durk [Sun, 17 Jan 2010 19:41:47 +0000 (19:41 +0000)]
John Denker:
Subject: [PATCH] Make terrasync more usable, more like a proper daemon.
-- writes pid to file
-- normally does not exit until signalled
-- uses blocking and non-blocking I/O as appropriate.
jmt [Sun, 10 Jan 2010 23:35:53 +0000 (23:35 +0000)]
GPS: improve reliability (at the expense of some in-development features) for pending release.
- default to an invalid altitude in routes, instead of cruise altitude (temporarily)
- only set an altitude on the autopilot, if valid
- only add departure airport/runway to the route, if not airborne
ehofman [Tue, 5 Jan 2010 11:24:15 +0000 (11:24 +0000)]
Tatsuhiro Nishioka:
These patches fixes minor bug in addition to providing --with-openal-framework and --with-cocoa-framework.
Now you can use your own version of OpenAL.framework for selecting various audio output device.
Plus, you can build FG/SG on Snow Leopard with cocoa configuration.
Note that Python's XML parser is picky about accents. Choose encoding
ISO-8859-1 for latin1 accents or UTF-8 for UTF-8 accents. (Don't use latin1
comments in UTF-8 encoding!) Use xmllint to check for the validity of your
XML files (or let ./scripts/tools/fg-check do that for you):
$ xmllint a26-yasim.xml
a26-yasim.xml:113: parser error : Input is not proper UTF-8, indicate encoding !
Bytes: 0xE8 0x73 0x20 0x6C
moment d'apr?s la doc : ________________________ = __________ =
^
Csaba Halasz [Sun, 20 Dec 2009 15:17:14 +0000 (16:17 +0100)]
Use the _controlled member instead of the tower pointer itself to check for valid tower in AILocalTraffic as tower doesn't always reflect current state.
Tim Moore [Thu, 12 Nov 2009 09:58:42 +0000 (10:58 +0100)]
Attach an empty shader program to the scene graph root
This establishes a default state of using the OpenGL fixed function pipeline.
Otherwise a shader program from a state set can leak into state sets that don't
use shaders at all.
James Turner [Thu, 17 Dec 2009 12:36:51 +0000 (12:36 +0000)]
Rewrite the spatial index to use a sparse octree on the cartesian coordinates of items.
* Fixes errors selecting 'nearest' items, including duplicates
* Stable behaviour around the poles and dateline
* Decently efficient, even for 'everything within 500nm' queries
ehofman [Tue, 8 Dec 2009 09:48:00 +0000 (09:48 +0000)]
John Denker:
Suggestion: It might be helpful to promote each of
the .c files to .cxx.
Rationale:
-- The configure/makefile system handles CFLAGS
differently from CXXFLAGS.
-- It is important for the *info programs to
compiled and run in exactly the same environment
as the main fgfs program. Some users depend on
compiler or linker flags such as rpath that
strongly affect the results of the *info programs.
-- There is no downside; you code compiles just
fine as-is under the c++ compiler.
second part of the original commit, for ehofman/sound
ehofman [Tue, 8 Dec 2009 09:48:00 +0000 (09:48 +0000)]
John Denker:
Suggestion: It might be helpful to promote each of
the .c files to .cxx.
Rationale:
-- The configure/makefile system handles CFLAGS
differently from CXXFLAGS.
-- It is important for the *info programs to
compiled and run in exactly the same environment
as the main fgfs program. Some users depend on
compiler or linker flags such as rpath that
strongly affect the results of the *info programs.
-- There is no downside; you code compiles just
fine as-is under the c++ compiler.
timoore split the original patch so that alcinfo.cxx can be commited to
the ehofman/sound branch seperately.
ehofman [Tue, 1 Dec 2009 09:08:08 +0000 (09:08 +0000)]
The problem with the current test is that HAVE_CULLSETTINGS_CLEAR_MASK is always defined, either with a value of 1 or a value of 0 while the test in the code checks for '#if defined(HAVE_CULLSETTINGS_CLEAR_MASK)'