Celestia 1.7.0 Development Thread

[onetwothree]

A new snapshot 1.7.0~git20190409+ge422fe is available at https://dl.bintray.com/celestia/celestia-builds/.

Changes made since the previous snapshot:
* Names with Greek letters are stored in a processed format, e.g. α instead of ALF/ALPHA, while it seems that all possible bugs caused by this change are fixed, we won't be surprised if somebody finds a new one ;
* In Qt5 version it's possible to copy from and paste into a search console;
* Bugs in celestia:rotation() and celestia:setorientation() CELX methods causing Celestia crash are fixed;
* «Closed» exoplanet α Cen Bb is removed from an exoplanet database;
* A lot of tools were ported to currently used Eigen3 and Qt5 libraries, this snapshot is first one to contain the following tools: 3dstocmod, cmodfix, cmodsphere, cmodview, qttxf, txt2cmod, vsoptrunc-rect, vsoptrunc-sph;
* Dead code removal.

That's all so far

[Lafuente_Astronomy]

Alright, going to try this one out

[Lafuente_Astronomy]

Good afternoon Celestians and Devs

A member of the Discord asked this. The question is within the red box I placed in this image:

Thanks in advance

[john71]

New snapshot 1.7.0~git20190409+ge422fe bugs:

Earth has only nightlight textures from every angle when ambient light is off (zero).

With ambient light there is no dark side, the whole planet is illuminated equally from every angle.

The Moon is still invisible.

So the usual - very annoying - problems persist.

Is it me or it is a common Windows 10 64 bit problem?

[pirogronian]

@john71
I didn't noticed it yet, however I use Linux 64. I'm not sure if I tested yet pure build version or just one of my local branches...

[john71]

1.) The invisible Moon problem is a persistent bug in Windows 10, under any circumstances.

2.) The Earth illumination problem cannot be detected with the "base" low definition Celestia textures. But if you use HD Earth virtual textures (normal, surface, night, cloud maps), it resurfaces.

Added after 3 minutes 13 seconds:

no moon.png

Added after 8 minutes 22 seconds:

earth illumin.png

[pirogronian]

Hmm, just tested. No problems with any textures resolution. However, I'm not technically skilled with textures, just set via qt settings interface.

[Lafuente_Astronomy]

1.) The invisible Moon problem is a persistent bug in Windows 10, under any circumstances.

Concerning the Invisible Moon, go check if an asteroid and a dwarf planet have the same invisibility(For example, go to Ceres, then to Pluto). If they have, just uncheck "Moons", "Asteroids" and "Dwarf Planets" in the "View Options". Only then will they be visible

That is actually still a problem as the negative value leads to a positive outcome, which is not the way it should be.

If none of that works, then it's most likely your Windows program

[john71]

Thanks, Lafuente_Astronomy!

In the "qt" version, it works now! I just had to CHECK the "Moons" box! Strange.

The "win" version exe file does not work well in this regard, there are no moons in it.

By the way, why are there 2 exe files?

+ new bug: I have big cities on the surface of moons. They simply disappeared.

[onetwothree]

By the way, why are there 2 exe files?

Qt version is our main cross-platform version, it's the most featureful. But It actually requires OpenGL 3 for Qt itself. Native win version drags behind as we don't have any windows developers in our team and sometimes it has its own bugs like the one you have. But it works on older computers with GL2.1 only.

+ new bug: I have big cities on the surface of moons. They simply disappeared.

Please provide a test data and more information to reproduce the bug.

Added after 4 hours 5 minutes:

Concerning the Invisible Moon, go check if an asteroid and a dwarf planet have the same invisibility(For example, go to Ceres, then to Pluto). If they have, just uncheck "Moons", "Asteroids" and "Dwarf Planets" in the "View Options". Only then will they be visible

I'm unable to reproduce this under wine.

I suspect that the problem doesn't exist in clean installations. Could you guys clean your windows registry (Software\Shatters.net\Celestia) and check the result?

[john71]

Onetwothree, yes, of course, I will try to do that!

Right now I'm looking for more bugs.

+ 2. bug: a Dyson-sphere 3d model with 1 AU radius totally disappeared, it is invisible (in the 'qt' version).

[cartrite]

Concerning the Invisible Moon, go check if an asteroid and a dwarf planet have the same invisibility(For example, go to Ceres, then to Pluto). If they have, just uncheck "Moons", "Asteroids" and "Dwarf Planets" in the "View Options". Only then will they be visible

That is actually still a problem as the negative value leads to a positive outcome, which is not the way it should be.

If none of that works, then it's most likely your Windows program

c465ab26 was my last build of the win version on VMWare win7. The check boxes seem to be fixed in that version. Unchecking dwarf planets makes Ceres invisible as it should. Checking it makes it visible. Ceres is now considered a dwarf planet in this version.

One thing I noticed a while back, when I was building old versions trying to figure out the normal map problem, I downloaded lua tools along with the Realistic MilkyWay addon for it. The ssc file had a line

Code: Select all

Class "diffuse"

and I needed to comment it out to get the model visible. I'm not sure when this needed to be done or what commit lost support for the diffuse class. But commenting it out makes the model visible. Maybe this is causing invisible addons.

[john71]

I hope the diffuse function is still working, it is essential for building large scale diffuse structures around other models.

[Lafuente_Astronomy]

Good evening from the Philippines Celestians!

According to some astronomical information based on the Transit of Venus in 2004, there is going to be a very rare event in which Venus and Mercury transit the Sun simultaneously. Those would happen on 2 different dates: On July 26, 69,163 and on March 29, 224,508. At least, that's what this article states, which was among the astronomical information regarding the transit: http://www.solexorb.it/SolexOld/Simtrans.pdf

So, out of curiosity, and wanting to see such a simultaneous transit that would happen a very long time after we're all dead, I used Celestia to check if it would happen. For July 26, 69,163, this is what I got:

Since Venus wasn't shown in the first screenshot, this is where it was located:

For March 29, 224,508, this is what it looked like:

In either of those dates, a simultaneous transit of Mercury and Venus of the Solar Disk wasn't shown.

To make sure, I did the same thing in another planetarium software: Stellarium. This is what the first date looked like in Stellarium. I was only able to get the first date of the "double transit" of the Sun because Stellarium has a limit of years up to 100000 AD, meaning that I could never go to the second one.:

Notice that Celestia's position of the planets in that specific day is different from what was predicted in the article, and it was also greatly different from what Stellarium showed.

So perhaps the question is: Which has the most accurate calculations of the orbits and movements of the planets? And if so, are there ways to make them all the more accurate? Either way, I do believe that many things in Celestia can be improved or added like Proper Motion for both Stars and DSO's, Precession and more accurate calculations for the orbits of stars, galaxies and other celestial objects as well.

Thanks in advance

[onetwothree]

Celestia supports VSOP87 only, see https://en.wikipedia.org/wiki/VSOP_(planets)

[Lafuente_Astronomy]

Well, it looks like there are more accurate versions of VSOP that could be used, provided they are published. And perhaps it could be possible for the devs to update Celestia to support more than just VSOP87, because despite VSOP87 being more highly accurate than most other models, there are other numerical calculations that come into play as well, and while not noticeable, or barely present in modern-day estimates, it may over time become more pronounced and obvious that the theoretical models are becoming not accurate. Since the timetables of the possible double transits occur many thousands of years from now, it may be that the inaccuracies become obvious in those periods of time, not to mention that there's no proper precession in Celestia as well.

But it's entirely up to you. Accuracy is one thing, and I like it if stuff in space is accurate. But the other features would be important as well, and if you think it's more important to develop those instead, then I'll accept that.

So for now, I'll just wait patiently until the next update of Celestia 1.7.0. More power to you guys!

Added after 3 minutes 37 seconds:
Also, aren't there also the SPICE kernels for the planets or are they only limited to non-planetary solar system objects?

[Janus]

One of the things that that is commonly missed in calculating positions of planets, is it little different than those of stars.
To see how accurate Celestia can be, you need to calculate how far Venus & mercury will go in that time.

At 15.8Ly of travel in a circle around the sun, chosen for ease of illustration.
The maximum resolution of of celestia's internal calculations will be 1AU.
It doesn't take very much of an AU to be way off in an orbit, especially that close in.

I leave the actual math to you, but that is how much drift there is in SP positional math.
Does anyone know if VSOP in any version has more digits?


Janus.

[john71]

There is definitely a problem with surface features in the new build.

I made a small ssc file with a cmod cone to prove it.

Extract the attached file in the extras folder.

Go to Earth, hit enter and type "one1".

Celestia will find the object, but it is invisible.

Added after 24 seconds:

problem.7z

[Lafuente_Astronomy]

I think that problem's encountered before. Is it the one in which moons, dwarf planets and asteroids are invisible even if in the "View Options" window, their boxes are rendered as checked and thus, visible?

[john71]

Yes, I wrote about it a few weeks ago. This is the same problem.