Pages: << Previous 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 Next >>
Summary Period: 2000-04-15 to 2000-04-08 (Commits 2145-2194 of 12744)
Added example directory for world randomizer (data/dungeon) to VFS.
Added config file for DUNGEON example application.
    
2 lines of code changed in:
Added world randomizer interface.
16 lines of code changed in:
Backed out modification for defining OS_NT4/"WinNT"
0 lines of code changed in:
Fixed a typographical error from phase 1.  (I'm surprised GNU make didn't
complain about it.)
    
1 lines of code changed in:
*** Phase Six of Eric's Big Documentation Overhaul(tm) ***
-*- Completely eliminated all remaining warnings which were reported by
    TeX.  Mostly these involved overfull and underfull 'hbox'.  Now all
    conversions (TeX, HTML, Info) build warning-free.
-*- Added a 'pdfdoc' target to the makefile which converts the Crystal
    Space manual from Texinfo to PDF.  This complements the other output
    formats, PS, DVI, HTML, and Info.  (PDF generation still requires a bit
    more tweaking.  As it stands, it does not preserve the print margins,
    thus the printed material appears immediately adjacent to the border of
    the sheet of paper.)
-*- Overhauled the NeXT compilation and NeXT detailed information documents
    so that they make sense in the context of a larger manual.  Also
    sensibly merged the old INSTALL.NeXT and INSTALL.NeXT-Binary documents
    into a single section within the manual.
-*- Performed an overhaul of texinfo/intro/contrib.txi, the Crystal Space
    "contributors" document.  Unified its layout with the layout of
    authors.txi.  Also subjected it to a much more thorough Texinfo
    conversion.
-*- Overhauled resource.txi.  Provided much more up-to-date information
    regarding accessing Crystal Space in its various forms (such as CVS
    snapshots, etc.).
-*- Completely overhauled faq.txi to bring it much more up-to-date.  It was
    extremely outdated.
-*- Overhauled coding.txi and cvs.txi within the "Contributing" chapter.
    Some of the information in cvs.txi was also relevant to coding.txi in
    one form or another, so I added the information there as well.
-*- Major cleanup of a significant number of Texinfo document files.  (I
    would name them here, but the list would be too long :-).  Also
    performed a much more thorough Texinfo conversion on many of the
    existing documents.
-*- Finished overhauling the Simple tutorial so that the discussion now
    agrees with the sample code; rather than often being independent of it.
-*- Upgraded to texi2html version 1.62.  This version fixes six or seven
    additional bugs which I reported to the maintainer.  Unfortunately, it
    also introduces several new bugs (some of which are severe).  I have
    sent off appropriate bug reports and am awaiting response.  In the
    meantime, I manually patched the most severe problems myself.
-*- Way too many additional documentation changes to even begin to
    summarize them here.  You'll just have to take my word for it that a
    _lot_ has changed. :-)
    
3609 lines of code changed in:
Fixed a REAL bug in FastSqrt() that caused sqrt(32) to return 4e18.
In fact, any sqrt(power-of-two) was incorrect.
    
1 lines of code changed in:
New texture explo.jpg, (C) W.C.A. Wijngaards, have fun with it.
0 lines of code changed in:
Showing off more explosion options, new texture explo.jpg.
4 lines of code changed in:
csSprite now supports iParticle
65 lines of code changed in:
Using csVectors and iParticles now. csParticleSystem also implements iParticle.
187 lines of code changed in:
iParticle interface, added by W.C.A. Wijngaards.
9 lines of code changed in:
Update.
24 lines of code changed in:
Reverted back to 5MP cache (I accidentaly commited the 2MB setting)
1 lines of code changed in:
Removed ambient_white variables from csLight class. Now csWorld just adds
the AMBIENT_WHITE value to csLight::ambient_r/g/b and that's all.
    
18 lines of code changed in:
Added a bunch of missing historical entries regarding Seth's work on Squawk
51 lines of code changed in:
Patch for Mingw to work around unknown escape sequence error
   - Moved #define COMPUTE(comp) comment to position one line
     before void complete().
   - Minor modification to comment to maintain accuracy of
     commented documentation.
    
3 lines of code changed in:
Added RemoveHalo function.
12 lines of code changed in:
Bugfix in destructor, have to delete halo too.
2 lines of code changed in:
Update.
30 lines of code changed in:
Replaced 4-space tabs with spaces in one comment
4 lines of code changed in:
Implemented a killer enhancement on the texture cache, curtesy of
John Carmack :-) It improves performance a lot when all the visible
textures don't fit at once into the texture cache (e.g. with complex
levels, also those sudden slowdowns that happens sometimes while
you're walking). Some numbers, just to give you the feeling:
Entry data:
2.2Mb textures visible at once (e.g. flarge.zip, near the staircase).
First with texture cache of 10Mb (e.g. far enough for current frame):
11.976048 fps in both cases (old code and new code).
Now setting texture cache of 2MB:
5.025126 fps without improvement and 11.415525 (!!!!) with improvement.
Now setting texture cache of 1MB:
5.025126 fps without improvement (naturally) and 6.752194 fps with it
For real-world cases it should give even MORE performance: generally the
more the texture set occupies, the bigger is improvement. For example, with
the default cache size of 5MP (e.g. 10MB for 16bpp modes), if you ever
encountered the "too much textures in view" problem, you should notice
a huge performance increase.
    
54 lines of code changed in:
csVector::Insert() no longer asks memmove() to move zero bytes.
3 lines of code changed in:
csVector::Delete() no longer asks memmove() to move zero bytes.
4 lines of code changed in:
Update.
2 lines of code changed in:
moved recalculation of clipper from Draw to a separate method
2 lines of code changed in:
moved recalculation of clipper from Draw to a seperate Method
7 lines of code changed in:
Small fix.
1 lines of code changed in:
- Fundamental API change for sprites. Move() has been replaced with
  MovePosition() (relative move). SetMove() has been replaced
  with SetPosition() and GetOrigin() has been replaced with
  SetPosition(). This is both for 2D sprites and 3D sprites.
  Additinally GetPosition (previously GetOrigin) and
  MovePosition (previously Move) have changed direction!!!
  They are now correct.
- Also use SetPosition and MovePosition for csCollection and
  csThing now (to be consistant).
- Changed HitBeam to use 'const' input parms.
    
84 lines of code changed in:
Minor textual changes.
14 lines of code changed in:
Fixed an additional CD bug in WalkTest.
1 lines of code changed in:
- Fixed old bug with bots (divide by zero for Norm ()).
- Fixed another old bug with the missile getting stuck if you
  immediatelly fire after starting walktest.
- Cleaned up WalkTest a little (not much).
- Added walkcd.cpp file where all collision detection code
  related to WalkTest is placed.
- Added walkdemo.cpp file where all special demo code in WalkTest
  is added (ghost, skeletal tree, bots, missile, particles, ...)
- Added walkcmd.cpp file where all console command stuff is
  housed.
- Moved constructor and init stuff of WalkTest from keys.cpp
  to walktest.cpp where it belongs.
- keys.cpp is now much closer to really doing what the name
  suggests: handling keys, events, and movement.
- Fixed a bug in collision detection for WalkTest. The code
  to handle the resulting collisions incorrectly looped on 'hits'
  instead of 'num_our_cd'. The result of this was that it was
  rather possible to walk through walls or fall through floors.
  This occured very easily in STATBSP type levels but could
  also affect portal levels.
    
2427 lines of code changed in:
Fixed AddColor. Fixed destructor, UnlinkSprite is already done by ~csSprite().
1 lines of code changed in:
Fixed the ColorChange functionality
5 lines of code changed in:
Small comment change.
2 lines of code changed in:
- Moved deletion of particle systems before sprites so that
  we can't have crashes because of this.
- Did a little cleanup of the colldet stuff (i.e. mostly coding
  style and Doc++ comments).
- Fixed a bug with DrawPolygonFX/software which would crash
  for polygons that exceed screen size. It is still strange that
  this occurs though. Normally the engine should not let that
  through.
- Fixed csSector::IntersectSegment to support octree for faster
  checking. This also fixes FollowSegment.
- Rewrote csSector::HitBeam() to use the new IntersectSegment
  better. This function should now work better and faster than
  before. This is basicly the much requested TraceLine.
  HitBeam() also supports warping portals correctly now.
- Changed WalkTest to use the new HitBeam() function to click
  on objects.
    
185 lines of code changed in:
Created a custom PHP script (cvssnap.php) which allows access to the daily
CVS snapshots via HTTP.  Previously snapshots were only accessible via FTP.
Snapshots can now be downloaded from either of these two URLs:
    http://crystal.sourceforge.net/cvs-snapshots/
    ftp://crystal.sourceforge.net/pub/crystal/cvs-snapshots/
    
20 lines of code changed in:
Fixed SHM problem on solaris, when SHM fails it will retry without SHM.
14 lines of code changed in:
-*- Added CAMERA keyword to the new loader. Not to the old one though
-*- Added named camera positions support in the engine. Now START (...)
    is just an alias for CAMERA 'Start' (POSITION (...))
    
426 lines of code changed in:
- Finally fixed the bug in the dynamic object support for
  polygon trees. Sprites are now finally working correctly in
  STATBSP levels.
- Fixed a memory leak as well so that the polygon stubs are
  correctly freed when a sprite is removed.
    
18 lines of code changed in:
- Added 'DETAIL' keyword for THING and SIXFACE and CS_ENTITY_DETAIL
  flag to csThing. With this keyword you can indicate detail objects
  in CS. Currently the only effect this will have is excluding the
  thing from the octree and rendering it true the new and
  experimental DrawPolygonMesh. So don't start using it already
  unless you are interested in beta testing :-)
- Several objects implemented a 'CheckFlags' in the engine. I fixed
  a comment in CheckFlags(). The comment said that ALL flags in the
  given mask had to be true but the function itself returned true
  if SOME of the flags were true. I fixed the comment and added
  a new CheckAllFlags() function that really checks if ALL flags
  are set. This affects csThing, csPolygon3D and csLight.
- Added csFlags class to csutil. This is a very small class to make
  abstraction of the above settings for objects in the engine.
- Fixed bug in DrawPolygonMesh where num_vertices was used instead
  of num_polygons to loop over all polygons.
- Fixed bug in DrawPolygonMesh where the array of txt_handle
  should be iTextureHandle** instead of iTextureHandle*.
  Same for poly_texture.
- Implemented needed support for poly_texture in DrawPolygonMesh.
- Mapping of vertices to G3DPolygon structure was completely wrong.
- Several other fixes to DrawPolygonMesh to make it really working.
- Implemented first primitive support for rendering detail objects
  through DrawPolygonMesh. This is initial support for purposes
  of debugging DrawPolygonMesh only. It does not yet support all
  features (like fog and clipping) and it will certainly be
  slower than the old way to render.
- Changed iGraphics3D::SetTransform to use a csReversibleTransform
  instead of a csTransform.
- Changed iGraphics3D to use a csPlane3 instead of the current
  old structure.
    
353 lines of code changed in:
Fixed some Texinfo issues.
6 lines of code changed in:
Added some missing cross references.
3 lines of code changed in:
Update.
30 lines of code changed in:
- Added Dumper::dump for csBspTree2D.
- Implemented missing csSegmentArray::Get().
- Fixed a small bug in 2D bsp.
- Fixed back2front and front2back traversal in 2D bsp.
- Started further debugger on system to select octree splitter
  based on solid space intersection.
- Drasticly improved octree building but NOT with the solid
  space splitter. I still need to investigate that one.
  The new octree center selection just tries to minimize on
  polygons splits like before but instead of choosing random
  points only vertices of the polygons in the octree are
  considered. This tends to minimize splits greatly.
  Lighting improved about 20% with this. Rendering could have
  improved approximately the same amount.
    
54 lines of code changed in:
*** Phase Five of Eric's Big Documentation Overhaul(tm) ***
-*- Overcame a shortcoming in nodefix.pl where it was unable to repair a
    @menu if the menu was not preceded by a @node directive in the same
    file.
-*- Fixed a bug in nodefix.pl where it would neglect to blank out the
    "next" pointer of the last node at a particular depth.  This could be
    an especially serious problem if it was pointing at a completely
    defunct node.
-*- The nodefix.pl and docproc.pl scripts are now run with Perl's -w
    (enable warnings) option.  Eliminated several warnings produced by use
    of this switch.  Improved command-line option processing.
-*- The docs.mak makefile no longer passes --no-validate to 'makeinfo'.
    Now that my nodefix.pl script generates completely correct @node and
    @menu directives, --no-validate is no longer necessary.
-*- Restructured the Applications chapter in order to keep it from severely
    polluting the table of contents with useless junk and, to restrict the
    number of itsy-bitsy, teeny-weeny, nearly-useless HTML files which were
    generated by it.  Performed a more thorough Texinfo conversion on it.
-*- Overhauled dirhier.txi, the directory hierarchy documentation.  Brought
    it much more up-to-date.  Eliminated a significant amount of redundant
    information.  Performed a much more thorough Texinfo conversion on the
    document.
-*- Broke the (library) Dependencies node out of dirhier.txi into its own
    file (depend.txi) and its own section (rather than being a subsection
    of Directory Hierarchy).  Conceptually it is completely independent of
    Directory Hierarchy, so I am not sure why it was placed there
    originally.
-*- Upgraded to a newer revision of texi2html 1.61.  This revision
    addresses all of the pending bug reports which I had sent to the
    maintainer.
-*- Removed all of the uninteresting options from Crystal Space's custom
    texi2html.init.  This makes it much easier to see exactly which options
    are important for generation of Crystal Space's HTML documentation and
    also makes it vastly simpler to update to a newer version of texi2html.
    (Note that a copy of the original .init file is always textually
    incorporated into texi2html, itself, thus one can still see a list of
    all options if desired.)
    
427 lines of code changed in:
Fixed several errors in texinfo/build/platform/beos.tex, the BeOS 'build'
document and also rewrote several sections for increased lucidity.
    
107 lines of code changed in:
Added initial DrawPolygonMesh() code to 3D drivers.  Tested on all Linux
drivers.  Shouldn't interfere with anything since it is not called yet, but
dpmesh.cpp will have to be added to all 3D drivers for non-UNIX Makefile
builds.
    
81 lines of code changed in:
Added initial DrawPolygonMesh() code to 3D drivers.  Tested on all Linux
drivers.  Shouldn't interfere with anything since it is not called yet, but
dpmesh.cpp will have to be added to all 3D drivers for non-UNIX Makefile
builds.
    
178 lines of code changed in:
Animate your particles! Use the SetChangeXXX() functions to have the rotation, size, color, alpha change over time. For the explosion made lightfading and spritefading independent. Removed UpdateAll/DeleteAll, since a list of particlesystems is kept in the world.
137 lines of code changed in:
Added a Uint GetMixMode() function.
3 lines of code changed in:
Pages: << Previous 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 Next >>
![]()
Generated by StatCvs v0.2-dev