Create OMR::Node::storeToAddressField() helper

[gita-omr]

• create a helper for a common but not trivial code sequence

[gita-omr]

Removed unnecessary change.

[gita-omr]

Removed unnecessary change.

[gita-omr]

@hzongaro kindly agreed to review.

[hzongaro]

I had a couple of minor comments. My only real concern about this is proposal is whether adding this as a method in OMR::Node is appropriate. It feels like a handy utility method that should be added elsewhere.

I'd like to ask @0xdaryl for his thoughts about it.

[gita-omr]

I had a couple of minor comments. My only real concern about this is proposal is whether adding this as a method in OMR::Node is appropriate. It feels like a handy utility method that should be added elsewhere.

I'd like to ask @0xdaryl for his thoughts about it.

We considered adding it to J9::TransformUtil but since it just returns a node it seemed more appropriate for it to be one the static factory functions in TR::Node.

[0xdaryl]

While there may not be a lot of application of this API within OMR itself, downstream projects might benefit from it. I think it is OK to have this function in the Node class because it is using the right APIs for determining write barrier usage and it is properly documented.

My main concern is the use of TR::comp() in this code. I realize that pretty much every other static TR::Node function uses TR::comp(), but they shouldn't. Instead, they should accept the compilation object as a parameter (see #4106). While you are probably justified in using it for symmetry until the other functions are modified, you are leaving a bit of compile-time performance on the table in exchange and I would re-consider your use.

[gita-omr]

Added TR::Compilation *comp parameter.

[hzongaro]

Looks good. May I ask you to squash the two commits, and add a sentence to the commit message describing the motivation for the addition? Thanks!

[gita-omr]

Squashed and added a description.

[hzongaro]

Jenkins build all

[hzongaro]

A number of .exe files appear to be missing from the Windows build, but it's not clear why. It doesn't look like they could be related to this change. Rerunning. . . .

Jenkins build win