CHARM2 measured stellar radii (DO NOT USE WITH NEW STARDB)

[ajtribick]

ajtribick,

I tried your addon. The stars don't use any texture at all. There should be something in the STC file so they can use the default textures, as declared in the new config file for 1.5.0. Currently, they are all dull glowing spheres

There is no texture override in the file, so there's no reason why the default texture shouldn't be used... however as far as I can tell you are incorrect to say the stars use no texture - instead the default A star texture (astar.jpg) is used.

A quick check suggests this is a problem common to ALL stars defined in .stc files (including the defaults supplied with Celestia) without a Texture definition, not just my add-on... looks like the .stc parser has not been updated to take these preferences into account. I'll file a bug on the tracker.

[chris]

ajtribick,

I tried your addon. The stars don't use any texture at all. There should be something in the STC file so they can use the default textures, as declared in the new config file for 1.5.0. Currently, they are all dull glowing spheres

There is no texture override in the file, so there's no reason why the default texture shouldn't be used... however as far as I can tell you are incorrect to say the stars use no texture - instead the default A star texture (astar.jpg) is used.

A quick check suggests this is a problem common to ALL stars defined in .stc files (including the defaults supplied with Celestia) without a Texture definition, not just my add-on... looks like the .stc parser has not been updated to take these preferences into account. I'll file a bug on the tracker.

I think there's a bug lurking here somewhere, but the stars in the CHARM2 add-on look just fine for me: the default textures as specified in celestia.cfg are being used.

--Chris

[ajtribick]

Can you give an example of a working CHARM2 star? The ones I've tried (Betelgeuse, Lalande 21185, Epsilon Eridani) all seem to be using astar.jpg...

There seems to be something weird going on.

It seems certain properties in .stc files wreck the texture matching.

EllipticalOrbit seems to be one.
I'm going to guess that CustomOrbit is another, given that the Sun seems to be using astar.jpg.
On the basis that most of the CHARM2 stars are using astar.jpg, I'd guess Radius is one.

In fact, the only stars that I've seen respecting the textures in .stc files (mainly I'm testing on M stars because it's most obvious) are the ones that use only the "vanilla" properties that can be defined in stars.dat.

[Cham]

After some tests, I think there's clearly a new bug here. All my standard stars do have textures, but the stars from the CHARM2 STC file don't.

[Cham]

I think I've found a source of the bug : RotationPeriod.

I'm having a star defined in some STC file, without a rotation period. It's texture is showing. Adding a simple rotation period, and the same star lose its texture !

[chris]

Never mind . . . it turns out that I wasn't actually using the CHARM2 add-on when I did my check. I was able to reproduce the bug and find it's source: a small typo in star.cpp:

Code: Select all

Index: star.cpp
===================================================================
--- star.cpp   (revision 4057)
+++ star.cpp   (working copy)
@@ -761,7 +761,7 @@
     bolometricCorrection(sd.bolometricCorrection),
     knowledge(sd.knowledge),
     visible(sd.visible),
-    texture(texture),
+    texture(sd.texture),
     model(sd.model),
     orbit(sd.orbit),
     orbitalRadius(sd.orbitalRadius),


This is probably also the source of cartrite's Sun-looks-like-Mars bug.

--Chris

[Cham]

Okay, this is systematic and very easy to reproduce (on my system) :

This star do have a texture :

Code: Select all

1009544774 "TET1 Ori B:41 Ori B:BM Ori:HD 37021"
{
   RA 83.81695833
   Dec -5.38527778
   Distance 1399.97
   SpectralType "B0V"
   AppMag 7.96
}


This same star, with an RotationPeriod added, doesn't have a texture !

Code: Select all

1009544774 "TET1 Ori B:41 Ori B:BM Ori:HD 37021"
{
   RA 83.81695833
   Dec -5.38527778
   Distance 1399.97
   SpectralType "B0V"
   AppMag 7.96
   RotationPeriod 1
}

[ajtribick]

It isn't just rotation period. Of the following four stars, only Test 1 gets the correct texture...

Code: Select all

"Test 1"
{
   RA 0
   Dec 0
   SpectralType "M5V"
   Distance 1
   AbsMag 6
}

"Test 2" # add in a radius definition, things go wrong
{
   RA 0.01
   Dec 0
   SpectralType "M5V"
   Distance 1
   AbsMag 6
   Radius 696000
}

"Test 3" # apparently even just being used as a barycenter is bad news
{
   RA 0.03
   Dec 0
   SpectralType "M5V"
   Distance 1
   AbsMag 6
}

"Test 4" # orbitbarycenter stuff isn't good.
{
   OrbitBarycenter "Test3"
   SpectralType "M5V"
   AbsMag 6
   FixedPosition [0.5 0.5 0.5]
}

[chris]

Correct. Any extended star attribute will trigger the bug. "Extended" means anything except coordinates, spectral type, and magnitude.

--Chris

[chris]

I committed my fix to SVN. If you update, the bug should be gone.

--Chris

[Cham]

Updated the code from SVN, recompiled, and the bug is gone !

[ajtribick]

I'm starting to wonder just how many more Celestia bugs this add-on is going to drag out of the woodwork, the tally's currently somewhere around 3-4.

*sigh*

[ajtribick]

Just to let you know you shouldn't use this add-on with the new star database as it will restore all affected stars to their Hipparcos-1 parallaxes instead of using the new updated ones.

A new version is in development but will probably only be released after 1.6.0 is available.