New Celestia-1.4.0pre-FT1 (galaxies) for Testing

[t00fri]

Fridger,

Sorry I wasn't clear: I meant that there must be buggy changes in the C++ code of Celestia which is related to the vertex shaders. Among other things, it certainly is no longer properly detecting what functions are available in the card.

The vertex shader code that you are describing is the OpenGL code that gets loaded into the graphics card.

Selden,

sorry, I wasn't clear

I understood your above remark perfectly well, but my previous post was NOT directly related to it. There I just examined and described an issue that possibly could have been relevant. But it wasn't.

Bye Fridger

[selden]

OK!

[t00fri]

Hi,

if anyone here happens to know how one would correctly prepare a complete, selfextracting Windows Celestia-win32-1.4.0pre-FT1.1.exe
distribution from the sources that uses the standard install script, PLEASE let me know! What tools do I need?

It's so easy in UNIX, LINUX...

Bye Fridger

[Brendan]

I looked on http://www.freewarehome.com/, which has lots of free Windows stuff, and found some such as this http://www.jrsoftware.org/isinfo.php I didn't try any of them before. Celesita is linked to from the freeware site too.

[t00fri]

I looked on http://www.freewarehome.com/, which has lots of free Windows stuff, and found some such as this http://www.jrsoftware.org/isinfo.php I didn't try any of them before. Celesita is linked to from the freeware site too.

Many thanks, Brendan!

Let's see how isetup works for us...

Bye Fridger

[Boux]

Hi t00fri!
I have a question.
I am more a Linux guy than a Windows one.
And compiling stuff is so much easier under Linux, well, you know...
Why stars in Linux-KDE are only rendered as "fuzzy" and not as plain vanilla points?
I hate this fuzzy rendering which does not make any sense.
Where in the code does this thing is hiding itself?

[t00fri]

Hi t00fri!
I have a question.
I am more a Linux guy than a Windows one.
And compiling stuff is so much easier under Linux, well, you know...
Why stars in Linux-KDE are only rendered as "fuzzy" and not as plain vanilla points?
I hate this fuzzy rendering which does not make any sense.
Where in the code does this thing is hiding itself?

Boux,

yes Linux is much more fun, of course

But the type of the star rendering can be cycled by repeatedly pushing CTRL-S, independently of the OS. If you like solid disks -- no problem. I do, too, on my notebook, for example.

Bye Fridger

[Boux]

Thanks for the hint
I am digging into the source here and there to look for stuff and try to understand the logic and try to get the big picture.
It's pretty huge! Need to read alot and learn about the ogl rendering side.
Hehe. Brainwash sort of. Lot of fun.

[t00fri]

Someone (cartrite?) who is compiling the code both under Linux and Windows could perhaps check whether for the free Windows compiling toolkit, my proposal (outlined above) works:

We shall release the next FT1.1 sources after 'make distclean' under Linux. Then all Makefiles are gone. The Windows batch compile then works for me, but should be once cross-checked (cartrite?).

For Linux building people then simply type as usual (CVS)

> make -f Makefile.cvs

Then for KDE:

> ./configure --with-kde --with-lua --disable-debug

for Gnome

> ./configure --with-gnome --with-lua --disable-debug --with-included-gettext

for gtk

> ./configure --with-gtk --with-lua --disable-debug --with-included-gettext

and finally,

> make

Thanks,
Bye Fridger

[maxim]

Maxim,

if you take "screen size =screen height" then you lost that bet by far. Here are three of many examples. But the most striking is Stephan's Quintet, i.e 5 ~ equally large galaxies /VERY close/ together. In the lower two images, it even doesn't matter what you chose as screen size .

Well, I see
What's your estimate then - about 5?

What are the additional costs for depth sorting those prominent ones on the screen?
I'd suppose coordinate transformation of the z-component and z-value sorting is the main task. What are the /usual/ costs - coordiate transformation of the x- and the y-component !? So let's assume O(coord transformation) = O(value sorting), then it would equal to another five galaxies beeing shown on the screen - is that a valid assumption?

Sorry to bother you on that, but the galaxies are looking so good now, that I think adding the dark bands would bring it to a almost perfection.

maxim

[ElChristou]

Curiosity:
Is Chris aware of this major improvement of Celestia?
Has he react in any sort?

[t00fri]

Curiosity:
Is Chris aware of this major improvement of Celestia?
Has he react in any sort?

NO. Not at all. No time for it. We are working very hard on FT1.1

Cheers,
Bye Fridger

[cartrite]

cartrite here,

I ran a make distclean on the ft1 source with linux, then repaired engine.mak then compiled with ms visual c++ toolkit 2003 (freebe) and every thing compiled with no problems. I think thats what you were asking?

I got a question. Is the rendering of files from the extras folder in the cvs version disabled? I can't see any galaxies or nebula. They work with 1.4pre-6.

I put the old pts files in the root dir. The basic galaxies get drawn. Red though. Is that what is happening with everyone?

cartrite

[cartrite]

cartrite again

Another question. I can compile the ft1 and cvs versions with debugging but celestia will not run. It wants a file called msvcrtd.dll. It will not run without it (Debug Version). I'm assuming that the ft1 debug version needs it too.
One comes with microsoft sdk platform (64 bit) but it says it not a valid Windows image. I ran a dependency checker. I think it wants a win32 version of the file. I got a 32 bit cpu. Anyone know were to get the Win32\msvcrtd.dll file?

cartrite

[cartrite]

Cartrite here

I found that msvcrtd.dll file in a vc98 folder on a visual studio 6 installation disk I had and the debug version runs now. I did a make distclean on the source ft1 with cygwin and it compiled alright. Both the release and debug versions.
The Ft1 debug version runs too. If you got a copy of msvcrtd.dll, ( my version is 6.0.8468.0), maybe it should be included with ft1.1. The file that comes with the sdk platform is 5.2.3790.1830 and that does not work.

I'm still wondering if the rendering is disabled in the cvs version for extras folder files. Anyone?

cartrite

[t00fri]

Cartrite here

I found that msvcrtd.dll file in a vc98 folder on a visual studio 6 installation disk I had and the debug version runs now. I did a make distclean on the source ft1 with cygwin and it compiled alright. Both the release and debug versions.
The Ft1 debug version runs too. If you got a copy of msvcrtd.dll, ( my version is 6.0.8468.0), maybe it should be included with ft1.1. The file that comes with the sdk platform is 5.2.3790.1830 and that does not work.

I'm still wondering if the rendering is disabled in the cvs version for extras folder files. Anyone?

cartrite

Cartrite,

thanks for testing. The result is as I hoped for. The msvcrtd.dll issue is somehow known to me from Version 6.x of VC++ . However I don't think we shall attach one to the sources. There is no problem with VC++.net, but there might be an interference and incompatib;e versions...

Yes, the rendering of the extras folder was disabled. It will work again in FT1.1 . It's due any time.

Bye Fridger

[cartrite]

Fridger,

I ran into another small issue compiling FT1 with linux.

With suse 9.3 I compiled with gcc 3.3.5-5 and everything works well as far as I can tell. With suse 9.2 I compiled with gcc 3.3.4-11 and it compiled. When I ran it all the galaxies were gone. no trace at all. The title still appears at the top left corner. I tried the latest driver for linux. nvidia 7676 still nothing.
Yesterday I tried the source 1.3.2 with ms vc++ toolkit and that would not even compile. Fatal errors. Maybe a lot of the issues people have are really compiling issues?
I think I'm going to run the versions that work for a while and see if I can find any bugs. Installing that driver 7676 for linux took its toll on me. It started giving me problems and had to be removed. My poor computer almost experienced my wrath!!!!!!!!!!!! 5 hours what a mess that was.
Suse povides 6111 for 9.2 and thats what I'm sticking with.
Later for now,

cartrite

[cartrite]

follow up for the post above.

GCC 3.3.4-11 compiled it alright. The make install command failed to install the pts files in the models directory. don't know why. At 1st glance everything else seemed alright now. sorryjavascript:emoticon(':o')

later,

cartrite

[t00fri]

follow up for the post above.

GCC 3.3.4-11 compiled it alright. The make install command failed to install the pts files in the models directory. don't know why. At 1st glance everything else seemed alright now. sorryjavascript:emoticon(':o')

later,

cartrite

Sorry for being late in answering. Unfortunately I am insainly busy (professionally) for the next 1-2 weeks...

Anyhow my suspicion was right: If you don't see galaxies, the *.pts templates are typically not where they are supposed to be...

The lacking installation of the *.pts guys is probaly a flaw, since I did nothing yet in the Makefile to include them. They still have to be copied into 'models' by hand.

Bye Fridger

[cartrite]

They were included (installed) when I ran make install from gcc 3.3.5-5.

I Think, I'm going to recheck that. I may have copied them manually and forgot.

Both of these vesions are developement versions according to the gcc home page. Mabey something is wrong with them?

I built gcc 4.0 but haven't got it working right yet. Can't wait.

cartrite