Fixes #416.

Fixes #471.

...as present in MS TrueType files. Since this table is optional,
the new stbtt_GetFontVMetricsOS2 has a return value and can fail.
This is a replacement for pull request #463.

Fixes #463.

This incorporates #462, but also factors everything into one
function that is shared between 8-bit integer, 16-bit integer, and
float pixels (vertical flip operates on rows of bytes and doesn't
really care), and finally always uses a 2k on-stack buffer without
dynamic memory allocation, doing multiple memcpys per row if
necessary. Not only does this remove an out-of-memory failure mode,
it is also preferable for large images, since it's more
L1-cache-firendly this way.

Fixes #462.

1. const char* for __FILE__ (string literals are const)
2. Use %zd to print size_t where available; the only real problem
   here is Visual C++. Use long long on the VC++ vers that support
   64-bit targets but not %zd, int on the even older 32-bit-only
   VC++ vers that don't support "long long" either.

Fixes #459. I think. (It's hard to be sure since the issue doesn't
state the exact warning message.)

This is the same method as in pull request #455, but using integer
arithmetic instead of converting to float.

Fixes #455.

Merge branch 'fix-vs2015-warnings' of https://github.com/Reedbeta/stb into Reedbeta-fix-vs2015-warnings

Merge branch 'load_16_variants' of https://github.com/poppolopoppo/stb into poppolopoppo-load_16_variants

Fixes issue #461.

req_comp is now desired_channels and *comp is *channels_in_file.

Fixes issue #466.

We used to require exact match between img_len and raw_len for
non-interlaced PNGs, but the PNG in issue #276 has extra bytes
(all zeros) at the end of the compressed DEFLATE stream.

The PNG spec doesn't have anything to say about it (that I
can tell), and if libpng accepts this, who are we to judge.

Fixes issue #276.