perf(QTree): rewriting the inner strategies of QTree to use hashes instead of greedy iterations (fix #17941)

[gilmar-sales]

What kind of change does this PR introduce?

• [ ] Bugfix
• [ ] Feature
• [ ] Documentation
• [ ] Code style update
• [x] Refactor
• [ ] Build-related changes
• [ ] Other, please describe:

Does this PR introduce a breaking change?

• [ ] Yes
• [x] No

The PR fulfills these requirements:

• [x] It's submitted to the dev branch (or v[X] branch)
• [x] When resolving a specific issue, it's referenced in the PR's title (e.g. fix: #xxx[,#xxx], where "xxx" is the issue number)
• [ ] It's been tested on a Cordova (iOS, Android) app
• [ ] It's been tested on an Electron app
• [ ] Any necessary documentation has been added or updated in the docs or explained in the PR's description.

Other information:

• This PR will solve the bad performance to find the expanded/ticked node when the big tree is loaded, but not solves the rendering overhead when full expanded.

[gilmar-sales]

I'll finish this PR by refactoring the meta computed that is spending so much time with large data sets, I'm thinking in remove the computed up-bottom strategy and do bottom-up travel from the edited node to recompute only the necessary metas, instead of recomputing the entire meta every time a node gets expanded/ticked.