Required some tinkering, hope I didn't mess the logic up.
This now requires the config section to be set for the header
file, and different sites that include it should agree on what
the values are. This is kind of iffy but hard to avoid.
Fixes issue #647.
To pacify GCC warnings at -Wimplicit-fallthrough=4. Why the
all-caps version works and the others don't, I'm not sure; the
GCC manual page lists regular expressions that GCC is checking for
but does not say which regular expression syntax variant it's using,
whether it's looking for a substring match or a full match for
the comment, and what exactly the text being matched against is
for a single-line comment. Sigh.
Fixes issue #507.
Was using 4-character spaces and otherwise formatted unlike the
rest of the code, fix this. Also get rid of the outer switch in
GetGlyphGPOSInfoAdvance with just one case; just use an if.
No behavioral changes.
Glyphs not assigned a glyph class should default to class 0 as
per the OpenType spec. Also, change the code to treat malformed
data as an error to be handled, not an assert, and change the
way the version checking works.
Fixes the issue mentioned in PR #1035. Also, this part of the
code is indented incorrectly; will fix that in a subsequent
commit.
If the data was uncompressible and this deflate implementation expanded
by more than the overhead of simply storing it uncompressed, fall back
to deflate's uncompressed storage mode. This bounds the maximum deflated
size at the original size plus an overhead of 6 fixed bytes with another
5 bytes per 32767 byte block.
Fixes issue #948.
These mostly add very little and have caused problems for people,
nor does it make sense to require this when the underlying
computations are performed in floating-point arithmetic depending
on ratios of user-passed in image dimensions.
Arbitrary absolute epsilons here would just be garbage; we could
try and compute desired relative error bounds based on the
determined scale values, but this still leaves the questions of
what purpose this would even serve, which is unclear.
Leave the filter kernel asserts as comment for documentation
of what the behavior would be with exact math, but don't actually
bother asserting here.
Fixes issue #736.
We've established the signs of values before so we can carefully
jiggle the expressions to be guaranteed overflow-free; the tests
for <0 here were meant to check if the value was "still negative",
i.e. if the sum did not underflow below INT_MIN, and we can
rewrite that using algebra to be overflow-free.
We need an extra case in the Euclidean dvision for
INT_MIN / INT_MIN which is a bit annoying but trivial; finally,
for two's complement platforms, note that abs(x) = (x > 0) ? x : -x
does not work (since for x=INT_MIN, -x still gives INT_MIN),
but the equivalent formulation for _negative_ absolute value
(x < 0) ? x : -x is always in range and overflow-free, so
rewrite the relevant expressions using that negative absolute
value instead.
Fixes issue #741.
Re-added calls to `stbte__hittest()`, fixed some compiler errors.
Also fixed some GCC warnings about unused variables when
STBTE__COLORPICKER and STBTE_ALLOW_LINK is not defined.
Confirmed from the OpenType spec that there's nothing missing
here. (These were just annotations listing the total sizes of
the tables, but this is not used for anything here.)
Fixes issue #704.
As per MS's own docs, should ignore the r/g/b bitmasks in the
header unless BI_BITFIELDS compression is selected. Factor out
setting the default masks since that now exists in two branches.
Add some more checking for unsupported compression formats and
illegal bpp/compression combinations while I'm at it.
Fixes issue #783.
As per issue 634, Stua "will be deleted from stb.h" soonish. That
was 3 years ago, which should be plenty of warning, and the
language has been de-facto orphaned and undocumented for a
good while longer than that.
Fixes issue #634.
Released Clang 12 generates bad code for the original loop in here.
While this is a compiler bug plain and simple, we still have to deal
with it.
This is related to the SLP vectorizer, and in particular the two
reverse subtracts in the butterflies for the second half to avoid
unary negates.
Use the more regular dataflow that has the unary negates in it
(we can at least fold one of them into a constant, namely for A2)
and introduce a few temporaries that also make alias analysis (and
possible block-level vectorization) a whole let easier while I'm at
it.
This fixes the codegen issues on Clang 12, which now produces a
working decoder, and I expect the single unary negate that we
actually gain per iteration of this loop is not a significant
perf concern. (There are bigger fish to fry here regardless.)
Fixes issue #1152.
Some parameters do not get used, or only when certain config
defines are set. Explicitly mark them as unused to make compilers
happy.
Fixes issue #396.
This is definitely unnecessary, or at least I can't find anything
in the Vorbis spec that would indicate anything special happening
here.
Fixes issue #816.
Fabian Giesen