If you've been following my git repositories on http://gitorious.org/fg/, you may have noticed that,
while the tip of the next branch is identical to cvs, the history is rather complicated,
with lots of merges. As people have been making commits to CVS, I've been assembling them
in branches and then merging them to the next branch. This allows me to merge these
branches individually to a more stable "master" branch when it seems like the changes have
settled down and have been tested in next. Here's the first report since I started
doing this on September 12.
First is a list of branches that have been merged to the next branch, but not to master.
They aren't in master because they aren't "mature" enough yet, IMHO. My criteria for
merging changes to master are:
* should be in the next branch for at least a week
* or fixes an obvious bug on the master branch
* and isn't causing a massive flame war (OK, discussion) on the mailing list.
The branches are named with the original author of the changes and some descriptive
blurb. As changes come in to CVS I've been giving them a once-over: if they seem to
fit with an existing branch, they go there; otherwise I start a new branch.
[Holding in 'next']
* curt/radial (2009-09-30)
Smooth out the time-to-intercept radial computation so it's a bit more useful.
* jmt/brakes (2009-09-22)
Auto-brake enhancements: support for JSBSim aircraft, Airbus-style combined MAX/RTO modes, correct disengage behaviour and more disengage conditions.
Update MSVC project files for the autobrake files
Don't crash with UFO FDM, etc, or other FDMs that don't define global props needed - just go into passive mode.
Don't look for /velocities/groundspeed-kts during initialisation.
A generic, configurable autobrake system. Not fully features yet, but works quite nicely in the 777, including RTO modes and auto-disengage on pilot input.
* jmt/positioned (2009-10-01)
Use boost::to_lower_copy instead of function from simgear
Positioned support for adding user waypoints.
Further extensions to FGPositioned to support ongoing GPS work.
Extend FGPositioned to allow mapping from a string to a type.
* jmt/xmlauto (2009-09-23)
Refactor the XMLAuto parsing code to no longer warn about every node handled by the base class (FGXMLAutoComponent). As a side effect, issue better warnings when an unrecognised node is encountered.
* luff/kln89 (2009-09-24)
Remove the GPSPage base class, that only KLN89Page was derived from.
* topic/atis (2009-09-19)
use boost::tie instead of make_ref
fix ATIS on 64bit machines
MSVC fixes for atis merge.
ATIS overhaul by John Denker, adapted to trunk by me.
I understand that this patch is stable, except for missing voice elements in the data file?
* vivian/trainz (2009-09-30)
Try to improve frame rate and reduce jitter
Add support for tunnel, y-offset and limit update rate
These changes have been merged into master since September 12:
[landed in 'master']
Logging: less verbose panel loading.
Logging: less verbose traffic-manager startup.
Logging: less verbose about navaids we didn't align.
Fix the has-gs flag when switching to a VOR (thanks Torsten), and expose GS validity via a new gs-in-range boolean prop, as suggested by John Denker.
navradio: fix crash spotted by Torsten, when nav range is unspecified.
Work around the penaltyForNav broken-ness, by ensuring GS behaves sanely even if non-matches LOC and GS transmitters are selected. Which happens. Yuck.
One more divide by zero fix from Torsten.
Make FGRunway track reciprocal runways.
Associate non-ILS localisers with their runway.
Different approach to handling the paired-ILS-on-opposite-runways situation. This uses additional information we now store (in FGRunway) to be more explicit about the behaviour (i.e only look at paired navids), and crucially, does not misbehave near either threshold. Instead it always switches about the runway midpoint (so back-course/missed-approaches are broken as before), leading to stable behaviour when touching down.
NavRadio DME in-range flag, and audio code improvements. By John Denker, adapted to current trunk by me.
NavRadio: By way of an experiment, make false courses/lobes disable-able via a property.
Header clean-ups in viewer/view-mgr.
Drat, this is really needed. Yuck, yuck, yuck.
Put back an unfortunate include, until ViewMgr can be fixed to not inline FGViewer destructor calls.
FGFX clean-ups, conversion of one more sgVec user.
Break viewmgr.hxx dependency on on viewer.hxx.
Remove audio update from main.cxx, back to FGFX (but keep the timing-dependent piece done artificially late)
Remove use of plig/sg from render-area-2d.
Remove <plib/sg> includes.
Replace one SGD_ constant with a local one, in LaRCsim.
Remove plib/fnt.h from panel header.
Allow airports system to function without an environment manager - useful in some tests I'm writing.
fgpanel is shouldn't have escaped (yet)
forgotten ones from Tat's patch to enable make builds on Macs
and the (hopefully) last one from Tat's Mac-can-run-make patch
unintentional introduction of a not yet ready util.
only complain about binary record length mismatch if mode is binary.
Fix the temperature computation.
Catch a possible NAN at insanely low but greater than zero speeds
Alex Buzin: Fix wrong calculation of temperature and dewpoint