Celestia Forums

Thanks.

Awesome! Now I just have to wait for Janus to release it outside the torrent file.

Celestia Origin has some neat addons indeed. It is very up to date, and I would love to add as many as possible, though it would be tedious to add them all to my already heavily modified Celestia installation.

To receive a direct link to the archive, please contact me in private messages.

To anyone who is interested, or who does not, can not, or chooses not to use, bit torrent.

I have uploaded the new version in my downloads section.

http://celestia.simulatorlabbs.com/Downloads/

If you look in the Celestia-Origins directory you will find

celestia origin 171230.7z

Along with a few par2 files.
They can verify the download, and repair a minor glitch in it to prevent having to redownload the whole thing for those who have datacaps.
quickpar or multipar either one will work for this, generated with quickpar.

I hope this helps everyone.


Janus.

Ah, thanks!

Added after 3 hours 34 minutes:
Actually. Janus, can you upload the new version as a zip archive, and not 7z?

LukeCEL wrote:Added after 3 hours 34 minutes:
Actually. Janus, can you upload the new version as a zip archive, and not 7z?

Sure, I normally do zip, but I was in a hurry and forgot to check.

... ... ...

Uploading now.

3.4Gb plus par2 files.


Janus.

Hi Janus, thank you for your precious work.
But this time, alas, I have problems with .zip files, too.
When I launch winrar 5.5 I receive this message:
"The archive is damaged or in unknown format"
It's my business only, or someone else has the same problem?
Thank you for the help, and Happy 2018 to all of you!
Goofy

@Goofy
I did say uploading.

If the zip you grabbed is less than 3,439,555,185 bytes, you should try again.

And if someone's last post says uploading now, check first, or at least check a couple of times and see if the file size changes..

Also, did you check integrity using the par2 files?


Janus.

Thank you Janus, really my file was about 3,100,000,000.
Downloading it again.
Bye
Goofy

I remember seeing that the whole of the release is around 4.3G in size...Whatever happened to the extra gigabyte of stuff? Also, what do you do with these so-called par2 files once you have them?

CM1215 wrote:I remember seeing that the whole of the release is around 4.3G in size...Whatever happened to the extra gigabyte of stuff? Also, what do you do with these so-called par2 files once you have them?

The torrent file is a collection of individual files as a directory tree.
I zipped them up into a single archive like any other directory tree.

As for the par2 files, those are called parchives.

https://en.wikipedia.org/wiki/Parchive

http://www.quickpar.org.uk/

https://multipar.eu/

Parity for archives is the simplest way to label them.

They use a galois 2^16 field/reed solomon/Crc16/recursive blockchain system of checking data integrity.
For ensuring data sets are intact, and are used internally in many applications, including some file systems, virtual or otherwise, when the size of the filesystem and data set exceeds the bit error rate of the HDDs that it is on.
All of the above are the same thing by the way, just labelled differently because they are called different things by different fields and sub fields of mathematics.
They used to confuse me a lot because of the way so much of the terminology is defined circularly.
In the end though, it all devolves to assembly, and though tedious, assembly hides nothing from the patient debugger.
And now that I have implemented them in so many languages/assembly languages/systems, I no longer think about them at all.

A shorthand method of explaining them is thus.
Take the 3500Mb archive above.
Get the MD5 of it, store it.
Divide the file into 5Mb pieces, and a parchive set can cover more than one file by the way.
Assign each a block number and an MD5 of its own.
Xor all of those blocks against each other such that

parblock[0][0] = block[0][0] xor block[1][0] xor block[2][0] ... block[i][0]
parblock[0][1] = block[0][1] xor block[1][1] xor block[2][1] ... block[i][1]
parblock[0][2] = block[0][2] xor block[1][2] xor block[2][2] ... block[i][2]
...
parblock[0][J] = block[0][J] xor block[1][J] xor block[2][J] ... block[i][J]

I wrote that C styled, but it has the same basic structure in most languages.
It works on word/uint16 though, not bytes.

That gives you parblock[0]

Now if you xor all the blocks, excluding say block[7], but including parblock[0], you will have block[7]

Using a principal called relative primaty, a tree is established for how many bits to shift each word for the following steps.
0x0001 becomes 0x0002 etc, just like it does with crc etc.
Apply this to each block, to create parblock[1] and however many you want.

Now you can recover two blocks, then apply a backshift 0x0002 becomes 0x0001 and so forth, you have another recovered block.

Much of sourcecode is hard to read because it is written for brevity, not readability.
Why so many programmers live their lives in perpetual preparation for the C/C++ obstufication contests is beyond me.
But that rant is for another day.

I can do a real in depth write up if needed.
But this should convey the basic idea.

If you have the par2 files, you can recover/recreate as much of the main file as they represent.
In this case, each block is good for 5Mb.
Brave souls can go mess up a few sections, add a megabyte of zeros somewhere in the middle, and watch it be recovered if they like.

I hope this answers what par2 files are for, and why I use them.
It is not that I distrust tcp, it is that I know what irrecoverable bit error rates are, and how much data is flowing around out there.


Janus.

Thank you, but, not being the coding genious that you are, I cannot make out what you are trying to say...

Perhaps what I am trying to ask is this: are they necessary, or is simply downloading the zip file all I need to do to get all of Celestia Origin?

@Cm1215

Believe me, I am no coding genius.
I struggle many days, but my googlefoo is strong, and I am unafraid to wade though reference material, RFCs and assembly.

It is discouraging how often programming problems come down to some genius grabbed a sample from somewhere, changed all three variable names relavent to their code, then complained when it didn't compile, or did something weird.
Ignoring of course, the fourth or fifth variable(s) that have a shared name, and they ignored the redeclaration with the same type warning, assuming they were even compiling with warnings turned on.
Somehow, the idea that a redefined variable is not blanked does not occur to so many programmers.

However, in this case, only the first par2 file is actually relevant if you have a stable connection, and even then only if you want to double check.
Everything will be fine for most people, but I always include error correction when sending data to customers, and they have gotten to be a habit.
Also, a lot of people, especially in the US, have data caps, which are pure B.S. by the way.
So them being able to grab a fix saves them data.
100Mb is a lot smaller than 3500Mb, which is why I included them.


Janus.

I like the exoplanet textures, but there is a problem with a few of them. You took satellite images of many countries and just pasted them in there, which I find abnormal.

Joey P. wrote:I like the exoplanet textures, but there is a problem with a few of them. You took satellite images of many countries and just pasted them in there, which I find abnormal.

Authors of these textures are not us, but it is very difficult to identify a certain country in them. For example, I did not notice this before.

Okay, for those who have trouble setting up Celestia-Origin, here is something for you.
In my downloads is a new archive.

Celestia_Origin_2017_12_30.zip

Notice the extra underscores to distinguish it from the others, and it says 2017, which the others abbreviate.

This archive contains an already installed Celestia-Origin with a celestia fork for windows in place, ready to run.
The exes are x86 & x64 both, VS2013 compiled & completely statically linked, it should not require even the VS redists.

The fork is one I am working on.
It displays Ra/Dec based on Celestia-DateTime for Stars and most solar system bodies.
This means if you select Venus or Mars or whom ever, you can get its real time coordinates, or where it will be at the date/time you set.
The Observer distance has been moved to the lower right, which this package covers.

It also includes a tweak to load everything in alphabetical order.
It reads directory contents, separating the subdirectories from the files.
Sorts both non case sensitive.
Parses and loads the files first in order.
Then goes into the subdirectories, also in order.

However, the regular celestia will also work just fine.
Tweaking the bat file for install is hard some, especially those of us who do not use our C drive for anything except the OS.
I am simply trying to save some people some work and hassle.

Loading is slow, this is a huge package, but very well assembled.

I hope this does someone some good.


Janus.

Mac version ?

bh wrote:Mac version ?

In our team everyone works only with Windows. It is well known that the collection works in Linux under Wine. If there is an analog in the Mac, then try it.