Performance tweaks and a few bug fixes contributed by Autodesk and an ObjexxFCL update

[DeadParrot]

This is a small set of contributions from Autodesk:

• Performance refactoring in a few hot spots
• Bug fixes for undefined variables detected by Clang undefined santizer runs
• ObjexxFCL updates including an Array1D 2-iterator constructor
• GCC std lib debug mode define added for debug GCC builds (for detection of things like std::vector bounds violations)

[Myoldmopar]

@DeadParrot These changes look really good and clean. CI is so happy... 😉 I'm OK with this merging at will once it comes back all clean unless you have additional work.

[DeadParrot]

@Myoldmopar Thanks. I think given the 8.4 release window this is a good place to stop and merge so no additional commits expected.

• The GCC debug flag probably uncovered a bug in that one Ubuntu case failure: Let me know if you'd like me to debug it as part of this PR
• I had added the Defect tag because there are some bug fixes in here but it is fine with me to drop it since they are ~minor

[Myoldmopar]

@DeadParrot Thanks for the response. That failure almost flew past me hidden up there. Interesting failure though:

Error copying file "/home/nrelci/ci/builds/EP-29ddf1242f-x86_64-G4.8ICD/build/Energy+.idd" to "/home/nrelci/ci/builds/EP-29ddf1242f-x86_64-G4.8ICD/build/tst/api_callback/Energy+.idd".

Then the vector subscript error after that. I'm building locally to test it out.

As for the defect tag, the way our changelog is created, the script doesn't want to decide which box to put a merged pull request into, so right now it requires exactly one of five tags on every merged pull request. I'll probably improve that one way, but for now it is working since usually new features are defects are separate (but not always like this one).

[Myoldmopar]

Looks like this may have had an issue for a while and this just revealed it. I'm testing some solutions locally. @DeadParrot If I figure it out, can I push directly to this branch or do you want me to do it separately then you pull those changes into this branch?

[DeadParrot]

@Myoldmopar OK, I'll try to remember to pick just the best tag for each PR. Normally I'd try to avoid mixing performance and bug fixes -- the fixes just fell out of testing for this branch so I bundled them in since it became a kind of grab-bag branch anyway.

Yes, I think the GCC debug flag I added uncovered an existing vector bounds violation. I'm fine with you pushing to this branch if you have a fix. Thx.

[Myoldmopar]

Problem Found

@DeadParrot OK, I found the problem. It's because when I call E+ as a DLL, I created a null set of command line arguments to pass to the command line argument processor. (The processor is still called to ensure all the right flags are set up inside E+.) However, the command line processor doesn't actually understand this zero-argument condition, and attempts to access the first entry that would normally always be there.

Problem Solved

I made a change in a commit I am pushing shortly that initializes the command line arguments with a single argument as would be expected when calling the program with only the program name and no subsequent arguments. I also had to update the path to the IDD because it couldn't find the idd to copy into the directory.

Solution Review?

• @nealkruis Could you verify the changes in 66d52ae make sense to you for the CLI option parser you added? I ran it locally and it now works, where I had confirmed it was broken before the fix.
• CI will also verify this...

Wrap up

If @nealkruis likes the changes, and CI comes back clean, this should be about ready. @DeadParrot it would seem a merge conflict has arisen, so that will ensure the branch is up-to-date anyway, and we can merge.

[DeadParrot]

@Myoldmopar Great. Do you want me to fix the merge conflict?

[Myoldmopar]

Yes, please. I'm off fixing other merge conflicts as we speak. It may be super minor on this branch, I didn't do an attempt.

[DeadParrot]

@Myoldmopar Merge conflict cleaned up.

[Myoldmopar]

Thanks @DeadParrot, this should be good to go once CI is happy.

[nealkruis]

@Myoldmopar looks good to me.

[Myoldmopar]

Thanks @nealkruis. OK, @DeadParrot, if I see CI coming back happy after your merge commit, I'll merge this tonight.

[Myoldmopar]

And there's the result I was looking for the most. IntegrationCoverage now runs all green. Should be mergeable this evening.

[DeadParrot]

@Myoldmopar Great. I assume you understand the HeatPumpProportionalControl_DCV diff (extra alpha field?).

[Myoldmopar]

I assume that will go away now that the branch is up to date with develop.

[Myoldmopar]

Yep, all green there :)

[Myoldmopar]

@DeadParrot could you rename the pull request so that users understand what came in with the changes? It's merged, thanks!

[DeadParrot]

@Myoldmopar Thanks! Renamed.