[9.x] Remove reduceWithKeys and merge the functionality with reduce

[mokhosh]

This moves reduce out of Collection and LazyCollection and puts it into EnumeratesValues as suggested by @JosephSilber.

The existing test for reduce has be adapted to contain asserts for both situations, with and without $key.

I've also moved tap one method down to be alphabetically correct.

@JosephSilber while I was checking the code in EnumeratesValues I got an idea. Can we make reduce higher proxy-able as well? I don't know how to do that now, but I'm willing to do it if you give me a little guidance.

[driesvints]

@JosephSilber you didn't get pinged by GitHub yet because we haven't merged the CODEOWNER changes yet into master so pinging you manually here.

[driesvints]

@mokhosh there's a merge commit in this commit history. Can you rewrite the history so it's clean?

[mokhosh]

Sure

[JosephSilber]

This moves reduce out of Collection and LazyCollection and puts it into EnumeratesValues

This was just done for 8.x, which is periodically merged with master. No need to do it here separately.

I've also moved tap one method down to be alphabetically correct.

Should probably go in a separate PR.

Can we make reduce higher proxy-able as well?

We can't. Higher order proxies work by calling a method on each item, and returning its values. Since reduce needs an additional parameter (the $carry), I believe it can't ever work with higher order proxies.

[mokhosh]

This was just done for 8.x, which is periodically merged with master. No need to do it here separately.

There's no diff there actually with the 8.x PR.
I should just say "this removes reduceWithKeys", I thought people who see this now might be in a 8.x mindset so I explained a bit more.

Should probably go in a separate PR.

I think it's already merged in the 8.x PR, so should we just leave it? I'll keep this in mind for the future contributions.

We can't. Higher order proxies work by calling a method on each item, and returning its values. Since reduce needs an additional parameter (the $carry), I believe it can't ever work with higher order proxies.

You're right. Thanks.

[mokhosh]

@driesvints I'm not sure why git rebase isn't showing me the merge commit so I can squash it. I red and watched a dozen tutorials on rebase, but they all worked on simple examples with one branch and no merge commits.

Can you please guide me through this?

[gocanto]

@driesvints I'm not sure why git rebase isn't showing me the merge commit so I can squash it. I red and watched a dozen tutorials on rebase, but they all worked on simple examples with one branch and no merge commits.

Can you please guide me through this?

@mokhosh
You could do this from your CLI

git checkout master
git pull
git checkout mokhosh:master
git pull
git rebase master -i

#edit vim CL view
pick `add foo`
s `merge commit to be removed`
pick testReduce`

then, you can choose the commit message you fancy better for those changes. 👍🏻

[mokhosh]

@gocanto Thanks for your reply. The third command gave me this error:

error: pathspec 'mokhosh:master' did not match any file(s) known to git

[gocanto]

@gocanto Thanks for your reply. The third command gave me this error:

error: pathspec 'mokhosh:master' did not match any file(s) known to git

mokhosh:master should be your master branch reference. Whatever the name or source that might be! :)

[driesvints]

@mokhosh I'd just check out a new branch and do the code changes anew as the current timeline is a bit messed up now. Then force push to your branch here.

[mokhosh]

@driesvints Thank you, and sorry for the trouble. I'll be cleaner next time.