GitLab

This CL is a refactoring on the way RenderText do
ItemizeTextToRuns for the corner cases where a run
should be split in grapheme clusters.

The previous code was naming these cases:
  * Unusual characters
  * Special characters
  * Non-regular characters
But they should be more specific on the purpose of splitting the runs.

Also, the algorithm used for splitting a sequence of codepoint
was based on the code block of the corresponding codepoints
and few cases where trying to merge adjacent codepoint in
a grapheme. The algorithm was incorrect multiple cases (e.g. emoji).

Unicode provide a way to split a sequence of codepoints into
grapheme and grapheme clusters.
They provide a state machine which is using the codepoint proporties
to decide if the current location is a grapheme boundaries.

The ICU library is providing an API over the character properties
to help iterating over graphemes.
  * ubrk_open
  * ubrk_first
  * ubrk_next
  * ubrk_close

The class base::i18n::BreakIterator(..., BREAK_CHARACTER) is
providing an easy to use wrapper over that API.


The current CL is replacing the previous characters based splitting
algorithm by the graphemes based version.



See emoji sequence:
  http://www.unicode.org/reports/tr51/
  1.4.5 Emoji Sequences

The full emoji list:
  * http://unicode.org/emoji/charts/full-emoji-list.html
Emoji data, used to make our unittests:
  * http://www.unicode.org/Public/emoji/12.0/emoji-data.txt


see: UNICODE TEXT SEGMENTATION (http://unicode.org/reports/tr29/)
see: https://cs.chromium.org/chromium/src/third_party/icu/source/common/unicode/ubrk.h
see: https://cs.chromium.org/chromium/src/base/i18n/break_iterator.h
Change-Id: I6b9a9c79021f2ce0e2db7cdefdd0838b5911f445
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1788804
Commit-Queue: Etienne Bergeron <etienneb@chromium.org>
Reviewed-by: Alexei Svitkine <asvitkine@chromium.org>
Reviewed-by: Robert Liao <robliao@chromium.org>
Cr-Commit-Position: refs/heads/master@{#712704}