Eclipse Finder locks-up

[DOJOMO]

I have noticed that eclipse finder when set to lunar locks-up when certain addons are included in the extras folder. The addons are all superb. The Hale telescope, DSS-24 "paper" radio telescope and the v large ISS model. I think the latter is a special case caused by me.
Very tentatively I think the reason is the use of the "component" definition in the ssc files. Is each "component" being seen as a moom? Possibly switching to "spacecraft" may eliminate the problem. I haven't tried this yet.
I had previously changed the ISS ssc's definitions from "spacecraft" to "component" to remove the "glow" effect. I am possibly utterly wrong about this but any solution would be gratefully received as these models are great favourites of my grandchildren.
This issue seems to happen in all versions of 1.6.0.
Regards, David.

[DOJOMO]

I have since tried out what happens when changing the ssc's definition of component to spacecraft. I suggested this in my first post. The answer is nothing i.e. the problem remains.

Regards David.

[DOJOMO]

Apologies for wittering on and having a conversation with myself but I think I have been able to see what the problem is. It is the use of the "invisible " Class for a body on or orbiting a planet. The invisible class seems to define the body as an eclipsing type as well.
This is interpreted by the eclipse finder and gives a wacky result or it locks-up.
I tested this using Selden's Paper radio teleecope. Changing the single reference of Class "invisible" to say "spacecraft" corrects the problem with the eclipse finder. Unfortunately the telescope is somewhat damaged.
Is this a bug?
Regards, David

[selden]

I think that Class "invisible" causing problems for the eclipse finder is a bug.

In order to minimize the damage to the Addon, you can try reducing the radius in the SurfaceObject definition:

Code: Select all

SurfaceObject "vla-00" "Sol/Earth"
{
#   Class "invisible"
   Radius   0.0001 # was 0.0125

    FixedPosition { Planetographic [ -107.6177275  34.0787492  2.124]}
   FixedRotation {}
}


[DOJOMO]

Thanks Selden. I will try your suggestion again as it didn't work with ssc file that I have. Perhaps I have an outdated version as the original radius is .04 and the object is defined as "dss-24" "Sol/Earth".
Amending your superb Hale telescope and the ISS will, for me prove a bit of a task.
I will post this as a bug. after I get your solution to succeed.

Regards, David.

[selden]

oops. Sorry. I did two radio telescopes recently and confused the two. Still, the principle is the same: make the radius of the reference object tiny so that it's hidden within the model's axis.

Which version of the Hale Telescope are you using? I think V1 should be straight forward, similar to the radio telescope. V2 would be difficult, and V3 impossible

[DOJOMO]

Selden, I can't yet get your solution to work for the "Paper Telescope" but I will keep trying and come back to you when I do. I have ver 3 of your Hale telescope installed.

Many thanks for your help, David.

[DOJOMO]

Selden, I have got your solution to work but it has thrown up another aspect. This is what I did:-
I substituted both "invisible" and "component" class definitions with a "spacecraft " class. As you suggested I decreased the radius to .0001. The telescope displays ok and the eclipse finder works correctly.
This seems to imply that both "invisible" and "component" classes exhibit the same bug effect on the eclipse finder.
I think I am way out of my depth here and no longer sure that I have found the true cause. Rather than waste your time on this I will report it as a bug. I will twiddle with it for a bit.

As always many many thanks, regards David.

[DOJOMO]

In the first post of this thread I included in the addons that cause the eclipse finder to fail linuxmans wonderful ISS model. Apologies, as this is not correct. The distributed version does not cause this to happen and everything works ok. The reason my version caused the failure was that I had altered all the "spacecraft" class to "component" class. This caused the failure.
I made the change to remove the glowing blobs that demark spacecraft in celestia and occur a number of times on the ISS model. I have now reverted to the distributed version.
Many apologies, David.

[Marco Klunder]

Well, I seem to have the same problem here.

I just released the new "Moon Phenomena v2.0" script for Celestia v1.6.1 and after the real Lunar Eclipse two days ago, my daughter asked me when the next one will occor. No problem, I thought as a Celestian
But starting-up Celestia and running the Lunar Eclipse Finder for Earth also caused this hang-up.
Then I was triggered by this thread.

You guys talk about "invisible" and "component" classes exhibit the bug effect on the eclipse finder.
Here is the SSC file I use for the compass cmod, used by the script:

Code: Select all

SurfaceObject "compass_MFK_EN" "Sol/Earth"
{
   Class "surfacefeature"
   Emissive true
   Visible true
   
   Mesh "compass_MFK_EN.cmod"
   Orientation [90 1 0 0]
   Radius 0.155

   FixedPosition {Planetographic [ 4.89358 52.38449 0.2075 ]}
   FixedRotation {}

}

In the above code, replacing "surfacefeature" by "spacecraft" seem to circumvent the problem with Eclips Finder and the script still runs fine...

Besides V1.6.1. I've also SVN running with the same problem on board.
Is the Eclips Finder problem as mentioned in this thread already reported as a (possible) Celestia bug to the development team?

I'll continue testing on this (temporary) solution for the "Moon Phenomena" script and when no further issues occur on this one, I'll soon release "Moon Phenomena v2.1" to circumvent the problem...

Marco

[Fenerit]

Hi! SurfaceObject has the class "surfacefeature" as default. If I completely remove the directive, with the Marco's add-on, the eclipse finder sometime works sometime not, instead of not working at all. Even for me changing the class as "spacecraft" works fine. Maybe this kind of class is most powerful, that is, it does wipe out and supersede "minor" classes?