Celestia 1.7.0 Development Thread

[Lafuente_Astronomy]

The dist info normally shown is from the observer.
The dist I added is for how far from earth.
It uses the same routine as observer for adjusting units of distance, DistanceLyToStr.

Ahhh. So that's how it works.

For example, select alpha centauri, goto it, then zoom out.
The dist value shown in the upper left will change.
The dist value I just added however, will not.

Well, I encountered a problem that needs fixing but probably it'll be fast to you:

Upon clicking Alpha Centauri, the Dist showed the info as you said:

But upon coming closer, the "Dist" info disappears:

Well, there's always room for improvement, unless that was intended.

[Janus]

I had not retested deeply enough.
With recent changes to text handling, things have changed.
I was using a text search for earth to get an offset.
However, I had not realized that searches are now limited to the nearest solar system only.

There is no longer a universal search everything, which I intend to replace with a new name for my own use, even if no one else wants it.
The change is annoying since it alters a baseline behavior, and may explain some other problems I have been having with celestia.

It also changes how stars are displayed.
Star discs now change at solar system boundaries, at the same time as searches.
That is why star sizes twitch when you go to a star.

So here is a fix for that issue.
Let me know if this does better.



Janus.

[Lafuente_Astronomy]

There is no longer a universal search everything, which I intend to replace with a new name for my own use, even if no one else wants it.

I tested it. Everything's good so far. But what's specifically wrong with the Universal search that you feel you need to replace it?

[Janus]

Not replace, supplement.
What I ran into is that once you are away from SOL, once it is no longer the closest star, searching for "Earth" gives you a blank result.
For instance, go to wolf 359, look around, then hit enter, type in earth, see what matches.

I am trying to figure out how to make a search from wolf 359 be more informative.
For instance, searching for earth from there could show 'Sol:Earth' in the list.
Perhaps I could type in 'Sol:' to search for things in just our solar system, then 'Alpha Cen A:' to search only there.
Still not a universal search really, but a nice way of better presenting what is already present.

Celestia keeps several what amount to lists in memory.
Such as:Deep space objects, Stars, Planets, Moons, Minor moons, Asteroids, Comets, Spacecraft, etc.
Search used to search them all sequentially.
When I tried that after some of the recent changes, once SOL was no longer the closest star, null results, which is what made the Dist go away.

What I had to do to keep Dist is this:

Sol_Object = universe::find("Sol");
Earth_Object = universe::find("Earth",&Sol_Object,1);
if (selection.body() != EarthObject.body()){ Do Dist calculation and display here.}

The third line simply checks if the current selection is earth, and proceeds if it is not.

What I will be doing is making a searchall that returns an array of all matching objects.
To be supported in Celx of course.
Though this will not something I do immediately.


Janus.

[Lafuente_Astronomy]

Not replace, supplement.
What I ran into is that once you are away from SOL, once it is no longer the closest star, searching for "Earth" gives you a blank result.
For instance, go to wolf 359, look around, then hit enter, type in earth, see what matches.

That has been a persistent problem in Celestia, though it could be easily solved by pressing "H", going there, and then typing Earth. But of course, if what you will do will greatly improve Celestia, then I'm up for it, and will support you in your endeavors.

It should also apply to searching for stars in other galaxies as well

[onetwothree]

I am trying to figure out how to make a search from wolf 359 be more informative.
For instance, searching for earth from there could show 'Sol:Earth' in the list.
Perhaps I could type in 'Sol:' to search for things in just our solar system, then 'Alpha Cen A:' to search only there.
Still not a universal search really, but a nice way of better presenting what is already present.

Just type Sol/Earth What you want is implemented years ago

[Janus]

Nice to know on the path thing, and I understand how I didn't know.
Typing slashes into text prompts is contraindicated, normally resulting in undefined behavior, so I never looked at it that way.
When I saw the path separator in the documentation, I took it as directories since that is how it is structured.
I never understood what those directories were, or where they were however.
For me, form follows function, so when I saw directory structured information, I treated it as directories.
Layered databases normally show layers like Sol:Earth:Luna or Sol.Earth.Luna to distinguish them.
The DOM of a webpage is a good example of how that works.

Still not quite what I meant though.
One has to know that Earth goes with sol first in order to do that.
That presupposes knowledge on the searchers part, which hides unanticipated corollary information.
What I am looking at is having sol:Earth or sol/Earth shown when just Earth is typed while away from Sol.
With any local matches shown first of course.
It shouldn't be to hard since bodies is its own structure, with parent links which will give you names.

This allows a search to match any part of any entry, making expansive results rather than just narrowing.
Expansive results encourage random exploration, which encourages learning.
They also help fill in missing or incomplete knowledge.

Now sure how to do it, but I will find it.
This is part of why I am tinkering with Celestia, to learn.


Janus.

[Lafuente_Astronomy]

Now sure how to do it, but I will find it.
This is part of why I am tinkering with Celestia, to learn.

And so far, your tinkering has lead to advances in Celestia. Imagine 100 GLY Star limit. That was not possible before because the limit was 10 Million LY. But you did it. So keep up the good work. Can't wait for your next release of the fork, either an improvement of 5441 or a new one altogether

[vpontin]

How can i install Janus's forks? Particularly the 100 billion LY limit? Should i get the data .zip from the celestia build repository?

[Lafuente_Astronomy]

Go to the previous 20 posts. Janus posted them as zip files

[vpontin]

Oh i'm currently using celestia-5441-RaDecDist-Fix.7z from Janus with Celestia 1.7 data, and sometimes when rotating bodies in outer solar system (aka a Planet Nine extra i have) the stars sometime flicks.

I got this bug in two machines: a desktop with AMD RX 560 GPU (drivers updated) and a laptop with Intel onboard HD Graphics (Lenovo G50-80).

[pirogronian]

Continuing topic about official Celestia & Celestia Origin integration... I imagined rather shipping CO in modular way, making avaliable to easly customize, what to include in personal Celestia installation. I already wrote about it in CO topic, but got answer that monolithic distribution i one of the project assumption... Someone else had to made it modular, but I don't know much about current status.

Also LUT can be hosted as separate packet, however still wihin official distribution.

[Lafuente_Astronomy]

Continuing topic about official Celestia & Celestia Origin integration... I imagined rather shipping CO in modular way, making avaliable to easly customize, what to include in personal Celestia installation. I already wrote about it in CO topic, but got answer that monolithic distribution i one of the project assumption... Someone else had to made it modular, but I don't know much about current status.

In short, you're going to make it customizable? Or is there more than just that?

[pirogronian]

I want it customizable, but can't do it myself, it's too big... task It would be useful, however, if CO team would expose its files on public repo (can be read only) to enable access for every one who want to customize it for self.
Meanwhile I have to focus on rewriting Celestia internals

[Janus]

@vpontin
I can confirm the texture glitch.
5436 appears clean, 5441 loses textures.

You can check it by selecting porrima A, going to ~3.5Ly, and rotating.
The body/texture will vanish, then reappear as you rotate.

I have no idea as of yet.
Though it may be related to the sudden displayed star size changes at ~1Ly.


Janus.

[Lafuente_Astronomy]

Good morning from the Philippines guys! Anything new today?

[vpontin]

@Janus

In fact a minor form of this glitch happens in the latest x64 build in BinTray, but only the Sun flicks.

[Janus]

Further testing and comparison.
A lot of stars have their textures go away.
In commit 5436, this does not occur, yet it does in 5441.
For comparison, I chose Porrima A, which consistently loses its texture.
Commit 5436 has roughly twice the frame rate on my system of 5441.
At ~18Ly, Porrima A shows in both, yet in 5441 the texture vanishes at ~14Ly, and the frame rate jumps.

Commit 5436, Porrima A
~18Ly -> ~360fps
~14Ly -> ~360fps

Commit 5441, Porrima A
~18Ly -> ~140fps
~14Ly -> ~155fps

Since the only changes in between commits are textual, it is confusing.
Yet also interesting, this is much closer to what I normally do, and hence more familiar.

I believe it is the greek letter substitution that is the problem, though I could be wrong, it simply feels like it.
I will have to wait for the memory call stack profiler to get done, then I will know more.
Does anyone know if the linux builds are doing the same thing?


Janus.

[Lafuente_Astronomy]

I believe it is the greek letter substitution that is the problem, though I could be wrong, it simply feels like it.
I will have to wait for the memory call stack profiler to get done, then I will know more.

Come to think of it, although most of the Constellation lines are back, some are not appearing possibly due to the Greek Letter substitutions. In my Celestia program, there was a problem in which Acrux and Gacrux didn't show the Constellation line that connected them. I had to change the names of the two stars in the asterisms.dat file to Acrux and Gacrux from Alpha Cru and Gamma Cru to make it work. This same problem also happened in some other constellations in my program.

So, perhaps a further refining of the Greek letter Substitution could do, or just to double check and make sure, a revising of the asterisms.dat file as well. But either way, it's your call.

[Janus]

Okay, here we go again.
I believe I have fixed the problem, not fixed, but bypassed.

I recreated my mod for RaDecDist from scratch in a clean copy of 5441, and now it all works.
Instead of posting it here however, I making a separate thread for it.
This is supposed to be about development, not snapshots of people tinkering, ideas.

Thus I am moving myself over to a "Forking committed thread." and posting there.
I am posting commit 5441 with RaDecDist right now.
For those who care, see you there.


Janus.