Refactor graphite2 to make ovleraps into negative width spaces

Fixes https://github.com/behdad/harfbuzz/issues/212

Very rudimentary right now, but will get kcc's bot going.

From
https://github.com/behdad/harfbuzz/issues/139#issuecomment-163580783

This speeds up shaping the Amiri font by over 15%.

This was primarily needed for my work on OpenType GX, since
we will be collecting only sublookups that are "active" for
current font instance; but it's a nice boost in general as
well.

We might, in the future, collect subtables in the lookup_accel.
That would also allow us to do a per-subtbale set-digest, which
should speed things up some more, specially for ContextChainFormat3
lookups...  Amiri, for example, contains one lookup with 53
subtables!

Fixes https://code.google.com/p/chromium/issues/detail?id=547912

Might add italic-angle, underline/strikethrough-position/thickness
in the future.  Do this before new struct goes into a release.

This happens with at least one test font I have.

See previous commit.

The font Garamond Premier Pro Caption (and possibly many other
Adobe fonts), have many FeatureParamsSize tables with the old
wrong offset.  We handle fixing those up, but they were still
contributing to edit_count, and when I reduced HB_SANITIZE_MAX_EDIT
from 100 to 8 in 14c2de32, these
fonts were now getting GPOS dropped and hence kerning disabled.

Fix, by not counting edits made towareds offset fix-up.  I'll
also increase edit count again, in the next commit.

[travis] Try building on Trusty

Trusty has gtk-doc 1.20 which has better markdown support which should
fix some of the brokenness in the newly added documentation. The build
seems to be as fast as the current container-based build on Precise.

Fixes failures in others' forks, like:
https://travis-ci.org/khaledhosny/harfbuzz/jobs/99719710#L3912

[docs] More buffers documentation

More docs

Remove unused HB_SHAPER_DATA_ENSURE_DECLARE lines.

The coretext_aat shaper delegates to the regular coretext_..._ensure() functions, so coretext_aat_..._ensure() functions defined by these macros are unused. The compiler warns about them, which in turn can confuse people to think that the coretext_aat_..._ensure() functions weren't called by accident.

Now it is just reporting one error about “503 undocumented or incomplete
symbols”, that will be the easy part I guess.

Remove unused `retry:` label.

Temporarily makes it rebuild docs on every change again.

Doc fixes

Some of it (create, reference, destroy) are adapted from Cairo docs.

To lower the number of gtk-doc warnings.

Fixes a -Wunused-label warning when building harfbuzz with clang -Wall.