Page 38 of 48

Posted: 21.01.2020, 22:19
by onetwothree
Any owners of old videocards here (old means with gl2.1 only)? could you test https://ci.appveyor.com/project/375gnu/celestia/builds/30267442/artifacts

Added after 4 minutes:
radeon >=9500, X8xx, X1xxx
Intel Pentium G6000/P6000/U5000, Celeron P4000/U3000. G35, G41, G43, G45, GM45, GL40, GS40, GS45, Q43, Q45, B43, 4 Series Express Chipsets

Posted: 11.02.2020, 21:25
by onetwothree
With this change - https://github.com/CelestiaProject/Celestia/pull/608/files i'm gonna disallow semi-axes for stars. Incorrect means any value > 1 or all < 1. Note: this doesn't affect planets. Any concerns?

Posted: 13.02.2020, 22:47
by SevenSpheres
onetwothree wrote:With this change - https://github.com/CelestiaProject/Celestia/pull/608/files i'm gonna disallow semi-axes for stars. Incorrect means any value > 1 or all < 1. Note: this doesn't affect planets. Any concerns?

Since all responses have been on Discord I might as well port them over. I agree with Luke/Fafers/Pedro here.

(Edit: 200th post!)

Lafuente_Astronomy wrote:Wait, so what would happen to stars with correct semi-axes as given?
I have just recently started making binary black hole addons, and I am still learning what to do with some of those measurements, and data, including semi axis
SevenSpheres wrote:@onetwothree no real concern but I don't see the point? It doesn't fix any bug, just prevents it from happening.
LukeCEL wrote:@onetwothree possible concern: backwards compatibility
onetwothree wrote:
Lafuente_Astronomy wrote:Wait, so what would happen to stars with correct semi-axes as given?
@Lafuente_Astronomy Nothing changes for them.
SevenSpheres wrote:@onetwothree no real concern but I don't see the point? It doesn't fix any bug, just prevents it from happening.
@SevenSpheres let's be on the safe side :)

LukeCEL wrote:@onetwothree possible concern: backwards compatibility
@LukeCEL please see https://en.wikibooks.org/wiki/Celestia/1.6.0_Features: "For stars, we specify the shape with SemiAxes [ x y z ] and this is multiplied by the radius.", "even if the radius is omitted in the stc file, a star always has an implicit radius based on the blackbody calculation."
for me that clearly means that for stars semi-axes no values greater than 1 are allowed and at least one of them must be equal to 1

otherwise we have an absurd case when actual star is larger or smaller then its bounding sphere defined by its radius
Fafers_br wrote:@onetwothree One alternative to disallowing could be normalizing the values, dividing the three of them by the greater one. Then it would work with whatever values are in the Semiaxes. Well, providing no one starts using negative values. You know, the users allways can push the developers beyond their limits... :smile:
onetwothree wrote:that's was one of my idea, but i rejected it cause better to clearly manifest error, so that it will be noticed and fixed
pedro_jg wrote:Should we really break things on purpose? Other than render bugs, over-unity semiaxes worked fine. Also, this change only hides the bug instead of fixing it.
LukeCEL wrote:That is what I was thinking. Also, some add-ons like this one (http://www.celestiamotherlode.net/catalog/show_addon_details.php?addon_id=1552) have definitely used over-unity (this one uses SemiAxes [ 1.5 1 1 ])
pedro_jg wrote:That's why the idea suggested by Fafers_br would be the best solution IMHO, since it would automatically fix the problem, as well as keeping compatibility

Posted: 14.02.2020, 04:52
by Lafuente_Astronomy
Well, I have to agree with all of the above. But perhaps we may also have to consider the Dev's position as well? We will await their response

Posted: 14.02.2020, 21:40
by onetwothree
I strongly oppose idea that Celestia code should have fixes and workarounds for incorrect addons. So, taking into account that the community is against the proposed change, it won't be implemented.

Posted: 15.02.2020, 06:00
by Janus
A suggestion.
Go ahead and test for out of range options, and log them if logging is turned on.
That way, all a developer has to do is turn logging on to test data and rendering.
No need to restrict it to just this, anything out of range or any bad combination could trigger a log note.
Include the entire entry that causes the issue to make it easier to find.


Janus.

Posted: 15.02.2020, 18:16
by SevenSpheres
Just so everyone knows; in response to this post, onetwothree has made Askaniy, LukeCEL, and myself part of the dev team. Askaniy and I are the closest to Windows maintainers Celestia has at the moment.

Posted: 15.02.2020, 21:57
by Lafuente_Astronomy
SevenSpheres wrote:Just so everyone knows; in response to this post, onetwothree has made Askaniy, LukeCEL, and myself part of the dev team. Askaniy and I are the closest to Windows maintainers Celestia has at the moment.

I jguess we just have to wait for Alexell to put you all in the Developers role in the Forum.

Posted: 15.02.2020, 23:44
by SevenSpheres
Suggestion: Chris Laurel's other astronomy program, Cosmographia, apparently has an iOS version. Since there has been some discussion about porting Celestia to iOS (and Android) the developers might want to look at Cosmographia's code for pointers.

Posted: 16.02.2020, 14:15
by Anthony_B_Russo10
That would be a good idea SevenSpheres.

Posted: 16.02.2020, 14:41
by onetwothree
Chris' cosmografia doesn't support ios. Only later versions were ported, but their source is not available.

Posted: 05.03.2020, 17:58
by onetwothree
today i applied small but important change which affects sprite models rendering
please test sprite models using files from https://ci.appveyor.com/project/375gnu/celestia/builds/31271178/artifacts
or build them from fix/glsl-particles-drawing branch

Posted: 08.03.2020, 15:15
by SVision
onetwothree wrote:today i applied small but important change which affects sprite models rendering
please test sprite models using files from https://ci.appveyor.com/project/375gnu/celestia/builds/31271178/artifacts
or build them from fix/glsl-particles-drawing branch


Hello onetwothree :smile:
I decided to test this update on my nebula.
Everything works, but what are the updates?

If the changes concern the rendering speed, no .It's the same.
The visibility of the right-hand models remains the same.
1.png
open cluster nebula 1860-20

Posted: 08.03.2020, 17:15
by SevenSpheres
SVision wrote:Everything works, but what are the updates?

When I tested that build your black hole disk/jets models didn't render, as I said on the pull request page.

Posted: 08.03.2020, 21:25
by onetwothree
SVision wrote:Everything works, but what are the updates?

The old code uses GL1 way to apply textures to points. The new one is GL3/ES compatible. Also that GL1 compatible code has issues on macOS.

SevenSpheres wrote:When I tested that build your black hole disk/jets models didn't render, as I said on the pull request page.

This is fixed already.

Posted: 08.03.2020, 22:12
by Lafuente_Astronomy
SVision wrote:Hello onetwothree
I decided to test this update on my nebula.
Everything works, but what are the updates?

If the changes concern the rendering speed, no .It's the same.
The visibility of the right-hand models remains the same.

Would that somehow be related to the reason why jets are not rendered at very high radii?

Posted: 09.03.2020, 05:59
by onetwothree
i have no idea what are you talkin' about. any screenshots?

Posted: 09.03.2020, 09:09
by Lafuente_Astronomy
onetwothree wrote:i have no idea what are you talkin' about. any screenshots?

Sorry for late reply. But here it is:
Jet Problem.jpg


Because 3C 273's jet is measured at around 60 kpc long, I gave it that estimate in its radius. But because of that, it is no longer rendered

Added after 1 minute 20 seconds:
Anyways, this was before I gave it the correct inclination, so do not worry

Posted: 16.03.2020, 21:34
by onetwothree
please test this build, especially if you have an old video card.
@Anthony_B_Russo10 this is something for your laptop collection :)

Posted: 16.03.2020, 23:05
by selden
It has problems on my laptop:

1. no GUI (corner) text is visible whether terse or verbose is selected.
2. There's a gap in the orbit next to the Earth.
3. Note that the OpenGL info window is improperly formatted, too.

I copied all of the files from the .7z archive into (a copy of) the same directory where I'd been running Janus' ttf and txf versions, so I don't think any fonts are missing.

I do have an old video card, which supports only up to OpenGL 3.3. (see the OpenGL info in the attached screengrab.)

capture_005_16032020_190006.png
screengrab showing problems