There's currently a regression bug in the tessellation code from
switching to the new tessellator. The bug is caused by
confusion in the comparator used to order events when there are
degenerate edges. This test is derived from the zrusin-another
performance test case.
Thanks to Mathieu Lacage for the recommendation of:
"Computational Geometry, Algorithms and Applications", M. de
Berg, M. van Kreveld, M. Overmars, M. Schwarzkopf
The perceptualdiff program was written by Hector Yee with contributions
from Scott Corley and Mick Weiss. It is hosted at http://pdiff.sourceforge.net
The source code added here was obtained by:
svn co https://svn.sourceforge.net/svnroot/pdiff/trunk pdiff
which gave revision 22 of the source code.
The perceptualdiff program is available under the terms of the GNU GPL,
so I've added a note to COPYING about this program, (and the fact that
it is auxiliary only, and does not affect the license of the implementation
of cairo itself).
It was added in commit 157663e1fd because there
was a serious bug in the pkg.m4 file from pkg-config 0.18. However, it was
fixed two days later. So, instead of requireing 0.18.1 or greater which is
not available on Nokia 770 SDK, we just ignore the broken pkg-config 0.18.
The problem only happens if one runs autogen.sh with against a broken
pkg-config. Configuring tarballs is not affected.
This patch removes a redundant call to skip_list_find()
that was being used to detect duplicate intersection events.
Instead, skip_list_insert() now takes an additional parameter
letting it know what to do with duplicates.
This fixes the failures of the new tessellator with the 3 tests:
bitmap-font, rectangle-rounding-error, and close-path
The problem was that identical edges from separate polygons
were not being added to the event queue, (because of a check
that was actually only intended to prevent an intersection
event from being scheduled multiple times).
This is the implementation as it cooked in the new-tessellator branch
available from:
git://people.freedesktop.org/~cworth/cairo
The file here comes from commit eee4faf79900be2c5fda1fddd49737681a9e37d6 in
that branch. It's sitting here not hooked up to anything in cairo yet,
and still with a main function with test cases, etc.
The files here are copied directly from the standalone skiplist module
available from:
git clone git://cworth.org/~cworth/skiplist
In particular the files come from the double branch and the following
commit on that branch:
8b5a439c68e220cf1514d9b3141a1dbdce8af585
Also of interest is the original skiplist module hosted by Keith Packard
that is the original implementation on which these files were based.
Since the cworth/skiplist branched off of keithp's, Keith has also
now implemented a doubly-linked variant which might be interesting for
further simplification of the code. See:
git clone git://keithp.com/git/skiplist
and the double-link branch there.
We'll need to rewrite this test (and the way cairo is compiled for
testing) before it can exercise the bug of interest. In the
meantime, let's just give it an "expected failure" status.
This is the same fix as in the recent commit (166dffc96a)
but for the same idiom that appears in the operator-clear,
operator-source, and unbounded-operator tests. This fixes
the failures that were introduced by the recent switch to
_cairo_lround.
After changing _cairo_gstate_show_glyphs and _cairo_gstate_glyph_path to use
this function, we see a significant speedup due to the elimination of redundant
FP calculations.
