GitLab

1. Previously, our implementation of `ExpandToEnclosingUnit` didn't take
   into account that, since we changed the behavior of text positions to
   be able to reach the maximum text offset of its anchor, there are
   multiple text positions that are equal when compared, but are located
   in different anchors.

   For example, the end position of an anchor and the start position of
   the next text anchor. This could get even trickier with empty text
   anchors between characters in the text representation.

   Such scenarios are relevant to some assistive technologies since the
   anchors that the enclosing range spans affect `GetEnclosingElement`,
   ultimately breaking functions of these ATs.

   Implementing two new methods `AsPositionBeforeCharacter` and
   `AsPositionAfterCharacter` in `AXPosition`, which we will use to
   normalize positions between characters of the text representation.

2. Adding unit tests for both new methods in `AXPositionTest`.

3. Improving `TestITextRangeProviderExpandToEnclosingCharacter` coverage
   in `AXPlatformNodeTextRangeProviderTest` to check that the enclosing
   element resolved from the expanded range matches AT expectations.

4. Addressing a bug in `ExpandToEnclosingUnit` where the position at the
   end of the last anchor of the tree does not expand to the last
   character of its text representation.

Bug: 961893
Change-Id: Ib168d680578a3b460b57b7ca9808987568743a95
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1614940Reviewed-by: Nektarios Paisios <nektar@chromium.org>
Reviewed-by: Kurt Catti-Schmidt <kschmi@microsoft.com>
Commit-Queue: Ethan Jimenez <ethavar@microsoft.com>
Cr-Commit-Position: refs/heads/master@{#660881}