machines: simplify MACHINEOVERRIDES definitions

[zach-welch-aquabyte]

Prepending new definitions should be preferred to assignment, as it is simpler and more conducive to defining new machine configurations that reuse these configurations.

I posted an initial draft of this patch (with the required Signed-Off-By line) to the yocto mailing list as well, but I thought a PR might elicit a more expedient response.

Also, I tested this on the kirkstone branch with the raspberrypi{3,4}-64 machines included in my own custom machine configurations, but these changes should apply cleanly on all active branches and be safe to use with all configurations.

[shr-project]

@agherzan @zach-welch-aquabyte was this intentionally only for kirkstone? I guess it should be cherry-picked to langdale and master as well as description mentions it does apply cleanly.

[zach-welch-aquabyte]

Yes, I intended for this to be cherry-picked to newer branches. AFAIK, the GitHub interface does not allow for a multi-headed PRs, and it did not seem appropriate to file one PR per branch.

[kraj]

Yes, I intended for this to be cherry-picked to newer branches. AFAIK, the GitHub interface does not allow for a multi-headed PRs, and it did not seem appropriate to file one PR per branch.

a hint in subject could be good for maintainers. I think ideally fix should go into master first ( if needed there ) then backported to branches, not vise-versa

[zach-welch-aquabyte]

@agherzan Today I noticed that this patch still has been merged only to kirkstone and not the newer active branches (as intended). Can I do anything to help get it merged on those newer branches?

[agherzan]

@zach-welch-aquabyte, it would be helpful if you could raise an MR against master with this cherry-picked.

[shr-project]

I've added PR for mickledore now as well as mickledore is still supported and received some updates recently as well.