Second attempt to enable custom tree for 'zinit ls' command

[nbrown]

Enable users to specify the zinit ls command via ZINIT[LIST_COMMAND] and
provide default commands for exa and tree. Prefer exa if it is installed.

Addresses issue #170

I cherry-picked the original changes to a clean branch and then updated the docs to attempt to address the documentation action issues.

[vladdoster]

oof, this zshdocs bug is a doozy

[alichtman]

Whatever. Let the test fail.

[alichtman]

Haven't tested it, but looks good to me

[alichtman]

Haven't tested it, but looks good to me

[nbrown]

Well, I managed to make it work by using a docker container used in https://github.com/nektos/act

My conclusion is that there is something different between my macOS en_US.UTF-8 and the one used in the github actions containers.

[alichtman]

This PR shouldn't contain a massive docs update. That should already be on master. I don't have time to jump into debugging it but I won't merge it with an ~800 line addition diff

[vladdoster]

@nbrown great job on getting this to pass

[vladdoster]

@nbrown Since you mentioned this, I did some digging and think I found the root issue.

I went to our Dockerfile and saw we don't explicitly set the locale, which seemed odd. A few GitHub issues later and I found this issue:

dotnet/runtime#962

Your hypothesis seems to be spot on and created an issue to track it.

[nbrown]

This PR shouldn't contain a massive docs update. That should already be on master. I don't have time to jump into debugging it but I won't merge it with an ~800 line addition diff

Sorry for the large doc change, I don't know how you want to handle it but the two choices as I see it are minimal doc changes that don't pass the documentation check or a large change that passes the check. I'll gladly revert it to not pass that check if that's what is preferred.

[alichtman]

Yeah, let's keep the PR tight and drop the docs change. We'll get the docs test sorted out soon (maybe, idk -- no promises) but I'm fine with it failing here.

[nbrown]

Yeah, let's keep the PR tight and drop the docs change. We'll get the docs test sorted out soon (maybe, idk -- no promises) but I'm fine with it failing here.

Done. I reverted the doc fix so it's now a minimal change.

On the docs test, I'll leave some info on that bug. I think it primarily comes down to a sorting (collation) inconsistency between Linux and macOS/BSD based on my reading and testing.

[vladdoster]

This check evaluates to true even though I have exa, but no tree.

[vladdoster]

Closing in favor of #221 due to lack of update on the PR.