]> git.mxchange.org Git - simgear.git/log
simgear.git
20 years agoFix a typo.
ehofman [Mon, 11 Oct 2004 07:56:08 +0000 (07:56 +0000)]
Fix a typo.

20 years agoFrederic: Ignore display lists when using the blend animation.
ehofman [Mon, 11 Oct 2004 07:37:42 +0000 (07:37 +0000)]
Frederic: Ignore display lists when using the blend animation.

20 years agoAdding the panel was a step too far for Linux, causing a segfault.
ehofman [Sun, 10 Oct 2004 19:38:46 +0000 (19:38 +0000)]
Adding the panel was a step too far for Linux, causing a segfault.

20 years agoRemove a phantom makeDList call (probably an old one from my code.
ehofman [Sun, 10 Oct 2004 19:16:25 +0000 (19:16 +0000)]
Remove a phantom makeDList call (probably an old one from my code.

20 years agoCheck for the plib version when using display lists, just to be sure.
ehofman [Sun, 10 Oct 2004 19:05:25 +0000 (19:05 +0000)]
Check for the plib version when using display lists, just to be sure.

20 years agoFrederic: Include FGPanelNode in the display list generation process.
ehofman [Sun, 10 Oct 2004 18:43:11 +0000 (18:43 +0000)]
Frederic: Include FGPanelNode in the display list generation process.

20 years agoFrederic: Use display lists for 3d models also.
ehofman [Sun, 10 Oct 2004 17:49:27 +0000 (17:49 +0000)]
Frederic: Use display lists for 3d models also.

20 years agoReverse the declaration order. jpgRenderFrame (formerly known as trRenderFrame) is...
ehofman [Wed, 6 Oct 2004 09:57:31 +0000 (09:57 +0000)]
Reverse the declaration order. jpgRenderFrame (formerly known as trRenderFrame) is now declared as a NULL function pointer and assignment of the proper function is now done in FlightGear (jpgRenderFrame=FGRenderer::update).

20 years agoDavid Luff:
ehofman [Thu, 30 Sep 2004 09:43:11 +0000 (09:43 +0000)]
David Luff:

The one-liner removes a lot of re-definition warnings on Cygwin.

20 years agoSmall update for future use.
ehofman [Sun, 19 Sep 2004 09:08:12 +0000 (09:08 +0000)]
Small update for future use.

20 years agoExpose the size() method for locked and blocking thread queues.
curt [Wed, 15 Sep 2004 15:28:58 +0000 (15:28 +0000)]
Expose the size() method for locked and blocking thread queues.

20 years agoFix another case where the direction vector is not initialized which can
curt [Wed, 15 Sep 2004 15:28:14 +0000 (15:28 +0000)]
Fix another case where the direction vector is not initialized which can
lead to openal "inrange" assertions, crashing FlightGear.

20 years agodirection vector needs to be initialized, otherwise garbage data could cause
curt [Fri, 10 Sep 2004 20:44:42 +0000 (20:44 +0000)]
direction vector needs to be initialized, otherwise garbage data could cause
openal to generate an assertion, aborting the top level app.

20 years agoTweaks for 0.3.7-pre1
curt [Fri, 10 Sep 2004 18:16:06 +0000 (18:16 +0000)]
Tweaks for 0.3.7-pre1

20 years agoAdd support for parsing an xml stream from an in memory buffer, rather than
curt [Fri, 10 Sep 2004 15:57:52 +0000 (15:57 +0000)]
Add support for parsing an xml stream from an in memory buffer, rather than
just from a specified file name.

20 years agoAdd support for audio orientation: direction and cone definition. This currently...
ehofman [Wed, 8 Sep 2004 11:15:57 +0000 (11:15 +0000)]
Add support for audio orientation: direction and cone definition. This currently only works for internal view and tower view because those set the listener position correctly.

20 years agoRemove a left over of ancient times.
ehofman [Thu, 19 Aug 2004 12:18:20 +0000 (12:18 +0000)]
Remove a left over of ancient times.

20 years agothis simple patch will enable the direct use of Point3D::get_n() instead of making...
ehofman [Tue, 17 Aug 2004 08:31:51 +0000 (08:31 +0000)]
this simple patch will enable the direct use of Point3D::get_n() instead of making duplications to call sgCartToGeod

20 years agothis simple patch will prevent to have fog in Clear cload layers.
ehofman [Tue, 17 Aug 2004 08:28:35 +0000 (08:28 +0000)]
this simple patch will prevent to have fog in Clear cload layers.

20 years agochange the log level of the initialization of OpenAL. This prevent a console popup...
ehofman [Sun, 15 Aug 2004 09:24:12 +0000 (09:24 +0000)]
change the log level of the initialization of OpenAL. This prevent a console popup for no reason on Windows.

20 years agoTweaks for official 0.3.6 release.
curt [Thu, 29 Jul 2004 21:36:41 +0000 (21:36 +0000)]
Tweaks for official 0.3.6 release.

20 years agoMake gcc 2.95.* compile again.
ehofman [Thu, 29 Jul 2004 08:30:10 +0000 (08:30 +0000)]
Make gcc 2.95.* compile again.

20 years agoUse the SimGear default notation.
ehofman [Wed, 28 Jul 2004 14:13:16 +0000 (14:13 +0000)]
Use the SimGear default notation.

20 years agoIRIX fixes (at least).
ehofman [Wed, 28 Jul 2004 13:59:57 +0000 (13:59 +0000)]
IRIX fixes (at least).

20 years agoIRIX fix (at least).
ehofman [Wed, 28 Jul 2004 13:16:54 +0000 (13:16 +0000)]
IRIX fix (at least).

20 years agoTweaks for 0.3.6-pre3
curt [Tue, 27 Jul 2004 21:18:14 +0000 (21:18 +0000)]
Tweaks for 0.3.6-pre3

20 years agoFix a minor warning message.
curt [Sat, 24 Jul 2004 19:21:15 +0000 (19:21 +0000)]
Fix a minor warning message.

20 years agoCorrect a typo that produces segfault during cleanup on some systems.
ehofman [Thu, 22 Jul 2004 08:04:18 +0000 (08:04 +0000)]
Correct a typo that produces segfault during cleanup on some systems.

20 years agoTweaks for 0.3.6-pre2
curt [Wed, 21 Jul 2004 21:22:21 +0000 (21:22 +0000)]
Tweaks for 0.3.6-pre2

20 years agoFrederic Bouvier:
ehofman [Wed, 21 Jul 2004 10:52:40 +0000 (10:52 +0000)]
Frederic Bouvier:

I just discovered this : state() is not valid when _level==0,
because it is doing:
{ return _state_stack[_state_stack.size() - 1]; }
and is returning a wrong index fetch ( 0 - 1 ) == -1

Moreover, st is not used when _level==0 so the line can be moved
to a safer place.

20 years agoTest for alutInit in libopenal.so and in libalut.so for Cygwin builds.
ehofman [Mon, 19 Jul 2004 14:59:51 +0000 (14:59 +0000)]
Test for alutInit in libopenal.so and in libalut.so for Cygwin builds.

20 years agoalut is now part of libopenal.so
ehofman [Sat, 17 Jul 2004 14:01:00 +0000 (14:01 +0000)]
alut is now part of libopenal.so

20 years agoWeak excuse, but it's getting late. Do clipping right this time.
ehofman [Fri, 16 Jul 2004 21:00:46 +0000 (21:00 +0000)]
Weak excuse, but it's getting late. Do clipping right this time.

20 years agoClip to volume to 1.0 if necessary.
ehofman [Fri, 16 Jul 2004 20:36:01 +0000 (20:36 +0000)]
Clip to volume to 1.0 if necessary.

20 years agoTweaks in preparation for the 0.3.6-pre1 release.
curt [Wed, 14 Jul 2004 22:01:15 +0000 (22:01 +0000)]
Tweaks in preparation for the 0.3.6-pre1 release.

20 years agoMake sure that a directory name containing a '.' doesn't screw up the ::base() and...
ehofman [Mon, 5 Jul 2004 16:39:02 +0000 (16:39 +0000)]
Make sure that a directory name containing a '.' doesn't screw up the ::base() and ::extension() functions.

20 years agoRemove the need to append a trailing / to the texture-path string.
ehofman [Mon, 5 Jul 2004 13:40:37 +0000 (13:40 +0000)]
Remove the need to append a trailing / to the texture-path string.

20 years agoAllow the use of a <texture-path/> tag to specify the directory containing the textur...
ehofman [Mon, 5 Jul 2004 11:42:19 +0000 (11:42 +0000)]
Allow the use of a <texture-path/> tag to specify the directory containing the textures for this model.

20 years agoUse #elif instead of #else if
ehofman [Sat, 3 Jul 2004 11:16:53 +0000 (11:16 +0000)]
Use #elif instead of #else if

20 years agoFrederic Bouvier:
ehofman [Sat, 3 Jul 2004 10:59:00 +0000 (10:59 +0000)]
Frederic Bouvier:

I have 3 issues that are fixed by this set of patches.

1. In extensions.cxx
  #else if !defined( WIN32 )  must be changed by
  #elif !defined( WIN32 ) because the text after #else
  seems to be ignored

2. banner is not available on windows, only cygwin

3. ANSI escape sequences are not interpreted on the
   windows console. We just have garbage that is hard
   to read.

20 years agoIf we pass in a position or velocity of nan, openal will generate an assertion.
curt [Thu, 1 Jul 2004 19:05:37 +0000 (19:05 +0000)]
If we pass in a position or velocity of nan, openal will generate an assertion.
Under rare circumstances we could encounter a non-cooperative external fdm
that could trigger this condition.  This patch catches the problem and returns
rather than letting FG get killed.

20 years agoSome small updates.
ehofman [Sun, 27 Jun 2004 08:09:56 +0000 (08:09 +0000)]
Some small updates.

20 years agoWork around a broken dlopen/dlclose Linux implementation.
ehofman [Sun, 27 Jun 2004 07:49:40 +0000 (07:49 +0000)]
Work around a broken dlopen/dlclose Linux implementation.

20 years agoFrederic Bouvier:
ehofman [Fri, 25 Jun 2004 10:59:59 +0000 (10:59 +0000)]
Frederic Bouvier:

this patch correct the cloud repositionning that
made them going against the wind. Now the clouds
and the windsock show the same direction.

20 years agoIgnore generated files.
david [Tue, 15 Jun 2004 12:47:52 +0000 (12:47 +0000)]
Ignore generated files.

20 years agoBe a bit anoying (and a tad more descriptive) about OpenAL volume errors for some...
ehofman [Sun, 13 Jun 2004 11:59:36 +0000 (11:59 +0000)]
Be a bit anoying (and a tad more descriptive) about OpenAL volume errors for some time.

20 years agoFrederic Bouvier:
ehofman [Sat, 12 Jun 2004 21:03:45 +0000 (21:03 +0000)]
Frederic Bouvier:

Do state sorting by material before adding primitives
in the tile branch. I thought I could see a bit of
improvement in framerate but it is not the case on my setup.
I don't see a degradation though

20 years agoFrederic Bouvier:
ehofman [Mon, 7 Jun 2004 18:49:32 +0000 (18:49 +0000)]
Frederic Bouvier:

I am still experimenting with the
code. Here is so far the best I could achieve. The
dark aspect of clouds at dusk or dawn is far better
than the problems of transparency of the previous
version.

20 years agoMingW fix.
ehofman [Mon, 7 Jun 2004 12:43:29 +0000 (12:43 +0000)]
MingW fix.

20 years agoAdd an 'add(str)' function that adds a search path separator and appends the str.
ehofman [Mon, 7 Jun 2004 09:50:32 +0000 (09:50 +0000)]
Add an 'add(str)' function that adds a search path separator and appends the str.

20 years agoTweak the doppler effect.
ehofman [Fri, 4 Jun 2004 16:59:12 +0000 (16:59 +0000)]
Tweak the doppler effect.

20 years agoRevert the previous patch. There was already a class availble that allows for that...
ehofman [Fri, 28 May 2004 08:42:59 +0000 (08:42 +0000)]
Revert the previous patch. There was already a class availble that allows for that. It's just that the description doesn't explain too much.

20 years agoTweak the bump-mapped 2d cloud color a bit.
ehofman [Thu, 27 May 2004 18:00:03 +0000 (18:00 +0000)]
Tweak the bump-mapped 2d cloud color a bit.

20 years agoAdd the possibillity to parse a user data pointer to getter and satter functions...
ehofman [Thu, 27 May 2004 13:03:01 +0000 (13:03 +0000)]
Add the possibillity to parse a user data pointer to getter and satter functions. This adds a convenient way to get the 'this' pointer to the static functions.

20 years agoUse a different coloring scheme.
ehofman [Tue, 25 May 2004 07:58:33 +0000 (07:58 +0000)]
Use a different coloring scheme.

20 years agoMacOSX fix that never got applied before.
ehofman [Fri, 21 May 2004 16:27:16 +0000 (16:27 +0000)]
MacOSX fix that never got applied before.

20 years agomingw32 fix
ehofman [Fri, 21 May 2004 15:07:25 +0000 (15:07 +0000)]
mingw32 fix

20 years agoUpdates from Frederic to use 2 texture units and a color blend function rather than...
ehofman [Fri, 21 May 2004 14:50:49 +0000 (14:50 +0000)]
Updates from Frederic to use 2 texture units and a color blend function rather than 3 texture units.

20 years agoFrederic Bouvier:
ehofman [Thu, 20 May 2004 14:18:15 +0000 (14:18 +0000)]
Frederic Bouvier:

Melchior spotted a problem where we can crash an airplane into the
beacon's beam. The patch below enable to mask out a branch from HOT
traversal, whatever the animation.

The beacon.xml file is also included. It has a
<enable-hot type="bool">false</enable-hot> in a halo branch

20 years agoMake sure there will be no previous declaration errors.
ehofman [Thu, 20 May 2004 14:02:40 +0000 (14:02 +0000)]
Make sure there will be no previous declaration errors.

20 years agoPatch from Frederic. Adds support for bump-mapped (multi textured) 2d clouds, include...
ehofman [Thu, 20 May 2004 13:24:48 +0000 (13:24 +0000)]
Patch from Frederic. Adds support for bump-mapped (multi textured) 2d clouds, includeing support code.

20 years agoFred: include more check against null pointers and a raise in log level for missing...
ehofman [Sat, 15 May 2004 12:45:51 +0000 (12:45 +0000)]
Fred: include more check against null pointers and a raise in log level for missing objects.

20 years agoSolve the endless loop problem for the DC-3 and prevent a potential segmentation...
ehofman [Fri, 14 May 2004 19:46:12 +0000 (19:46 +0000)]
Solve the endless loop problem for the DC-3 and prevent a potential segmentation fault.

20 years agoAdd a function to globally set volume (aka AL_GAIN)
curt [Fri, 14 May 2004 15:47:01 +0000 (15:47 +0000)]
Add a function to globally set volume (aka AL_GAIN)

20 years agoOops, fix a typo.
curt [Tue, 11 May 2004 22:39:21 +0000 (22:39 +0000)]
Oops, fix a typo.

20 years agoFrederic Bouvier:
curt [Tue, 11 May 2004 22:21:24 +0000 (22:21 +0000)]
Frederic Bouvier:

I was not very happy with the size of the halo, so
I created a new animation to control it. Now we can
control the scale value with the distance from the
viewer to the object. The towers are now beginning to
look good. They might need some tuning though. If
you want to play, locate in radio-*.xml this code :

 <animation>
  <type>dist-scale</type>
  <object-name>RedLight.2</object-name>
  <interpolation>
   <entry><ind>0</ind><dep>0.1</dep></entry>
   <entry><ind>500</ind><dep>0.3</dep></entry>
   <entry><ind>16000</ind><dep>3</dep></entry>
  </interpolation>
 </animation>

You get the idea ? ind is the distance, dep is the
resulting scale value.
The medium tower appears brighter than the tall one,
because the lights are closer to each other. Maybe
they need a smaller scale factor at distance. Feel
free to modify these values if you find a better
setup.

About the code : I renamed flash to custtrans because the
ssg branch is now less specialized. It needs a callback
to compute the so called 'custom transformation'. It can
be used for the SGFlashAnimation and the new
SGDistScaleAnimation. So please cvs remove flash.[ch]xx and
add custtrans.[ch]xx. I also undo some of the code I send
you yesterday that was totally useless. It is replaced by
something simpler and that works.

There is also a patch to matmodel.cxx. This is not related
and was something I forgot. Its purpose is to set the
alpha test on material billboard models that are likely to
be trees to lessen a transparency weird effect with clouds.

20 years agoI had overlooked a few memory allocation/deallocation issues for audio buffers.
curt [Mon, 10 May 2004 21:22:50 +0000 (21:22 +0000)]
I had overlooked a few memory allocation/deallocation issues for audio buffers.
Hopefully this helps clean those up.

20 years agoFrederic Bouvier:
curt [Mon, 10 May 2004 20:27:30 +0000 (20:27 +0000)]
Frederic Bouvier:

I modified the included animation.cxx to have a randomly displaced
time origin, to break the unison. And the flashing period is also
random as you noticed. I also put all the flashing light of the pole
in the same animation so they flash in the same rhythm.

20 years agoFrederic Bouvier:
curt [Mon, 10 May 2004 14:59:02 +0000 (14:59 +0000)]
Frederic Bouvier:

Fix a memory leak, and brownian animation, if not motion.

I have 2 new files : personality.[ch]xx . They store the personality
data that would be deleted when the object is destroyed, instead
of staying in the animation maps. I also manage the current animation
step better and the towers are not flashing randomly now.
Makefile.am is updated.

20 years agoFrederic Bouvier:
curt [Mon, 10 May 2004 14:35:58 +0000 (14:35 +0000)]
Frederic Bouvier:

modellib.cxx :
  Add a branch between the model and its transformation to add
 a unique identifier for the model. I called it "personality
 branch" and it maintains a "current_object" in SGAnimation.
 Animations specifically written to support it ( currently only
 the timed animation ) can add a degree of variety among the
 occurrences of the same model that would look otherwise cloned.

flash.[ch]xx :
  Better compute the view vector. The flash is now syncronized with
 its axis even at the edge of the screen.

animation.[ch]xx :
  Lots of changes :
 - add a condition to 'rotate', 'spin', 'translate' and 'range'.
  When a condition is specified *and* it evaluates to false, the
  animation becomes a no-op. Possible usage : no rotation during
  daylight, LOD range variable along the day, ...

 - use different durations for each branch in the timed animation.
  Enable the use of multiple <branch-duration-sec>, one for each
  <object-name> specified. Usage : strobes with flash light.

 - allow randomization of the <branch-duration-sec>, by using
  <random><min>_min_value_</min><max>_max_value_</max></random>.
  The value computed once is between _min_value_ and _max_value_.

 - implement model personality in timed animation. If
  <use-personality type="bool">true</use-personality> is specified,
  a different set of duration is created for every model in the
  scenegraph using this animation. Best if used with randomization.
  When using strobes where the population of the same object is
  dense, it avoids the "cheasy" clone effect.

20 years agoAdd fast functions for exp2, pow, log2, root, sin/cos/tan, asin/acos/atan along with...
ehofman [Sat, 8 May 2004 12:58:29 +0000 (12:58 +0000)]
Add fast functions for exp2, pow, log2, root, sin/cos/tan, asin/acos/atan along with abs, neg and sgn.

20 years agoFrederic Bouvier:
ehofman [Fri, 7 May 2004 16:42:59 +0000 (16:42 +0000)]
Frederic Bouvier:

this patch introduce a new kind of animation and ssg branch.
I called them flash animation, because they help me to
enhance the look of the rotating beacon and possible future
lighthouse. It computes the cosine of the angle between an
arbitrary axis, transformed by the current modelview matrix,
and the view direction. No trig involved, just a dot/scalar
product.

The computed value can be modified by three parameters,
power, factor and offset, according to the formulae :

 value = factor * pow( cosine, power ) + offset.

It is clamped between a minimum and a maximum.
This value is then used as the scale factor of a matrix
transformation applied to the children of the SGFlash
branch.

The xml syntax, with default values, is :

<animation>
 <type>flash</type>
 <object-name>HaloObject</object-name>
 <center>
  <x-m>0</x-m>
  <y-m>0</y-m>
  <z-m>0</z-m>
 </center>
 <axis>
  <x>0</x>
  <y>0</y>
  <z>1</z>
 </axis>
 <power>1</power>
 <factor>1</factor>
 <offset>0</offset>
 <min>0</min>
 <max>1</max>
 <two-sides>false</two-sides>
</animation>

20 years agoRobustify the SGTimerQueue destructor. There have been reports of
andy [Mon, 3 May 2004 18:43:53 +0000 (18:43 +0000)]
Robustify the SGTimerQueue destructor.  There have been reports of
crashes on deletion.

20 years agoAdd a destructor for SGEventMgr. We don't own the pointer, so we
andy [Mon, 3 May 2004 18:39:25 +0000 (18:39 +0000)]
Add a destructor for SGEventMgr.  We don't own the pointer, so we
can't free it.  Just zero it out.

20 years agoChanges to get SimGear to configure and compile out-of-the-box on
andy [Fri, 30 Apr 2004 00:44:04 +0000 (00:44 +0000)]
Changes to get SimGear to configure and compile out-of-the-box on
a MinGW target:

Link against alut.dll in addition to openal32.dll.

Remove some preprocessor defines from compiler.h that were
confusing the mingw and/or libstdc++ headers (I put the _isnan
one back in the only file it was used).

Hack a broken sleep() call into the OpenAL sample programs so
that they will compile (but not work) in a non-POSIX environment.

Change the header file ordering in sample_openal.hxx to get
around some really weird interactions between MinGW's windows.h
and the gcc iostream header.

20 years agoHopefully fix a chicken/egg linking problem for people who've never built
curt [Thu, 29 Apr 2004 21:14:44 +0000 (21:14 +0000)]
Hopefully fix a chicken/egg linking problem for people who've never built
or installed simgear before.

20 years agoAdd support for specifying a positional offset relative to the listener.
curt [Wed, 28 Apr 2004 20:37:49 +0000 (20:37 +0000)]
Add support for specifying a positional offset relative to the listener.
This allows us to "place" cockpit sounds.  For example, we can make left
engine sound come out of the left speaker, right engine out the right
speaker, etc.

20 years agoAdd default openal libs for cygwin.
curt [Wed, 28 Apr 2004 19:19:05 +0000 (19:19 +0000)]
Add default openal libs for cygwin.

20 years agoLower verbosity level.
curt [Wed, 28 Apr 2004 03:59:10 +0000 (03:59 +0000)]
Lower verbosity level.

20 years agoExpose the ability to specify how the sound volume fades relative to
curt [Wed, 28 Apr 2004 03:57:13 +0000 (03:57 +0000)]
Expose the ability to specify how the sound volume fades relative to
distance from the listener.  This let's us configure "interior" cockpit
sounds versus "exterior" engine type sounds.

20 years agoTweak default source audio parameters.
curt [Tue, 27 Apr 2004 21:11:21 +0000 (21:11 +0000)]
Tweak default source audio parameters.

20 years agoExpose some of the positional components of the OpenAL API.
curt [Tue, 27 Apr 2004 20:45:58 +0000 (20:45 +0000)]
Expose some of the positional components of the OpenAL API.

20 years agoMake sure the prototype definition of sleep() is found (at least for IRIX).
ehofman [Tue, 27 Apr 2004 08:59:31 +0000 (08:59 +0000)]
Make sure the prototype definition of sleep() is found (at least for IRIX).

20 years agoOops, one addtional tweak.
curt [Mon, 26 Apr 2004 22:07:19 +0000 (22:07 +0000)]
Oops, one addtional tweak.

20 years agoUpdate the SoundSample api so we can request that a copy of the sample be
curt [Mon, 26 Apr 2004 22:02:14 +0000 (22:02 +0000)]
Update the SoundSample api so we can request that a copy of the sample be
kept in memory and accessible.

20 years agoGive these two methods slightly less misleading names.
curt [Mon, 26 Apr 2004 16:17:21 +0000 (16:17 +0000)]
Give these two methods slightly less misleading names.

20 years agoDavid Luff:
curt [Mon, 26 Apr 2004 15:55:29 +0000 (15:55 +0000)]
David Luff:

Fix comments for two functions.

20 years agoMissed one fix for Mac OS.
curt [Mon, 26 Apr 2004 14:55:27 +0000 (14:55 +0000)]
Missed one fix for Mac OS.

20 years agoLower the verbosity in a couple other spots.
curt [Sun, 25 Apr 2004 02:30:52 +0000 (02:30 +0000)]
Lower the verbosity in a couple other spots.

20 years agoLower verbosity.
curt [Sun, 25 Apr 2004 02:02:35 +0000 (02:02 +0000)]
Lower verbosity.

20 years agoAdd support for the MacOS variations of OpenAL.
curt [Sun, 25 Apr 2004 01:48:08 +0000 (01:48 +0000)]
Add support for the MacOS variations of OpenAL.

20 years agoAdd correct openal libs for MacOS.
curt [Sun, 25 Apr 2004 01:41:24 +0000 (01:41 +0000)]
Add correct openal libs for MacOS.

20 years agoClamp pitch values rather than just dumping an error message.
curt [Sat, 24 Apr 2004 19:47:42 +0000 (19:47 +0000)]
Clamp pitch values rather than just dumping an error message.

20 years agoRewrite the entire audio support library on top of OpenAL rather than plib's
curt [Sat, 24 Apr 2004 19:02:29 +0000 (19:02 +0000)]
Rewrite the entire audio support library on top of OpenAL rather than plib's
sound manager.  The interface was simplified and cleaned up a bit, and I
haven't back ported these changes to the plib sound wrappers ... we could
I suppose if someone really had a problem, but I haven't seen anything so
far that would indicate the extra effort is worth it.

20 years agoBernie Bright:
curt [Thu, 22 Apr 2004 12:39:16 +0000 (12:39 +0000)]
Bernie Bright:

gcc 3.4 has changed the rules for unqualified template name lookup.  This
affects SGQueue.hxx.  The changes I've made are backwards compatible with
earlier gcc versions.  Everything else compiles pretty much okay except for a
few warnings.  The resultant executable seems a bit faster too.

20 years agoFrederic: The state selector was not referenced and got deleted as soon as the sky...
ehofman [Sun, 4 Apr 2004 15:35:28 +0000 (15:35 +0000)]
Frederic: The state selector was not referenced and got deleted as soon as the sky was rebuilt a second time with the metar code.

20 years agoAdded more generated files.
david [Sun, 4 Apr 2004 14:24:06 +0000 (14:24 +0000)]
Added more generated files.

20 years agoMake sure GLuint is known.
ehofman [Sun, 4 Apr 2004 13:46:26 +0000 (13:46 +0000)]
Make sure GLuint is known.

20 years agoFrederic Bouvier:
ehofman [Sun, 4 Apr 2004 13:41:53 +0000 (13:41 +0000)]
Frederic Bouvier:

This is a new patch that allow to define a different
texture for top and bottom of clouds. It uses the
overcast_top.rgb you made for me last time.

What the patch do is to install a ssgStateSelector
instead of a ssgSimpleState for each layer.
The SGCloudLayer::draw method is modified to accept
a boolean that will select the proper state: 0/false
for bottom, 1/true for top.

Then, in SGSky::drawUpperClouds and SGSky::drawLowerClouds,
SGCloudLayer::draw is called with false and true because
we see the bottom of upper clouds and the top of
lower clouds.

Only overcast has 2 textures, the other types share the
same state for top and bottom, but that could be modified
in SGCloudLayer::rebuild.

20 years agoPlib is willing callbacks to return 0, 1 or 2 and not simply a boolean
ehofman [Fri, 2 Apr 2004 19:48:50 +0000 (19:48 +0000)]
Plib is willing callbacks to return 0, 1 or 2 and not simply a boolean