Recent changes have added special cases for line breaks within URLs
and email addresses. Such breaks are undesirable when they can be
avoided, but at other times are needed to avoid huge gaps, or indeed
to make the line fit at all.

This patch assigns a penalty for such breaks, equal to the hyphenation
penalty. The mechanism is currently very simple, but would be easy to
fine-tune based on more detailed information about break quality.

Bug: 20126487
Bug: 20566159
Change-Id: I0d3323897737a2850f1e734fa17b96b065eabd9c

This change adds accceptable line breaks according to sections 7.42
(Dividing URLs and e-mail addresses) and 14.12 (URLs or DOIs and line
breaks) of the Chicago Manual of Style (16th ed.). In general, these
place breaks before punctuation symbols, and suppresses them after
hyphens.

Bug: 20126487
Bug: 20566159
Change-Id: I2d07d516b920a506a2f718c38fb435c5eb1ee1f8

Detect URLs and email addresses, and suppress both line breaking and
hyphenation within them.

Bug: 20126487
Bug: 20566159

Change-Id: I43629347a063dcf579e355e5b678d7195f453ad9

Implement a WordBreaker that defines our concept of valid word
boundaries, customizing the ICU behavior. Currently, we suppress line
breaks at soft hyphens (these are handled specially). Also, the
new WordBreaker class has methods that determine the start and end
of the word (punctuation stripped) for the purpose of hyphenation.

This patch, in its current form, doesn't handle email addresses and
URLs specially, but the WordBreaker class is the correct place to do
so. Also, special case handling of hyphens and dashes is still done
in LineBreaker, but all of that should be moved to WordBreaker.

Bug: 20126487
Bug: 20566159
Change-Id: I492cbad963f9b74a2915f010dad46bb91f97b2fe

Removing variables in main function of sample/example.cpp as
they are not used.

Bug: 26936282
Change-Id: I64ae0a455b413df333ddd4810a9e090d52322041

Bug: 24505153
Change-Id: If61c063c175086dec88cda187eafd9ce923e4cb1

It turned out that hb_font_t creation is not a lightweight operation.
Especially, Paint.hasGlyph creates hb_font_t for all existing fonts
every time. To improve the performance, cache hb_font_t instead
of hb_face_t.

Note that to calculate horizontal advance, MinikinPaint needs to be
associated with hb_font_t by calling hb_font_set_funcs. With this patch,
hb_font_set_funcs may be called multiple times for the same hb_font_t
object. However this is not an issue since MinikinPaint is unique
during layout.

Bug: 26784699

Change-Id: I516498ae9f0127d700fc9829327e9789845a1416

am: 3d15532a

* commit '3d15532a':
  Disable unsigned integer overflow sanitization until libc++ is fixed.

am: 4553fa41

* commit '4553fa41':
  Disable unsigned integer overflow sanitization until libc++ is fixed.

Bug: http://b/26781196
Bug: http://b/25884483
Bug: http://b/26432628

Although this issue was first only manifesting on Fugu, it now affects
N9 and N6p as well. This change disables unsigned overflow sanitization
on all platforms. The real fix for libc++ (r257368) can't be committed
until we have updated Clang at least one more time.

Change-Id: I71e9c50d25ae4566d4c06f348183c4b22a4bb60a

The intruduced method measureText can be used instead of
doLayout for text measurement purpose.

Bug: 24505153
Change-Id: Ic29bbb347daf18d1f6c13f86970dcdd11dd6a2bd

Reject fonts with invalid ranges in cmap am: ca8ac8ac am: 3034e8b4 am: a3df6da3 am: e82552ae am: 545b36d6 am: 98dd15e2 am: 97d72658  -s ours
am: f5dc6044  -s ours

* commit 'f5dc6044':
  Reject fonts with invalid ranges in cmap

Reject fonts with invalid ranges in cmap am: ca8ac8ac am: 3034e8b4 am: a3df6da3 am: e82552ae am: 545b36d6 am: 98dd15e2
am: 97d72658  -s ours

* commit '97d72658':
  Reject fonts with invalid ranges in cmap

Reject fonts with invalid ranges in cmap am: ca8ac8ac am: 3034e8b4 am: a3df6da3 am: e82552ae am: 545b36d6
am: 98dd15e2

* commit '98dd15e2':
  Reject fonts with invalid ranges in cmap

Reject fonts with invalid ranges in cmap am: ca8ac8ac am: 3034e8b4 am: a3df6da3 am: e82552ae
am: 545b36d6

* commit '545b36d6':
  Reject fonts with invalid ranges in cmap

am: e82552ae

* commit 'e82552ae':
  Reject fonts with invalid ranges in cmap

am: a3df6da3

* commit 'a3df6da3':
  Reject fonts with invalid ranges in cmap

am: 3034e8b4

* commit '3034e8b4':
  Reject fonts with invalid ranges in cmap

am: ca8ac8ac

* commit 'ca8ac8ac':
  Reject fonts with invalid ranges in cmap

The motivation of this CL is enhance the font fallback score design
to support multiple language font fallback.

This CL contains following changes:
- Break language based font score into two: script-based score and
  primary-language-based score.
- The primary-language-based score is 0 if the script-based score is 0.
  If the script-based score is not 0 and the primary language is the
  as same as the requested one, the font gets an extra score of 1.
- The language score gets a higher multiplier for languages higher in
  the locale list.

Bug: 25122318
Bug: 26168983
Change-Id: Ib999997a88e6977e341f4c325e2a1b41a59db2d5

am: 850fcc0b  -s ours

* commit '850fcc0b':
  Tailor grapheme boundaries so sequence emoji are one grapheme - DO NOT MERGE

am: 2e98eb6b

* commit '2e98eb6b':
  Tailor grapheme boundaries so sequence emoji are one grapheme - DO NOT MERGE

Make it so it's not possible to position the cursor inside an emoji
formed by a sequence including zero-width joiners.

Bug: 25368653
Change-Id: I67ec0874cd1505f3c82ab91492ffc3d39a52fae6

Make it so it's not possible to position the cursor inside an emoji
formed by a sequence including zero-width joiners.

Bug: 25368653
Change-Id: I67ec0874cd1505f3c82ab91492ffc3d39a52fae6

A corrupt or malicious font may have a negative size in its cmap
range, which in turn could lead to memory corruption. This patch
detects the case and rejects the font, and also includes an assertion
in the sparse bit set implementation if we missed any such case.

External issue:
https://code.google.com/p/android/issues/detail?id=192618

Bug: 26413177
Change-Id: Icc0c80e4ef389abba0964495b89aa0fae3e9f4b2

A corrupt or malicious font may have a negative size in its cmap
range, which in turn could lead to memory corruption. This patch
detects the case and rejects the font, and also includes an assertion
in the sparse bit set implementation if we missed any such case.

External issue:
https://code.google.com/p/android/issues/detail?id=192618

Bug: 26413177
Change-Id: Icc0c80e4ef389abba0964495b89aa0fae3e9f4b2

am: 0497c013

* commit '0497c013':
  Minikin: Disable sanitizer on x86

am: 72c02682

* commit '72c02682':
  Minikin: Disable sanitizer on x86

Disable unsigned-integer-overflow sanitizer on x86, as it crashes.

Bug: 25884483
Bug: 26432628
Change-Id: Ia658ed56a6c81660a36edf71f7116118056aa917

am: cf6ebc1e

* commit 'cf6ebc1e':
  Enable integer sanitization in libminikin

am: 7087da25

* commit '7087da25':
  Enable integer sanitization in libminikin

Enable signed and unsigned integer sanitization in libminikin.

Bug: 25884483
Change-Id: I98905827174d16138d20bb443fe2e1d7228ea1a3

am: 5255614d

* commit '5255614d':
  Revert "Enable integer sanitization in libminikin."

am: 342cd45a

* commit '342cd45a':
  Revert "Enable integer sanitization in libminikin."

This reverts commit 9c13a3dc.

Change-Id: Iecd1f61a5fdf5955d871a920cb243553857d46ff

am: ebe37167

* commit 'ebe37167':
  Enable integer sanitization in libminikin.

am: 9c13a3dc

* commit '9c13a3dc':
  Enable integer sanitization in libminikin.

The assertion for the lock state has now activated by
I9c4b1e1f09c6793e387fbdb8bb654cc0a13c65d5.
This CL fixes the assertion failure in the unit tests by acquiring
lock before calling the functions.

Change-Id: I6a6afefb4de01e8610c2abfe6c779afa9442cc67