It's considerably faster than the fallback implementation we had
previously!

The merger of normalizer and glyph-mapping broke shapers that
modified text stream.  Unbreak them by adding a new preprocess_text
shaping stage that happens before normalizing/cmap and disallow
setup_mask modification of actual text.

Ouch!

A *real* release this time, with NEWS, ChangeLog, etc.

The change is very subtle.  If we have a single-char cluster that
decomposes to three or more characters, then try recomposition, in
case the farther mark may compose with the base.

Essentially move the glyph mapping to normalization process.
The effect on Devanagari is small (but observable).  Should be more
observable in simple text, like ASCII.

Preparing for merging map_glyphs() and normalize().

To room for more allocations, coming.

This reverts commit 30ec9002.

See previous commit.

This reverts commit 0981068b.

I was confused.  Even if we access coverage[0] unconditionally, we don't
need bound checks since the array machinary already handles that.

And misc linking fixes.

Apparently even that doesn't make check-internal-symbols.sh happy with
mingw32.  Going to disable that for DLLs again, but hopefully the
export-file is doing *something*.

https://bugs.freedesktop.org/show_bug.cgi?id=53233

Implemented for Arabic, Hebrew, and generic marks.
Activated if no GPOS table present.

Used when there is no GPOS.

Fixes logic also, where before we were always matching if glyphs_len==1
and a ligature started with the glyph.