1.2.4 prerelease 4

[chris]

Here's the latest Windows EXE:

http://www.shatters.net/celestia/files/celestia-win32-1.2.4pre4.zip

As before, you'll need the data files from the prerelease 1 package:
http://www.shatters.net/celestia/files/celestia-win32-1.2.4pre1.exe

This version finally fixes the exploding orbits problem . . . The image capture issue on Windows is resolved as well.

Unless anyone finds any major bugs, this will be the final 1.2.4 prerelease . . . The only changes I want to make are to resize the Umbriel texture and include the new Triton texture.

--Chris

[Rassilon]

Ok Chris,

I ran through everything...Time limit past 10000th century etc...no more exploding orbits...Did several screenshots...no problems there...looks good.

The only other smaller issues still to be tackled I guess would be multiple lighting options(When adding Emissive to a planet, allow it to be a light source relitive to planet texture color or a color you set)...Planet clipping past a certian distance from its sun(Or clipping near additional star in the case of Toliman)...clipping of objects near the surface when zoomed out or moved away from at the horizion of stars or planets...but those are mostly cosmetic...

There is still the abilty to pass the allowed length of your time variable...which goes out of bounds in the negitive range...And sometimes -1K still shows for planets given high albedo values...

I noticed if you give them albedo values of 0 they disappear

Personally for Windows, unless there is some other compatability issues with other video cards, I think you have all the major bugs sorted out...

What were you thinking of adding in release 1.2.5?

Thanks again for your efforts!

[Rassilon]

I did notice a difference in zooming in on objects from planet surface...As the zoom gets closer in...the fps drops down to 1 frame per second...Was this to rid the engine of the jitter?

[Mikeydude750]

Actually, the orbits do change. Instead of exploding, after a while, they just go into the sun. I tried it. It probably can't be stopped, but I just wanted to bring this to your attention.

[chris]

Actually, the orbits do change. Instead of exploding, after a while, they just go into the sun. I tried it. It probably can't be stopped, but I just wanted to bring this to your attention.

How long is 'a while'? Is this a problem with overflowing double precision floats?

--Chris

[Mikeydude750]

Overflowing double precision floats? What the f*** are those?

Yeah, it happens after several tens of thousands of years.

[chris]

Overflowing double precision floats? What the f*** are those?

Yeah, it happens after several tens of thousands of years.

Ack! Only several tens of thousands of years? Something's messed up then . . . I can go millions of years into the future and the orbits remain stable. I'll have to take a look at this before releasing 1.2.4 . . .

--Chris

[Rassilon]

Actually, the orbits do change. Instead of exploding, after a while, they just go into the sun. I tried it. It probably can't be stopped, but I just wanted to bring this to your attention.

How long is 'a while'? Is this a problem with overflowing double precision floats?

--Chris

Even during the overflow I didnt experience this explosion on my system...

1.5 gug Intel (not celery)
384 RAM
GF 3 Ti 200 64 Meg
Win XP Pro

[ravinfinite]

hi guys

the orbits DO remain stable after a gazillion years. one thing that i noticed is the rate of speeding up time somehow affects the position of the planets.

let me clarify. if you make time runs about 10e10 times faster, everything remains perfect. however, when you make time run at 10e26 times faster, the planets simply disappear. only a few comets remain and after a while they disappear too.

i think the speed increase should be limited. for example, the maximum time should be 10e15 or 10e20 or something else less than 10e26 ~ 10e27. if you wait like 10 seconds, Celestia freezes. if you wait 5 more seconds, you're computer freezes. besides, it doesn't make much sense to view the universe go by that those speeds anyway. in just one second of going at that speed, in theory (someone correct me if i'm wrong), there should have been repeating big bangs, cooling of the universe, maximum proton lifetime reached, etc.

my system specs are:

Athlon 1.4 GHz
512 MB DDR RAM
Asus GeForce 3 w/ 64 MB DDR RAM
Win2k

someone verify this...

thanks

ravINFINITE

[Rassilon]

I just tried super fast time lapse as you suggested...well I think it was 10e25 or 10e30 and the planets were still there after several loops through the overflow and back...It must've went positive to negitive and back like 10 or 20 times before I stopped and I still couldn't reproduce the dissapearing planets or the lockups...Looks like this could be due to system configuration...etc...

I would suggest the above...limit time lapse and the year length as well...That should clear it up...

[ravinfinite]

hi people

i just tried it again and the same thing happens. as soon as celestia fires up, i go straight to the sun. i then zoom out with the sun in the middle looking at the solar system from above (like a bird-eye view). if I speed the time up to 10e26, everything in the solar system disappears except for the comet Ikeya-Zhang (or whatever). If I speed it up to 10e27, then it disappears too.

I think you're right about it being my system. Oh well, if the speed of time (sounds kinda funny!) is limited to a certain level, I think that should prevent this from happening. I guess I should just limit myself!

thanks

ravINFINTIE

[chris]

hi people

i just tried it again and the same thing happens. as soon as celestia fires up, i go straight to the sun. i then zoom out with the sun in the middle looking at the solar system from above (like a bird-eye view). if I speed the time up to 10e26, everything in the solar system disappears except for the comet Ikeya-Zhang (or whatever). If I speed it up to 10e27, then it disappears too.

I think you're right about it being my system. Oh well, if the speed of time (sounds kinda funny!) is limited to a certain level, I think that should prevent this from happening. I guess I should just limit myself!

thanks

ravINFINTIE

Thanks for all the testing . . .

It's easy for me to add a change that limits time acceleration to 1e15 or something . . . There's nothing interesting happening on that timescale in Celestia anyhow.

--Chris

[Mikeydude750]

Thanks. We like beta-testing.

You know, no one told me what overflowing double precision floats were

[Rassilon]

double is a floating point identifier used in C++ and other programming languages and in a range of 2.2 x 10^ -308 to 1.7 x 10^ 308...

[Mikeydude750]

Oh, thanks Rassilon. But can't they go any higher?

[Rassilon]

Yeah long double...

Heres the chart for standard sizes in C

http://va5-3.iso.port.ac.uk/docs/digital_unix_40D/AQTLTBTE/DOCU_031.HTM#sizes_sec

[Mikeydude750]

Oh, I see now. The more digits, the larger amount of RAM is required.

Well, I guess my dad's computer could handle a 512 bit number =P.

[Rassilon]

I doubt it as the number range for something like 512 bit would be ~1.7 x 10 ^ -5917353218191003681 - 1.7 x 10 ^ 5917353218191003681...

And first it would have to be defined in order to be used...

Besides that's as close to infinity as I would like to get mind you