The expected behaviour for masking in Cairo is to set the mask according
to the current active matrix and apply unchanged to the masked surface.
In SVG, the mask element is bound to the masked object and thus the local
matrix from that image object applies to the nested mask as well.
Attached is a small patch for substracting the matrix of a image
surface from the matrix of the mask to comply to Cairo's behaviour.
I did not test for other stuff like vectors or text and would expect this
part is incomplete.
Quote Kai-Uwe Behrmann:
"The expected behaviour for masking in Cairo is to set the mask according
to the current active matrix and apply unchanged to a to be masked
surface.
In SVG the mask element is bound to the masked object and thus the local
matrix from that image object applies to the nested mask as well."
This is a test case to exercise applying a mask to an image under
separate transformations.
Original patch by Kai-Uwe Behrmann, altered to run the test against all
backends (where it causes poppler to crash on my machine <evil grin>).
We now compile cairo-system.c twice when building static library, but
that's fine because cairo-systemc. includes no other code. Indeed
that's why cairo-system.c was born in the first place.
This is where DLL initialization/finalization should be done for example.
Moved the one for win32. For OS/2 just left a comment as the code needs
more work.
This change simplifies building shared and static libraries in the win32
makefiles.
When I designed this first I thought people can define make vars on the
command line to override default features. That works. However, it's
natural to want to define the list of features in the toplevel Makefile
and not on command line. So, we don't override user's var now. They
can set a make feature var to 0 to disable it.
Previously cairo-mutex-list-private.h assumed that every use of the file
will produce a statement for each mutex by including a semicolon after
each. But some uses (like enumerating all static mutexes in an array
for example, can't be implemented with the semicolon in place. So, move
the semicolon out to the users of the header file.
The macros CAIRO_BEGIN_DECLS and CAIRO_END_DECLS are declared in two
places: config.h and cairo.h. On Win32 build there is no config.h. So,
we can't rely on pulling CAIRO_BEGIN_DECLS from there. Hence, we now:
* Not add those declarations to config.h anymore,
* Test that every cairo private header includes some other cairo
header before any other includes.
These two are fairly enough to ensure that cairo.h is seen from all
private headers. There's still the case of headers not including
any header file at all. I'll fix that later.
Previously I made the mutex implementation code to use Windows mutexes
conditional on HAVE_WINDOWS_H. This was suppposed to be a more wine-friendly
version of checking for a Windows platform. However, it also meant that
when compiling without config.h, one had to stuff a #define for HAVE_WINDOWS_H
somewhere. Traditionally it was put in cairo-features-win32.h, and that
practice was carried over even after my build system rewrite.
Somewhere along the road I also decided that the mutex implementation header
does not need to include anything other than config.h. So the hacked up
HAVE_WINDOWS_H in cairo-features-win32.h was not being seen by its lone
user.
We now condition on any of HAVE_WINDOWS_H and _MSC_VER. So the hack can be
removed altogether.
The tessellator is well-proven now. However, the sweep-line validator
consumes around 50% of the total time required to draw the fractal
Pythagoras tree (the leaves are sub-pixel rectangles, so lots of edges
to sweep through). So disable the validator, but keep it available for
debugging.
Avoid the overhead in sorting the edges within
_cairo_traps_tessellate_convex_quad() by using our prior knowledge that we
have a simple rectangle and construct the trap from the extreme points.
Avoid the overhead of sorting the rectangle in
_cairo_traps_tessellate_convex_quad() by constructing the trap directly
from the line segment. This also has secondary effects in only passing
the non-degenerate trap to _cairo_traps_add_trap().
For rectilinear Hilbert curves this makes the rectilinear stoker over 4x
faster.
The rectilinear stroke finalized the cairo_traps_t passed to it - which
was then subsequently used without re-initialized. So instead of
finalizing the structure, just remove any traps that we may have added
(leaving the limits and memory intact).
Carl had this confusing problem today making the snapshot, which was
"make release-publish" erring. This commit reworks the release-publish
machinery to enforce serialization of some of the tasks such that it
works with make -j. Don't expect any speedup though as "make distcheck"
is inherently sequential.
The cairo_headers variable in Makefile.win32 was unused and
from before the new build system was put in place. One can
use enabled_cairo_headers if need be.
By introducing $(top_srcdir)/cairo-version.h I tried to make it
unnecessary to rebuild the entire library when the version changes.
That comes handy around releases and when bisecting.
However this was not working as the version info was also ending
up in config.h and hence forcing a full rebuild upon change.
I now undefine the PACKAGE_VERSION* macros that autoconf defines,
so config.h does not change when cairo-version.h changes.
Now that our build system can compute a list of all cairo files,
we prefer to use that in the tests instead of running "find".
However, we were not switching to $(srcdir) so the test was not
finding source files in out-of-tree builds. Noticed that from
the grep error in distcheck.
We used to do a full doc build to just to run the coverage test. That's
way too slow to expect people to run regularly. Instead now we just do the
source code scanning part of the doc build system that is just enough to
know if all symbols are documented.
A full doc build can be done as always by invoking "make doc", and indeed
will be called as part of "make dist" or "make distcheck".
Copy/paste error meant we were not checking cairo-undeclared.txt there.
Indeed the test was passing even though Carl forgot to remove lcd-fitlering
API from docs. The test fails now.
Note 1.7 API changes: Remove lcd_filter, cairo_has_show_glyphs, and
add cairo_text_cluster_flags_t. Also mention a few major bug fixes,
Behdad's build-system rework, and Chris's customary robustness
improvements.
During the cairo summit it was decided that this API is to freetype-
specific to be in the general cairo interface for now. This will
likely come back again soon as a cairo_ft-specific interface.
