This broke with the clone_similar optimization in
8d7a02ed58
The optimization added an interest rectangle to clone_similar,
but the acquire_surface path was neglecting to transform its
rectangle by the pattern matrix.
The test suite did catch this, but apparently we were too
distracted by the performance improvements to notice. Only
backends other than image that implemented clone_similar
would be affected by the bug, (which meant I only saw xlib
failures in my testing).
This fixes bug #8711
This fixes a huge performance bug (entire image was being pushed to X
server in order to copy a tiny piece of it). I see up to 50x improvement
from subimage_copy (which was designed to expose this problem) but also
a 5x improvement in some text performance cases.
xlib-rgba subimage_copy-512 3.93 2.46% -> 0.07 2.71%: 52.91x faster
███████████████████████████████████████████████████▉
xlib-rgb subimage_copy-512 4.03 1.97% -> 0.09 2.61%: 44.74x faster
███████████████████████████████████████████▊
xlib-rgba subimage_copy-256 1.02 2.25% -> 0.07 0.56%: 14.42x faster
█████████████▍
xlib-rgba text_image_rgb_over-256 63.21 1.53% -> 11.87 2.17%: 5.33x faster
████▍
xlib-rgba text_image_rgba_over-256 62.31 0.72% -> 11.87 2.82%: 5.25x faster
████▎
xlib-rgba text_image_rgba_source-256 67.97 0.85% -> 16.48 2.23%: 4.13x faster
███▏
xlib-rgba text_image_rgb_source-256 68.82 0.55% -> 16.93 2.10%: 4.07x faster
███▏
xlib-rgba subimage_copy-128 0.19 1.72% -> 0.06 0.85%: 3.10x faster
██▏
In particular, many possible error values on a surface provided
to cairo_pattern_create_for_surface were previously being swallowed
and a nil pattern was returned that erroneously reported
CAIRO_STATUS_NO_MEMORY.
Adds API functions for inspecting the current dash state, as well as
the contents of pattern objects:
cairo_get_dash
cairo_get_dash_count
cairo_pattern_get_rgba
cairo_pattern_get_surface
cairo_pattern_get_color_stop_rgba
cairo_pattern_get_color_stop_count
cairo_pattern_get_linear_points
cairo_pattern_get_radial_circles
A recent renaming from CAIRO_MAXSHORT also changed the range of some
values from INT16_MAX to UINT16_MAX. The test suite shows one case
where this introduces a bug, (push/pop_group with SVG and RGB24
surfaces).
I haven't tracked the bug down, (presumably there's overflow of the
unsigned value at some point), but I'm now reverting this to avoid
triggering the bug.
This rectangle has regular integer values, not fixed-point values.
So the old name was horribly wrong and misleading, (and yes I think
it was even I that had suggested it).
This patch was produced with the following (GNU) sed script:
sed -i -r -e 's/[ \t]+$//'
run on all *.[ch] files within cairo.
Note that the above script would have also created all the changes
from the previous commits to remove trailing whitespace.
This patch was produced with the following (GNU) sed script:
sed -i -r -e '/^[ \t]*\/?\*/ s/[ \t]+$//'
run on all *.[ch] files within cairo, (though I manually excluded
src/cairo-atsui-font.c which has a code line that appears as a comment
to this script).
Throw away old gradient code from within cairo-pattern.c in favor of using new support in pixman for gradients.
Update reference images for all tests involving gradients since the gradient output has changed slightly.
Add new function to be used instead of the abuse of pasing data=NULL to cairo_pdf_ft_font_write.
Just return a status now instead of a pointer to the written buffer, since cairo_pdf_ft_font_allocate_write_buffer should now be used instead when a pointer is needed.
Switch to use cairo_pdf_ft_font_allocate_write_buffer.
Fix use of uninitialized status value.
initialization just to keep the compiler quiet about possibly uninitialized variables.
Split CARIO_EXTEND_DEFAULT to indicate defaults for surfaces and gradients. The default for gradients is still the NEAREST beaviour.
Support the NEAREST mode, change NONE to be TRANSPARENT.
Update the extend mode code to use the pixman support.
Handle NEAREST, note that support for NONE is not yet handled for gradients.
Add a check for NEAREST but don't do anything useful.
Add a check for NEAREST but don't do anything useful.
Update a comment about the extend support in the image backend.
reviewed by: davidr, cworth, otaylor
Use _cairo_fixed_to_double and _cairo_double_to_fixed when converting between double and 16.16 fixed point. These functions round nicely, and solve some inconsistencies in rendering between my PPC and x86 machines. Thanks to Owen Taylor for suggesting the fix.
The above change does affect linear gradients slightly, but it seems worth it. Update the reference images to follow.
Don't change the status value if it already holds an error. This avoids destroying "first error" information and also avoids attempting to write read-only data, (eg. from nil objects).
Don't set an error on the same object if it is in an error state to begin with, just return immediately. This is intended to reduce (eventual) error prints to the first cause of an error for any given object, instead of a continual spew.
New function to return the current nil pattern depending on the status.
Add missing early bailout on surface->status with error propagation to the pattern.
Related cleanups for cairo_pattern_t:
Don't check other->status since this is a static function.
Add missing early bailout on other->status.
Cleanup identifier names.
Track rename of nil patterns.
Don't call _cairo_error for pre-existing errors.
Take care to initialize some fields to that _cairo_pattern_release_surface will work even after an error.
Track rename of cairo_solid_pattern_nil to cairo_pattern_nil.
New test to ensure that a file-not-found error will propagate from a surface, through a pattern, and onto a cairo_t.
Document cairo_pattern_add_color_stop_rgb, cairo_pattern_add_color_stop_rgba, cairo_pattern_set_matrix, and cairo_pattern_get_matrix.
Add note on user-space locking to the documentation for cairo_set_source().
_cairo_path_data_create_in_error _cairo_pattern_create_in_error _cairo_surface_create_in_error
Don't bother with extra check of other->status to anticipate and try to prevent cairo_surface_create_similar from returning through cairo_surface_create_in_error.
Just return &_cairo_surface|pattern|path_nil rather than _cairo_surface|pattern|path_create_in_error.
After checking surface->status from a cairo_<foo>_surface_create function, just return CAIRO_STATUS_NO_MEMORY since that's the only error we'll get from one of these create functions.
Remove unnecessary check for gstate == NULL;
Fix old check for image == NULL instead of image->base.status.
Add missing check of surface->image_base.status after creating surface->image.
Add missing check of surface->base.status after creating surface.
Change functions to return type of void:
cairo_scaled_font_extents cairo_surface_finish
Add new functions to query object status:
cairo_scaled_font_status cairo_surface_status
Implementation of new error handling scheme for cairo_surface_t and cairo_scaled_font_t.
Track change in return value of cairo_surface_finish.
Track change in _cairo_pattern_release_surface and also pass the appropriate pattern for each acquired surface. The previous backend mismatch was causing memory leaks.
Remove stale comment.
Add missing fclose to keep valgrind happy about memory leaks.
