Added print accessibility tree command. Find accessibility element command reworked

[gkassabli]

Added print accessibility tree command that is based on Apple internal methods. It could print whole tree or subtree of specified root element.
Find accessibility element command was reworked to use internal API and make use of only relevant and visible elements

[KingOfBrian]

This is fantastic. Great finds on the AX stuff!

[KingOfBrian]

It would be helpful to add the ability to print the accessibility value and traits if possible. If you had the values for them, adding an optional second argument to pa11y would be pretty easy.

[gkassabli]

Hey, @KingOfBrian.
That's a great idea. I haven't looked into that, but I will try to add this in next iteration of command

[kastiglione]

@KingOfBrian How does this pull request compare / contrast to #55?

[KingOfBrian]

Much better approach, with fewer round trips by a large margin. The different types of AX values can be added pretty easily. forceStartAccessibilityServer is awesome, I'm gonna add that to FRY to enable AX.

I've never determined traversal choices via a nil accessiblityLabel, and I think it may cut the tree off a bit early. A UITableView accessibilityLabel for instance is used to store the empty table view AX value, and without actually testing, I'm not sure how this would work. There are other instances, like a navigation bar item where the accessibilityLabel is duplicated in an internal view. It's not important to pick up that distinction, but for a debug tool, I think it'd be better to be more liberal. And, some magic inside of _accessibilityElementsInContainer may behave differently than I think.

Anyway, this approach is definitely the correct one. I don't have time to test it, as my AX work has passed for now.

[kastiglione]

It seems this will always be true, since foundLocally will always be false in this elif branch.

[gkassabli]

You're right. Did it correct way initially, but refactored wrong way

[kastiglione]

Were you intending to rectify this in 55cca66?

[gkassabli]

Yes, this should work properly now

[kastiglione]

👍

[kastiglione]

Thanks @a2, and apologies @gkassabli for the dropped ball. I had asked a question and was waiting to hear back, but perhaps it wasn't seen.

[gkassabli]

@kastiglione Yeah, I fixed it there. Apologies for not responding in thread