CVS bugs (stars, atmosphere,...)

[Malenfant]

Star name searches are always global. Planet searches are local to whatever system you're closest to or have selected.

--Chris

Hrm. That explains that then... However, could the star names be made local too? At least, can it be set so that if you type in a star name that exists elsewhere, it assumes you're looking for the one that is in the system you're in?

[Cham]

Malenfant,

there isn't suppposed to be two stars with the same name. I suggest you just rename one of the stars.

[Malenfant]

Malenfant,

there isn't suppposed to be two stars with the same name. I suggest you just rename one of the stars.

I guess you're right, you're supposed to call it "(Star name) A" and "(Star name) B" aren't you...

OK. Never mind then

[Cham]

I've just found another bug with the CVS version. This one may be minor, but I find it irritating. There are some random white dots appearing on many models (some 3ds models, and many 3ds converted to CMOD too). Those models were looking perfect in Celestia 1.4.1. Now, I get those "flashing", or "blinking" random dots on many parts of the models, like if the mesh wasn't perfect (I'm sure the mesh is okay). The flashing effect is caused by the rotation of the models relative to the observer. Here are some examples :



On a picture like this one, the dots looks tiny and insignificant. But live in Celestia, while everything is moving, they are perfectly clear, especially since they are "blinking", despite their small size (one pixel). They ARE NOT stars visible through the mesh. The dots tends to follow the mesh, and are frequently distributed along the models edges. Also, the dots seems to be scale independant. I can see many dots while far away from the model, and as much dots if I get close to the model.

[selden]

Cham,

a. Does that model contain one mesh or several?

b. Is that an SSC object Mesh or is it a DSC Nebula Mesh?

Problem almost confirmed: I just realized what's probably causing it on my system for the SSC Mesh shown below.

My models consist of a lot of separate meshes.
Meshes do not (yet) cast shadows on other meshes in Celestia.

The "hot pixels" in my model are most likely caused by seeing an edge of a mesh that is illuminated. It'll take me a while to verify this, but here's a picture of what I see now:

[Cham]

Selden,

all my models that are showing a "white dot" problem are normal SSC objects (no nebulae, since nebulae objects are usuallly very simple and set with Emissivity. I didn't noticed any problem with them yet).

Most objects contains several pieces, but some models don't. And I verified if those white dots were related to the separate pieces. Some yes, some not. The white dots are mostly concentrated on "spikes", hard edges and contours of the models. So I don't think it's related to the number of pieces included in the model (of course, if there are several pieces, then there are more "edges" and "spikes" on which the white dots could be concentrated).

I experienced some white dots on poorly done models in Celestia 1.4.1 before, but in those cases, it was evident that it was related to the "crudiness" of the model, which contained many holes, badly oriented normals, etc...

But now the problem appears on many good models which didn't had any problem in Celestia 1.4.1 before. This problem is "new" and appeared with the CVS version.

At first, I didn't noticed it, or didn't bothered much about that problem. The white dots are really tiny (but scale independant). But once I noticed it, it became "evident" and irritating. Now, I can notice the white dots all the time.

[Cham]

Here's another example of white dots on a mesh :



This is a large model. At very close range, the mesh doesn't show any defect (it's pretty smooth, actually). While looking at it from a distance, you can clearly see lots of small white dots on the edge (contour), and also on the surface. Moving around gives a bad feeling to the model.

[chris]

Here's another example of white dots on a mesh :

Is there somewhere I can download this model?

--Chris

[Cham]

EDIT :

Ok, here's a link to a small addon. A "funky" asteroid around Saturn (high resolution mesh).

Zip file is 2.7 MB (CMOD and a texture).

But I recall you that the problem occurs with many models of various types. Not just this one.

http://nho.ohn.free.fr/celestia/Cham/Te ... e-dots.zip

[cartrite]

Well I guess I got to say SORRY to Windows. I just did a CVS kde build for 9/26 and it broke the near mission. NEAR is MIA like in Windows. The same folder works with cvs 9/20 so something changed.

Code: Select all

Toolkit version: N0060

SPICE(SPKINSUFFDATA) --

Insufficient ephemeris data has been loaded to compute the position of -93
(NEAR) relative to 10 (SUN) at the ephemeris epoch 1979 DEC 01 00:00:00.000.

A traceback follows.  The name of the highest level module is first.
spkgps_c --> SPKGPS

I also seen this bright white sphere when starting the program and approaching earth. I think it's the moon. I took a screenhot but Image Shack is still having traffic problems. This was seen while the standard start.cel was running. Flew right through it. This is what I flew through



I still get a similar error message with the 9/20 version but it creates a log a mile long. The time is also current time not 1979. As soon as I set the time to 2001 then it stops. And everything works.

STRIKE THIS: I ereased my last edit. I was running the 9/20 build. Nothing works with 9/26 Near is MIA in both Win and Linux. Time for a break

[Sui Ota]

At specific viewpoint, specific fov, something like halation occurs:
Jupiter and its moons
Earth and Moon 1
Earth and Moon 2
Mars
with any star style.
This seems to be same thing as what cartrite mentions above.

System depended result:
On Windows XP SP2, Radeon 9200 SE, this occurs with any render path.
On Mac OS X 10.4.7, Mobility Radeon 9550, this occurs with Basic, Multitexture.

[Malenfant]

still got a big sun here... (with 1.5.0, 22/9/06) Looks kinda cool actually...

It only looks like this with Scaled Discs - as Points or Fuzzy Points the sun looks normal. While changing the FOV I notice that the size doesn't change at all at wider FOVs, but when I go smaller than default the solar disc suddenly gets a LOT smaller between 1.28x (20 degrees) and 1.34x (19 degrees) - but it's still larger than it should be there.

[bdm]

still got a big sun here... (with 1.5.0, 22/9/06) Looks kinda cool actually...

It only looks like this with Scaled Discs - as Points or Fuzzy Points the sun looks normal. While changing the FOV I notice that the size doesn't change at all at wider FOVs, but when I go smaller than default the solar disc suddenly gets a LOT smaller between 1.28x (20 degrees) and 1.34x (19 degrees) - but it's still larger than it should be there.

... and the last remnants of humanity desperately outrun the blast wave in their tiny spaceship as the sun goes supernova ...

[dirkpitt]

... and the last remnants of humanity desperately outrun the blast wave in their tiny spaceship as the sun goes supernova ...

Anyway, I think this bug has now been fixed.

[cartrite]

I don't think this is supposed to be happening. The full moon should be 180 degrees away? Screenshot taken with Oct/19/2006 cvs windows version.



A link to a celurl

Solarsystem barycenter was NOT being used to copy the url.

[cartrite]

I don't see this full moon in linux. Same url. No solar ststem baryenter.
The moon is still there with linux. It's just not lit up.
Ambeint Light was set to low
Turned it off and no full moon. Still, should the full moon be visable from the ground with ambeint light set to low.?