Melchior FRANZ:
The jitter is most likely caused by the irregular frame rate and CPU clock dependent intervals. There's no easy way around that. I tried some fancy interpolation and all that -- to no avail.
Improvements to the replay subsystem.
- Handle rotational interpolation across the "zero" point.
- Bug fixes to the rotational interpolator
- Change intervals for medium term and long term data recorders.
I have added a fledgling replay system that records flight data and control
positions during the flight.
I have added an internal command called "replay" which will trigger a replay
of the entire saved flight data set. This could be bound to a keyboard or
menu command, in fact this entire module is screaming for someone to build
a gui to control playback speed, amount of playback, etc.
This is the initial version so there are kinks that still need to be worked
out, please be patient.
Jim Wilson:
These changes should preserve previous functionality (with the exception of a
couple bug fixes).
Bugs fixed:
- AP no longer resets the error accumulator when switching altitude modes or
just closing the autopilot GUI. It will not be necessary to collect the barf
bags after selecting a new altitude anymore. Makes things much smoother.
- climb_rate calculation in the altitude hold mode included a factor that made
sense for the c172. It is now scaled according to the configuration's target
climb rate.
Additions:
Autothrottle (supports speed control only) is more configurable and accurate.
VerticalSpeed mode added (automatically arms to altitude if flown toward
altitude setting).
Exposed various properties, added new lock properties.
Melchior FRANZ:
The current chase view respects heading but ignores roll & pitch. And it follows heading without delay, which makes the viewer behave quite strange. This change makes the chase view feel more natural. You aren't fixed behind the plane, but follow all its movements with a delay.
Erik Hofman:
I've decided not to add the patch to preferences.xml in the base package because something feels funny with that. I think there needs to be some more discussion about it.
Add normalized control surface positions to net_fdm.hxx structure.
Convert several double values to float since the extra precision is not
needed and it just wastes network bandwidth.
Replace some of the most used fgGetXXXX() calls with a local variable. The next best thing would be replacing the fgGetXXXX()functions with cached pointers
david [Fri, 4 Jul 2003 19:13:03 +0000 (19:13 +0000)]
Rename some existing variables/methods in the navcomm module for
clarity:
nav_radial => nav_target_radial (same as selected, except for a LOC)
nav_heading => nav_reciprocal_radial
nav_magvar => nav_twist (it's not always the same as magvar)
nav_heading_needle_deflection => nav_cdi_deflection
nav_gs_needle_deflection => nav_gs_deflection
Added nav_radial back in, but now it shows the current radial from the
VOR, as one would expect. This value also appears in the
/radios/nav[*]/radials/actual-deg property.
Jim Wilson:
This patch exposes the nav_id--Navaid (VOR/ILS) IDs--in the property tree for use in EFIS displays. Both the string and individual integer (char) values are published.
Erik Hofman:
I have converted all sprintf() functions in navcom.cxx into snprintf() for some extra securety.
Jim Wilson:
This patch adds an "interval-sec" property which allows fixing an interval in seconds (or fraction of seconds) for the repeats for emulated axis controls (digital hats) on joysticks.
ehofman [Fri, 27 Jun 2003 08:46:57 +0000 (08:46 +0000)]
Melchior FRANZ:
These patches add a clock instrument, which allows to model failure ("serviceable") and to adjust the time independently of the system time (defaults to GMT). The main incentive is to make the p51d clock work and adjustable via the knob.
o Offers a time string ("12:03:15") for the LCD or for LED
clocks, or an empty string in case of failure/power off. The
instrument assumes that digital clocks are battery buffered,
so they will be updated even if there's nothing on the display.
o Offers the number of seconds since midnight for analog
clocks, like in the p51d. This number is not increased
if !serviceable. So the clock will stand still and continue
where it stopped when it's serviceable again.
I did not consider voltage yet, because the Mustang's clock will need a lot more current than the LCD clock. The instrument is updated 4 times per second but returns immediately if neither time nor offset changed. The function getGMTString() in fg_props.cxx could be removed after applying these patches.
curt [Wed, 25 Jun 2003 18:10:10 +0000 (18:10 +0000)]
Many ILS approaches do not have a listed GS elevation in the FAA or in the
DAFIFT data base. Use the field elevation as a reasonable approximation
when no other data is available.
curt [Tue, 24 Jun 2003 20:04:29 +0000 (20:04 +0000)]
Previously the weather dialog box and environment system had some confusion
between temperature at altitude vs. temperature at sea level. The dialog
box asked for temperature at altitude which makes sense, but all the
internal crunching expected temperature at sea level. However, it makes no
logical sense to specify the sea level temperature for different layers so
I changed the internal processing to work with temperature at altitude and
then derive an approximate sea level temperature at the end.
If you know the ground temperature, you can just enter this temperature
for the first boundary layer and the system should do the right thing.
daveluff [Wed, 11 Jun 2003 21:49:46 +0000 (21:49 +0000)]
Initial stab at making the AI plane take the user into account when flying a circuit. The AI plane will now delay it's turns to the various legs if the user has flown a wider pattern that it normally would in front of it. However, this is very much still work in progress - when the pattern gets extended the AI plane lands short of the runway, and it doesn't alter its speed around the circuit at all, only the turn positions. Still, its a start...
projects / flightgear.git / log