When rendering clusters that have colored and non-colored
glyphs, we were falling back to using show_text_glyphs,
rendering all glyphs in the cluster without color.
Instead, make composite_one_color_glyph smart enough
to handle non-color glyphs, and only skip clusters
that are entirely non-color.
Testcase: hb-view AmiriQuranColored.otf -u 627,64e
When the run contains a mix of color and non-color
glyphs, composite_color_glyphs returns the non-color
glyphs unhandled in the same array. In the rtl case,
the glyphs are processed from the end and the unhandled
glyphs are accumulated at the end as well, and we
need to move them to the beginning of the array before
we return. Add the missing memmove call to do that.
Fixes: #533
Symbols from cairo-tag-stack.c are used by cairo-tag-attributes.c, so
this should be in the same list of sources as that. Fixes this build
error:
```
/usr/bin/ld: src/libcairo.so.2.11705.0.p/cairo-tag-attributes.c.o: in function `parse_array':
/path/to/cairo/_build/../src/cairo-tag-attributes.c:347: undefined reference to `_cairo_tag_error'
/usr/bin/ld: src/libcairo.so.2.11705.0.p/cairo-tag-attributes.c.o: in function `parse_name':
/path/to/cairo/cairo/_build/../src/cairo-tag-attributes.c:359: undefined reference to `_cairo_tag_error'
/usr/bin/ld: src/libcairo.so.2.11705.0.p/cairo-tag-attributes.c.o: in function `parse_attributes':
/path/to/cairo/cairo/_build/../src/cairo-tag-attributes.c:410: undefined reference to `_cairo_tag_error'
/usr/bin/ld: /path/to/cairo/cairo/_build/../src/cairo-tag-attributes.c:431: undefined reference to `_cairo_tag_error'
/usr/bin/ld: /path/to/cairo/cairo/_build/../src/cairo-tag-attributes.c:441: undefined reference to `_cairo_tag_error'
/usr/bin/ld: src/libcairo.so.2.11705.0.p/cairo-tag-attributes.c.o:/path/to/cairo/cairo/_build/../src/cairo-tag-attributes.c:455: more undefined references to `_cairo_tag_error' follow
/usr/bin/ld: src/libcairo.so.2.11705.0: hidden symbol `_cairo_tag_error' isn't defined
```
This commit fixes the following warning from our check-def.sh:
PASS: check-def.sh
Checking documentation for incorrect syntax
./cairoint.h (523): ERROR: Get invalid doc id (should be 'cairo_...:')
./cairoint.h (534): ERROR: Get bad line: ' */'
./cairoint.h (534): ERROR: Get documentation comment not closed with **/
Signed-off-by: Uli Schlachter <psychon@znc.in>
The expression "1 << whatever" has type int, no matter what the type of
"whatever" is. Thus, "1 << 31" ends up overflowing an "int" and
undefined behaviour occurs.
The above happened in cairo-mempool.c. I saw the following line:
pool->free_bytes += 1 << (bits + pool->min_bits);
being executed with bits=15 and pool->min_bits=16, i.e. we had 1 << 31.
This ended up being INT_MIN due to the overflow. This was then promoted
to size_t and we ended up with a *huge* value being added to free_bytes.
This is obviously not the intended behaviour. Thus, this commit replaces
the "1" in all left shifts in cairo-mempool.c with "((size_t) 1)".
This fix avoids the integer overflow, but it does not fix issue #510,
because some allocation keeps the memory pool alive.
Related-to: https://gitlab.freedesktop.org/cairo/cairo/-/issues/510
Signed-off-by: Uli Schlachter <psychon@znc.in>
The FT change is because my MinGW build is using a more recent version
of FT.
Remove the disabled _cairo_win32_scaled_font_text_to_glyphs() code to
fix the defined but not used warning.
_cairo_win32_scaled_font_text_to_glyphs() was diabled in d9408041aa with
the comment:
"Currently disable the win32-font text_to_glyphs(), until that one
is updated. Or better yet, remove it and implement
ucs4_to_index(). It's the toy font API afterall."
_cairo_win32_scaled_font_ucs4_to_index() was added in d1c619bc7d.
In practice, the A and B images may be any mixture of RGB24 and
ARGB32 formats, so this change accepts all combinations of these
types, and converts the pixel values to a common (ARGB32) format
as needed.
Some of the newly added test failures are cases where the image
output is RGB24, but the matching reference image is ARGB32 with
noticeable transparency. Some of the newly passing tests are cases
where the unused 'alpha' channel of an RGB24 image was not equal
to 0xff, and the previous code had incorrectly used this channel
in max_diff calculations.
To avoid reading a potentially garbage alpha channel when users
of pdiff_compare pass in RGB24 images, if the format is RGB24,
force the alpha channel to be 0xff.
This commit also updates CI to adjust for the new tests that have
started/stopped failing. New failures often are cases where
the reference image has alpha transparency, but the test output
does not; new passing tests may indicate that the unused alpha
channel of an RGB24 image was garbage, but now is ignored.
The proper pthread check activated some tests that weren't active
before, which resulted in a test failure apparently caused by
inexact test rendering.
Update the quartz reference image accordingly.
COLR fonts can have a layer with the same color as the current text
color. This change passes the current color (if solid) through to
the font backend where it can be used to render color fonts.
scaled_glyph_lookup checks if the foreground color has changed (for
glyph that require it) and requests a new color surface if required.
This also fixes a bug where scaled_glyph_lookup would always request a
color surface for glyphs for glyphs in color fonts that do not have
color.
We are trying to get an approximate bounding box for
the extents of a glyph string and use it for bailing
early if there's nothing to do, but the code computing
that approximation makes invalid assumptions: It assumes
that a glyph never extends further than max-advance-width
before or after its position. That is not true in general,
as can be seen in https://gitlab.gnome.org/GNOME/gtk/-/issues/4132
In the example there, we have a monospace fonts with ligatures
that have a negative lsb of two or three times the glyph width.
The optimization here could theoretically be fixed by iterating
over the font once to determine suitable values for how far
glyphs can extend before or after their position, but this
patch just removes it.
