Unification of d3d8 and d3d9 implementation

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4329 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4328 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4327 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4326 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4325 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4324 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4323 dfc29bdd-3216-0410-991c-e03cc46cb475

Fix compiling with _IRR_WCHAR_FILESYSTEM on MinGW. Thx @  alexzk for reporting. Note that I tested this with gcc 4.7.0 and only compiling is tested.
Also added some documentation that UNICODE must be set as well for compiling with _IRR_WCHAR_FILESYSTEM ('cause I always forget to set that at first otherwise...).

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4322 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4321 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4320 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4319 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4318 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4317 dfc29bdd-3216-0410-991c-e03cc46cb475

- Fixed some minor problems with zfail shadow volumes (problem with an infinite matrix still exist).

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4316 dfc29bdd-3216-0410-991c-e03cc46cb475

- Fixed ZFail shadows problem related to second - unnecessary shadow on the scene (only minor problems related to less depth exist).

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4315 dfc29bdd-3216-0410-991c-e03cc46cb475

- Fixed reset material issues in D3D9 driver which solve shadow bug issue visible in example no. 8 (small, movable shadow bug still exist).

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4314 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4313 dfc29bdd-3216-0410-991c-e03cc46cb475

Fix problem with 64bit systems, due to wrong return types used. Problem still persists with mingw, where the return type is used as before and is kept via ifdef.

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4312 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4310 dfc29bdd-3216-0410-991c-e03cc46cb475

- Fix a bunch of off-by one errors in irr::core::string in functions equals_substring_ignore_case, findFirst, findFirstChar, findNext, findLast, findLastChar, replace, remove and removeChars. This prevents some potential memory access errors, find functions no longer try to find the \0, replace no longer replaces the \0 and remove no longer tries to remove it (which did remove the last character instead).
- Fix a few new warnings in gcc.



git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4308 dfc29bdd-3216-0410-991c-e03cc46cb475

Fix crash with null-driver in combination with IrrDeviceLinux caused by my last check-in. Thx @ all reporting that.


git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4307 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4306 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4305 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4304 dfc29bdd-3216-0410-991c-e03cc46cb475

Remove event handling in device constructor. Seems it was not really the cause for the problems, only the missing size update of the d3d9 driver. To avoid complications, only the least change was left in.
Also improved the userclipplane test to really show errors in case of the original problem.

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4303 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4302 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4301 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4300 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4299 dfc29bdd-3216-0410-991c-e03cc46cb475

First try for MSVC2012 project files. conversion seems to have worked, though some warnings for 64bit targets were shown.

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4298 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4297 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4296 dfc29bdd-3216-0410-991c-e03cc46cb475

Replacing matrix4::setRotationRadiansLH and setRotationRadiansRH by setRotationRadians. Old functions had left handed and right handed just the wrong way round and  contained an additional bug that did lead to wrong rotations (mixing up y,z). A single function for this is enough (people can just flip the axis) and this way people who already used the old functions (which only got added in trunk) don't just get new behavior but some compile error. Thanks to Randajad for finding the error with the wrong results, giving a testcase and insisting that there's really a bug when I didn't see it on first try :-)


git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4295 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4294 dfc29bdd-3216-0410-991c-e03cc46cb475

Add list of test casses with numbers for easier access to single tests. We can also add some descriptions there.

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4293 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4292 dfc29bdd-3216-0410-991c-e03cc46cb475

Fix user clip planes. For D3D only a security measure in case plane3d changes. For OpenGL we fixed the user clip planes to world coords now. This is coherent with D3D usage and more natural. Bug fix suggested by heda.
I've also removed the clip plane upload from the projection matrix setting. I don't think it's needed there. At least my test cases don't need it. If someone finds a problem with this change, please report in the forum.

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4291 dfc29bdd-3216-0410-991c-e03cc46cb475

Process events and resize operation upon startup once to initialize everything, fixing user clip planes which are defined before first device->run()

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4290 dfc29bdd-3216-0410-991c-e03cc46cb475

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4289 dfc29bdd-3216-0410-991c-e03cc46cb475

Skip rendering of lines, points, polygons, and triangle fans, as these lead to crahses due to wrong access to the vertex lists. A fix would need major rewrite of the vertex cache, or at least some other render methods.

git-svn-id: svn://svn.code.sf.net/p/irrlicht/code/trunk@4288 dfc29bdd-3216-0410-991c-e03cc46cb475