summaryrefslogtreecommitdiffstats
path: root/libdimension
Commit message (Collapse)AuthorAgeFilesLines
* Fix race when one worker thread errors out.Tavian Barnes2010-11-171-4/+13
|
* Make the bottom [0 0 0 1] of affine transformation matricies implicit.Tavian Barnes2010-11-163-76/+60
|
* End sentences with periods.Tavian Barnes2010-11-1518-83/+86
|
* Document libdimension with Doxygen.Tavian Barnes2010-11-1479-768/+1951
|
* Revert "Add support for jemalloc."Tavian Barnes2010-11-121-4/+0
| | | | | This reverts commit bac926b1e54696beee1d84834876cfadd9939132. Just use LD_PRELOAD if you want it.
* Standard compliance.Tavian Barnes2010-11-107-36/+59
| | | | Apparently ssize_t isn't a standard type...
* Use Newton's method when the root bound is degenerate.Tavian Barnes2010-11-102-5/+39
|
* Stability fix for dmnsn_bisect_root().Tavian Barnes2010-11-091-4/+9
| | | | | | | When one of the bounds is close to a different root, make sure the result is more accurate than that bound. Otherwise we find the wrong root, and eventually hang. This could be seen with a 1920x1080 render of demo.pov, for example.
* Don't just use quick_color for solid_color pigments.Tavian Barnes2010-11-092-3/+24
| | | | | quick_color should be able to be used to change an object's color for low-quality-renders.
* Add quick_color to pigments.Tavian Barnes2010-11-084-30/+27
|
* Add gradient pattern.Tavian Barnes2010-11-083-0/+52
|
* Fix double-init bug for inherited textures.Tavian Barnes2010-11-083-7/+18
|
* Optimize DMNSN_ARRAY_FOREACH_REVERSE() a bit.Tavian Barnes2010-11-081-3/+3
|
* Fix color_map values for negative parameters.Tavian Barnes2010-11-081-0/+4
|
* Add dmnsn_orange color.Tavian Barnes2010-11-082-1/+8
|
* Support checkered pigments in dimension.Tavian Barnes2010-11-071-1/+7
|
* Don't leak patterns in color_map pigments.Tavian Barnes2010-11-061-1/+10
|
* Add pattern framework, impelement checker pattern.Tavian Barnes2010-11-0610-1/+347
|
* Add more in dmnsn_line_add_epsilon().Tavian Barnes2010-11-061-1/+1
|
* Factor out transformation code from object callbacks.Tavian Barnes2010-11-069-64/+81
|
* Be more lenient about the root bracketing in dmnsn_bisect_root().Tavian Barnes2010-11-041-1/+14
|
* Fix ++i/--i thinko.Tavian Barnes2010-11-031-1/+1
|
* Pull < DMNSN_NBINOM test outside of loop.Tavian Barnes2010-11-031-20/+29
|
* Slight ray-AABB intersection test optimization.Tavian Barnes2010-11-021-1/+1
|
* Small polynomial.c cleanup.Tavian Barnes2010-11-011-4/+8
|
* Slight polynomial base case optimizations.Tavian Barnes2010-11-011-24/+25
|
* More numerical tweaks.Tavian Barnes2010-11-012-1/+12
|
* Numerical fixes for polynomial.c.Tavian Barnes2010-10-311-89/+109
|
* Be const-correct in polynomial.{c,h}.Tavian Barnes2010-10-302-11/+11
|
* Remove debug print from polynomial.c.Tavian Barnes2010-10-301-2/+0
|
* Fix dmnsn_print_polynomial again.Tavian Barnes2010-10-291-1/+3
|
* Only return non-zero roots of polynomials.Tavian Barnes2010-10-292-17/+16
|
* Bound torii like POV-Ray.Tavian Barnes2010-10-281-3/+51
|
* Handle narrow ranges in dmnsn_bisect_root().Tavian Barnes2010-10-281-0/+3
|
* Fix dmnsn_print_polynomial().Tavian Barnes2010-10-281-20/+9
|
* Precalculate some binomial coefficients.Tavian Barnes2010-10-281-9/+28
|
* Use the false position method to bisect isolating intervals.Tavian Barnes2010-10-281-30/+20
|
* Use Uspensky's method to find multiple roots at once.Tavian Barnes2010-10-281-57/+63
|
* Add polynomial test.Tavian Barnes2010-10-271-4/+3
|
* Fix typo in comment.Tavian Barnes2010-10-261-1/+1
|
* Add torii to libdimension.Tavian Barnes2010-10-263-1/+128
|
* Add numerical polynomial solver based on Uspensky's algorithm.Tavian Barnes2010-10-267-32/+411
|
* Be more conservative about when to invert by partitioning.Tavian Barnes2010-10-241-3/+3
|
* Correctly propagate errors in raytracing worker threads.Tavian Barnes2010-10-241-7/+3
|
* Constness fix for cylinder.c.Tavian Barnes2010-10-241-2/+2
|
* Fix inside calculation for cones.Tavian Barnes2010-10-221-2/+5
|
* Change initial array capacity to 2.Tavian Barnes2010-10-211-11/+11
| | | | | A quick experiment shows that arrays average less than 2 elements, and this helps a bit with the memory-hogginess of the parser.
* Don't call malloc() in the default error handler.Tavian Barnes2010-10-201-5/+1
|
* Add cone support to libdimension.Tavian Barnes2010-10-202-17/+45
|
* Missing `void' in argument list.Tavian Barnes2010-10-201-1/+1
|