This is a new improved patch for the previous tile manager fixes.
Rather than building dependencies between FGlocation or the viewer or fdm with
tilemgr what I ended up doing was linking the pieces together in the Mainloop
in main.cxx. You'll see what I mean...it's been commented fairly well. More
than likely we should move that chunk somewhere...just not sure where yet.
The changes seem clean now. As I get more ideas there could be some further
improvement in organizing the update in tilemgr. You'll note that I left an
override in there for the tilemgr::update() function to preserve earlier
functionality if someone needs it (e.g. usage independent of an fdm or
viewer), not to mention there are a few places in flightgear that call it
directly that have not been changed to the new interface (and may not need to be).
The code has been optimized to avoid duplicate traversals and seems to run
generally quite well. Note that there can be a short delay reloading tiles
that have been dropped from static views. We could call the tile scheduler on
a view switch, but it's not a big deal and at the moment I'd like to get this
in so people can try it and comment on it as it is.
Everything has been resycned with CVS tonight and I've included the
description submitted earlier (below).
Best,
Jim
Changes synced with CVS approx 20:30EDT 2002-05-09 (after this evenings updates).
Files:
http://www.spiderbark.com/fgfs/viewer-update-20020516.tar.gz
or
http://www.spiderbark.com/fgfs/viewer-update-20020516.diffs.gz
Description:
In a nutshell, these patches begin to take what was one value for ground
elevation and calculate ground elevation values seperately for the FDM and the
viewer (eye position). Several outstanding view related bugs have been fixed.
With the introduction of the new viewer code a lot of that Flight Gear code
broke related to use of a global variable called "scenery.cur_elev".
Therefore the ground_elevation and other associated items (like the current
tile bucket) is maintained per FDM instance and per View. Each of these has a
"point" or location that can be identified. See changes to FGLocation class
and main.cxx.
Most of the problems related to the new viewer in terms of sky, ground and
runway lights, and tower views are fixed.
There are four minor problems remaining. 1) The sun/moon spins when you pan
the "lookat" tower view only (view #3). 2) Under stress (esp. magic carpet
full speed with max visibility), there is a memory leak in the tile caching
that was not introduced with these changes. 3) I have not tested these
changes or made corrections to the ADA or External FDM interfaces. 4) The
change view function doesn't call the time/light update (not a problem unless
a tower is very far away).
Details:
FDM/flight.cxx, flight.hxx - FGInterface ties to FGAircraftModel so that it's
location data can be accessed for runway (ground elevation under aircraft)
elevation.
FDM/larsim.cxx, larcsim.hxx - gets runway elevation from FGInterface now.
Commented out function that is causing a namespace conflict, hasn't been
called with recent code anyway.
FDM/JSBSim/JSBSim.cxx, YASim/YASim.cxx - gets runway elevation from
FGInterface now.
Scenery/newcache.cxx, newcache.hxx - changed caching scheme to time based
(oldest tiles discard).
Scenery/tileentry.cxx, tileentry.hxx - added place to record time, changed
rendering to reference viewer altitude in order to fix a problem with ground
and runway lights.
Scenery/tilemgr.cxx, tilemgr.hxx - Modified update() to accept values for
multiple locations. Refresh function added in order to periodically make
the tiles current for a non-moving view (like a tower).
Main/fg_init.cxx - register event for making tiles current in a non-moving
view (like a tower).
Main/location.hxx - added support for current ground elevation data.
Main/main.cxx - added second tilemgr call for fdm, fixed places where viewer
position data was required for correct sky rendering.
Main/options.cxx - fixed segfault reported by Curtis when using --view-offset
command line parameter.
Main/viewer.cxx, viewer.hxx - removed fudging of view position. Fixed numerous
bugs that were causing eye and target values to get mixed up.
graph as the terrain, except for internal cockpit view. The SSG
scene-graph variables (except for the lighting root -- I'll get that
later) are now held in globals.hxx.
FGModelMgr::draw() is obsolete; I'll remove it in a future revision.
Description:
This update includes the new viewer interface as proposed by David M. and
a first pass at cleaning up the viewer/view manager code by Jim W.
Note that I have dropped Main/viewer_lookat.?xx and Main/viewer_rph.?xx and
modified the Makefile.am accordingly.
Detail of work:
Overall:
The code reads a little easier. There are still some unnecessary bits in
there and I'd like to supplement the comments in the viewer.hxx with a tiny
bit on each interface group and what the groupings mean (similar but briefer
than what you emailed me the other day). I tried not to mess up the style,
but there is an occasional inconsistency. In general I wouldn't call it done
(especially since there's no tower yet! :)), but I'd like to get this out
there so others can comment, and test.
In Viewer:
The interface as you suggested has been implemented. Basically everything
seems to work as it did visually. There is no difference that I can see in
performance, although some things might be a tiny bit faster.
I've merged the lookat and rph (pilot view) code into the recalc for the
viewer. There is still some redundancy between the two, but a lot has been
removed. In some cases I've taken some code that we'd likely want to inline
anyway and left it in there in duplicate. You'll see that the code for both
looks a little cleaner. I need to take a closer look at the rotations in
particular. I've cleaned up a little there, but I suspect more can be done
to streamline this.
The external declaration to the Quat_mat in mouse.cxx has been removed. IMHO
the quat doesn't serve any intrinsic purpose in mouse.cxx, but I'm not about
to rip it out. It would seem that there more conventional ways to get
spherical data that are just as fast. In any case all the viewer was pulling
from the quat matrix was the pitch value so I modified mouse.cxx to output to
our pitchOffset input and that works fine.
I've changed the native values to degrees from radians where appropriate.
This required a conversion from degrees to radians in a couple modules that
access the interface. Perhaps we should add interface calls that do the
conversion, e.g. a getHeadingOffset_rad() to go along with the
getHeadingOffset_deg().
On the view_offset (now headingOffset) thing there are two entry points
because of the ability to instantly switch views or to scroll to a new view
angle (by hitting the numeric keys for example). This leaves an anomaly in
the interface which should be resolved by adding "goal" settings to the
interface, e.g. a setGoalHeadingOffset_deg(), setGoalPitchOffset_deg(), etc.
Other than these two issues, the next step here will be to look at some
further optimizations, and to write support code for a tower view. That
should be fairly simple at this point. I was considering creating a
"simulated tower view" or "pedestrian view" that defaulted to a position off
to the right of whereever the plane is at the moment you switch to the tower
view. This could be a fall back when we don't have an actual tower location
at hand (as would be the case with rural airports).
ViewManager:
Basically all I did here was neaten things up by ripping out excess crap and
made it compatible as is with the new interface.
The result is that viewmanager is now ready to be developed. The two
preexisting views are still hardcoded into the view manager. The next step
would be to design configuration xml (eg /sim/view[x]/config/blahblah) that
could be used to set up as many views as we want. If we want to take the easy
way out, we might want to insist that view[0] be a pilot-view and have
viewmanager check for that.
What was happening was that we screwed up and scheduled tiles for
(lon,lon) rather than (lon,lat) ... note the typo. This generated
bogus tile id's which the system happily accepted, put into the tile
cache system, and attempted to load. The problem was that these bogus
tile id's were negative where as all valid tile id's should be >= 0.
These negative tile id's up the logic used to remove tiles from the
cache. When identifying tiles for removal, we look for the furthest
tile away from us by starting out the furthest id at -1 and if we find
something further, we update the furthest tile id. Then at the end we
check if the furthest tile id >= 0 to see if we found anything we
could remove. However, the furthest tile id was these bogus tiles
with negative tile id's so the system always assumed there was nothing
appropriate for removal. This made it impossible to ever remove a
tile from the cache meaning it quickly filled up and no more tiles
could be loaded.
I fixed the one instance of scheduling tiles for a bogus location, and
added a sanity check so if it ever happens again we'll bomb with an
appropriate error message.
but no entries qualify for removal. It will keep trying to schedule the
tile(s) until an entry frees up. Entries in the cache do not qualify for
removal if they are in the process of being loaded.
a crash when relocating to a new airport. Pending work from the old
area is now just completed as normal, rather than trying to empty the various
queues in their various stages when can lead to many problems in a threaded
environment.
to have attached an ssg loaded object to this branch, then plib will remove
it and all it's states (and textures) which will call opengl api commands
which will crash the program if run from a separate thread from the main render
thread.
- model loading deferred to primary thread
- tile removal deferred to paging thread
- other tweaks and rearrangments.
Airport signs
- first stab at some support for adding taxiway and runway signs. This
is non-optimal, but I'm under the gun for a demo.
He writes:
Here are the final changes to add threads to the tile loading. All the
thread related code is in the new FGTileLoader class.
./configure.in
./acconfig.h
Added --with-threads option and corresponding ENABLE_THREADS
definition. The default is no threads.
./src/Scenery/tilemgr
Removed load_queue and associated references. This has been replaced by
a new class FGTileLoader in FGNewCache.
Made the global variable global_tile_cache a member.
schedule_needed(): removed global_tile_cache.exists() tests since
sched_tile() effectively repeats the test.
initialize_queue(): removed code that loads tiles since this is now
performed by FGTileLoader.
update(): ditto
./src/Scenery/newcache
Added new class FGTileLoader to manage tile queuing and loading.
tile_map typedefs are private.
exists() is a const member function.
fill_in(): deleted
load_tile(): added.
./src/Scenery/FGTileLoader
The new threaded tile loader. Maintains a queue of tiles waiting to be
loaded and an array of one or more threads to load the tiles. Currently
only a single thread is created. The queue is guarded by a mutex to
synchronize access. A condition variable signals the thread when the
queue is non-empty.
CLO: I made a few tweaks to address a couple issues, hopefully what we
have is solid, but now we kick it out to the general public to see. :-)
FGOptions is history, and the modules are (starting) to use the property
manager directly. Let me know if I left any files out.
Inevitably, there will be some problems with broken options, etc.,
that I haven't found in my tests, but I'll try to fix them quickly.
We also need to stress that the property names currently in use are
not stable -- we need to reorganize them a bit for clarity.
structure. The new approach is simpler, more flexible, and more dynamics.
We can now dynamically size the tile cache up and down. Also, the range
of tiles to load is now dependent on visibility and is calculated to always
bring in enough tiles.
