "general help" is hanging forever.

[krzywon]

That shouldn't be triggering at that point. The doc generation should only be active for plugin models. I'm guessing this is on Mac?

[krzywon]

Merging this for now while I investigate the documentation issue

[krzywon]

Merging this for now while I investigate the documentation issue

Never mind - no approvals yet

[wpotrzebowski]

Tested the functionality, and it works fine. I cannot asses code quickly - sorry.

[lucas-wilkins]

So, it's my understanding that you've made it so GPU options go through the same kind of staging process as the other options. This seems sensible.

It looks like this will stop random errors from other preferences, but it will still cause errors if you change the GPU options. If this is acceptable, then it seems fine,

I've added some other changes, mostly stylistic. I would very much like to see the bias one done, and the reshape one would be much neater.

I'm going to approve so things are not held up though.

[krzywon]

Thanks for the review, @lucas-wilkins. I incorporated most of your suggestions. Have one last look and merge if you think it is ready.

[smk78]

Functionality test of actions/runs/11467760172 on W10/x64.
Tried single parameter fits with/out OpenCl and with/out DREAM. No sign of the errors reported by @wpotrzebowski . LGTM.