Activity
From 2006-10-04 to 2006-10-17
2006-10-16
- 00:57 Revision d528a05b (github): Added missing new line at file end.
- 00:21 Revision 4fd72f67 (github): Extracted the colour-keying and image alpha checking into separate routines and moved to gl_tex.c
2006-10-15
- 20:32 Revision 929600c1 (github): Began splitting gl_tex.c into smaller files. Separated all the texture management routines out and into gl_texmanager.c. gl_tex.c now consists entirely of image manipulation algorithms.
- 20:27 Revision c307326e (github): Changed to UTF-8.
- 16:57 Revision 3ec82c92 (github): Remove duplicate flare texture.
2006-10-14
- 23:43 Revision 1a6dfeeb (github): Fixed segv problem with "rend-tex 2" (thanks yagisan).
- Fixed lighting discrepancies (light adaptation was being applied twice for mobj sprites and models). Minor tweaks in ...
- 17:44 Revision 202f52db (github): Moved more decisions affecting rendpolys further up the logic hierarchy to make better use of surface_t.
- Fixes for bugs with shiny textures and polyobjects introduced by recent changes.
- 06:41 Revision 364ae81a (github): Implemented a new rendering mode to help debug lighting issues. Setting cvar "rend-tex 2"; Doomsday will render all world surfaces using the gray.png texture.
- 04:28 Revision 105added (github): Store pre calculated floating point coords for vertexes in seg_t. Save having to FIX2FLT() every time we need them (several thousand times per frame).
- 02:12 Revision d4009431 (github): Fixed an inaccuracy in the defined value of PI.
- 02:03 Revision 9882d13b (github): Replaced numerous instances of divide by 255 in time critical code with multiplication by reciprocal255 (the value taken for this is from the Risen3D source).
- 00:57 Revision a4194d89 (github): Changed to UTF-8.
- 00:49 Revision ab83fec3 (github): Use Memory management wrappers. Clean up.
- 00:06 Revision 7ddec61a (github): More fixes for warnings produced by gcc.
2006-10-13
- 20:22 Revision 65cdd2b9 (github): Fixed a few more compiler warnings (gcc).
- 19:41 Revision 54911b8b (github): Use the seg facing flag instead of recalculating in the dynlight code (optimize).
- 18:58 Revision aa568ddd (github): Fixed typo.
- 18:50 Revision 8368038f (github): Use Memory management wrappers. Clean up.
- 06:49 Revision 19a53221 (github): Moved (f/dd/div)vertex_t members (x,y (z)) into pos[VX, VY (VZ)].
2006-10-12
- 22:42 Revision c56d45aa (github): Shutup gcc (fixed compiler warnings). Cleanup decor light code.
- 21:07 Revision 357a101e (github): Cleanup.
- 21:00 Revision ff6b31a5 (github): Fixed "/tmp/doomsday/engine/portable/src/p_particle.c:279: error: static declaration of 'P_Uncertain' follows non-static declaration".
- 20:42 Revision 45238baa (github): Cleanup.
- 20:16 Revision fdcea1f6 (github): Changed to UTF-8.
- 19:35 Revision dcbd275d (github): Further clean up. Use unsigned types for indices where possible (self-document). Fixed signed/unsigned mismatch in comparison for checking the status of the matched word list.
- 19:30 Revision e0eefd2f (github): Fixed "/tmp/doomsday/engine/portable/include/con_bind.h:50: error: previous declaration of 'B_SetBindClass' was here...".
- 01:29 Revision 853acd59 (github): OPTIMIZE: Changed wall seg rendering. Previously we were calculating whether each wall seg, in all visible subsectors faced the viewer at least three times per seg per frame unnecessarily. This calculation is now done once per frame per seg in visible subsectors and seg_t->flags updated with SEGINF_FACINGFRONT if true.
- Cleaned up con_bind.c, event bindings are now put into separate arrays depending on event_t->type. This plus the othe...
2006-10-10
- 20:32 Revision 3edb118a (github): Enable dropengl to be fully MAO built
- 20:23 Revision cffa41ad (github): In a MAO build divide jhexen into about 4 'modules' roughly along subsystem lines. (too much effort at this time to make it fully MAO, but this should allow optimisation along subsystems).
- 18:50 Revision bd49b144 (github): In a MAO build divide jheretic into about 3 'modules' roughly along subsystem lines. (too much effort at this time to make it fully MAO, but this should allow optimisation along subsystems). p_oldsvg and m_random currently not MAO.
- 17:39 Revision be0d5914 (github): In a MAO build divide jdoom into about 3 'modules' roughly along subsystem lines. (too much effort at this time to make it fully MAO, but this should allow optimisation along subsystems). p_oldsvg and m_random currently not MAO.
- 12:17 Revision d89b9170 (github): In a MAO build divide doomsday into about 10 'modules' roughly along subsystem lines. (too much effort at this time to make it fully MAO, but this should allow optimisation along subsystems). Arch specific is not MOA.
2006-10-09
- 20:53 Revision 182f7b71 (github): Add Module At Once options for jDoom, jHeretic and jHexen.
- 20:39 Revision 64ff66a6 (github): These are more correctly known as lump groups
- 17:51 Revision ee7cea1b (github): Add file encoding and copyright info to the output. Results as always at http://www.dengine.net/audit
- 15:19 Revision 0c0d7f10 (github): Fix error: previous declaration of 'numSources' was here
- 15:06 Revision e8aad7d6 (github): Fix error: previous declaration of 'groups' was here
- 14:50 Revision c6a7d8f7 (github): Fix error: conflicting types for 'blockSize'
- 14:09 Revision cad963e0 (github): Fix error: conflicting types for 'freezeRLs'
- 13:55 Revision 0bef87a3 (github): Convert to UTF-8
- 13:45 Revision b9a4987f (github): Create a Module-At-Once build option. OFF by default. The purpose of this is to enable the compiler to see the whole source for a particular module and as such, make better choices when optimising. It no doubt will also help shake out bugs too
- 11:25 Revision aa45bb3d (github): Remove semi-function profile option
- 11:13 Revision 1c8d5b02 (github): cleanups
- 10:46 Revision 985caf5e (github): Finalise CFLAGS: Explain why chosen. Perhaps, Win32 and OSX might be interested in the rational and implement something similar for their packages
2006-10-08
- 20:31 Revision 09de61b1 (github): More rules file cleanup. Dump a truckload of warnings on suspect code. Try to extract every last bit of performce we can from gcc
- 20:14 Revision 23b1f93e (github): Fixed GCC compiler warnings.
- 19:46 Revision 5656a3df (github): Fixed compilation warnings.
- 19:10 Revision bf4df569 (github): Began dividing con_main.c into smaller chunks. We now have con_data.c which implements the various console databases (cvars, ccmds, knownwords, aliases) and rend_console.c which contains the console drawing routines.
- Since the various console databases were already being sorted I replaced their search mechanisms with those using a b...
- 19:03 Revision 35de1c0b (github): Optimization: Rendpolys are no longer allocated on the stack each and every time they are needed. Added a rendpoly allocation mechanism which allows the same rendpolys to be re-used whenever possible.
- Number of vertices are also allocated dynamically at this time, this meant I was able to remove the RL_MAX_POLY_SIDES...
- 18:54 Revision 6f9a3ee0 (github): Cleanup. Use the memory allocator wrappers (e.g. M_Malloc). Minor tweaks and optimizations.
- 18:34 Revision 6248c948 (github): Cleanup. Use the memory allocator wrappers (e.g. M_Malloc). Minor tweaks to reduce per-frame processing work.
- 18:29 Revision 31daa1f3 (github): Cleanup. Use the memory allocator wrappers (e.g. M_Malloc). Minor tweaks. Updated style.
- 18:27 Revision 47c2e478 (github): Cleanup. Use the memory allocator wrappers (e.g. M_Malloc). Minor tweaks. Updated style.
- 18:24 Revision be28d66a (github): Cleanup. Added function headers. Fixed issue of returned false positives when searching for nodes when a substring matches.
- 18:21 Revision 1c5d886a (github): Cleanup. Use the memory allocator wrappers (e.g. M_Malloc).
2006-10-07
- 17:42 Revision 2bda0b6c (github): fix bug report script. Start dragging the package kicking and screaming towards the beta5 release
- 08:48 Revision 9d047732 (github): Shove my CFLAGS down Cmakes throat
2006-10-06
- 19:27 Revision b26cce43 (github): zip is no longer a build dep
- 19:24 Revision 10e4f95f (github): remove typos that prevent building
- 18:28 Revision fe1fa371 (github): No - really make timidity a hard dependency this time.
- 18:25 Revision ae0be93b (github): Whoops, must not forget the packaging hooks or the jpacks will break before the snowberry migration. Not good :( fix up some thinkos in the reportbug script. Make timidity a hard dependency as per http://forums.newdoom.com/showthread.php?t=31639
Also available in: Atom