Omr: Make naming of CodeGenerator variable consistent

[0xdaryl]

The overwhelmingly common short form for a variable to store the TR::CodeGenerator object is cg (about 30000 instances). The next most common form is codeGen (about 400). We should use a single form for consistency and given the popularity of cg that should be the preferred name. Sweep through the OMR compiler and make that change.

Here is a pipeline of greps to help find some of the places to change:

grep -rn codeGen * | egrep -v 'codeGen[a-zA-Z0-9(]'

[pkrc267]

I would like to work on this. Can you please assign it to me?

[0xdaryl]

Thanks for your offer to help!

[pkrc267]

Thankyou so much 🙂

A quick question. After renaming all the instances, should I build and test the whole project?

[0xdaryl]

Generally, it is a good practice to test your changes to the extent that you can on your own before making a pull request (for example, on x86 Linux). The community certainly understands that you can't test them on every platform and architecture combination.

When you do make a PR (or any time that you push new commits), a Travis CI job will run automatically that will perform a build (x86 Linux only). Once a committer has reviewed your code, they will run a much broader set of sanity tests on all architectures the OpenJ9 community cares about that you will see in the "Checks" section of the PR.

[pkrc267]

Made the names consistent and made a PR.
@0xdaryl Please give feedback if I need to make certain things better or do anything different.