None that I know. The toolmakers seem to have vanished.

Report message to a moderator

OK, but may be there are still somebody around who knows the meanings of the bytes/bits in the Stars! file blocks? Or may be somebody at least have a copy of the PaulCr xml files that describes blocks format? So far, from the publically available sources, like this forum, I was able to collect info of Stars! encryption and 35 of 46 blocks structure, but 11 remains unknown for me. If there are somebody who can help with this?

PS: I'm still dreaming of Stars API, that allows reducing MM...

Report message to a moderator

Report message to a moderator

Yeah it would be nice to have those tools, I still have the encryption algorithm for anyone who's interested in writing those tools. Course you'll also have to decode what the stuff is once you decrypt it. But that'll be a lot easier when you can read it. I don't personally know C++ so the it's not a lot of use to me. (But I gave it to PaulC or whatever and it worked fine for him. At least I think that's what I did. It's been so long now. Whom ever I gave it to, made it work.)

I also have an exec that makes the files without encryption so you can test on it if you want.

(Also if anyone wants to write an AI for it please feel free.)

I can assist somewhat in what I know about how to read the data (I used to do it through the encryption without decoding it. I've got some charts about how to read some things that I've made. If anyone wants help deciphering what the data means I'll try and help.)

I think that the first thing to do is to decide what you want it to manage, I've spent time with the planet data, and I could go back and start looking at the unencyrpted version and try and give anyone who wants it a detailed summary of the data structure for planets. I've got some experience in reading the hst file (where I focused my attention) and I can probably find the ship data and figure that out for anyone who wants it. (I've also read the race data too, and well. I wrote a tool once to change a player's race through encryption without decrypting it. )

If someone wants to write the program such that the file structure data can be plugged in, I believe I can help them to read the files. (All my tools were written in batch code and are necessarily ridiculously slow as a result.)

[Updated on: Wed, 07 July 2010 19:17]

Report message to a moderator

OK, here is where I'm:

1. Decryption - I manage to found exactly the same that was used by PaulCr (I beleive it's exactly the same that Captin Maim have). And as I said above - at the moment Stars! encryption/decryption algorithm is publically available around that forum (I wouldn't describe here where it's exactly (Captin Maim if you want to know where and how I take it plz pm me), but beleive me if someone wants it - it's pretty easy to get it, I'd say it even easier then found blocks structure). As I understand from that forum, community most likely think that decryption shouldn't be public, so I don't know how bad it is. I have made a number of tests and they shows that it works perfect on any file types.

2. Blocks Structure - this one is tricky. There are a number of posts on that forum where PaulCr is describing blocks structure, so I was able to find out structure for about 36 of 46 blocks so far (some of them partially, but the stuff they contains could be skipped). What I need is to be able to load data from xy, m, x, h files, and then after making some changes to be able to write x file for beginning and h file later on. Currently xy file structure is completely decoded. m, x, h files is mostly decoded, but there are still some blocks that needs to be decoded, so any input on this is really appreciated.

I'm away from home now, so I'll post details later on, but the biggest issue I have at the moment is decoding Design and Partial Designs, also I can't understand how does Stars! match Designs to the specific ship in fleet (while I was able to completely decode fleet block), when there are 2 or more designs in m file with the same id (one belongs to you and another one belongs to your allyers/enemy). Also I'm going to spend some time on decoding race block (so far it's partially decoded), so if someone can give me a hint here, you are really welcome!

As for Server Messages block - I can decode it euristically (i.e. manually), but can't find how does Stars! do it automatically, since even for one message id there could be different agrguments length/structure.

Also I'm wondering what blocks 32 and 33 means. 32 seems contains number of followed planets and one more number that is uknown. 33 contains something that I cannot match to any type of info in Stars!, but it always placed after Planets blocks.

PS: I'm going to skip battles blocks, and wouldn't even try to decode it at least for now, since they are not important for my task. But Server Messages could be usefull.

[Updated on: Thu, 08 July 2010 08:08]

Report message to a moderator

XyliGUN wrote on Thu, 08 July 2010 11:22

As for Server Messages block - I can decode it euristically (i.e. manually), but can't find how does Stars! do it automatically, since even for one message id there could be different agrguments length/structure.

I'd hazard the guess that the Stars! exe stores a list of all message ID's and associated parameters/texts et al. Encrypted, of course.

Report message to a moderator

m.a@stars wrote on Thu, 08 July 2010 14:00

XyliGUN wrote on Thu, 08 July 2010 11:22 As for Server Messages block - I can decode it euristically (i.e. manually), but can't find how does Stars! do it automatically, since even for one message id there could be different agrguments length/structure. I'd hazard the guess that the Stars! exe stores a list of all message ID's and associated parameters/texts et al. Encrypted, of course.

Well, it's possible to collect all the messages and related IDs from Stars! (even manually, which will takes some time), but the problem here is that even for one ID (i.e. the same message) there are can be different arguments structure (i.e. 1 or 2 bytes for example), so my guess is that it somehow stores it's structure in the server messages block for each message (and there are several bytes in each message that definitely isn't arguments), but I cannot find out how.

Report message to a moderator

XyliGUN wrote on Thu, 08 July 2010 13:22

I'm away from home now, so I'll post details later on, but the biggest issue I have at the moment is decoding Design and Partial Designs, also I can't understand how does Stars! match Designs to the specific ship in fleet (while I was able to completely decode fleet block), when there are 2 or more designs in m file with the same id (one belongs to you and another one belongs to your allyers/enemy). Also I'm going to spend some time on decoding race block (so far it's partially decoded), so if someone can give me a hint here, you are really welcome!

Designs issue now solved thanks to Micha pointing me to StarsHostCreator in the another thread.

XyliGUN wrote on Thu, 08 July 2010 13:22

As for Server Messages block - I can decode it euristically (i.e. manually), but can't find how does Stars! do it automatically, since even for one message id there could be different agrguments length/structure.

Looks like I have to skip this one for now, but there are a lot of possibility for automation, so I'll return to this one later on.

So far of 46 blocks these 7 remain unknown: ## 11, 15, 18, 22, 25, 33, 41. Block 33 seems the only show stopper on the way to write h files. If you know somthing of these blocks plz let me know (does anybody have a copy of PaulCr xml files described these blocks?).

PS: I was surprised that Stars! sends name, which was assigned to fleet by player to host via x file and then host sends it back via m file, I was under impression that fleet name has no uses for host and that it only matters for player (like "Send x Pop to y").

Report message to a moderator

Report message to a moderator

Report message to a moderator

XyliGUN wrote on Thu, 08 July 2010 21:13

So far of 46 blocks these 7 remain unknown: ## 11, 15, 18, 22, 25, 33, 41. Block 33 seems the only show stopper on the way to write h files. If you know somthing of these blocks plz let me know (does anybody have a copy of PaulCr xml files described these blocks?).

Nope. Block #33 is empty and labeled as "unknown". The others aren't even listed.

Quote:

PS: I was surprised that Stars! sends name, which was assigned to fleet by player to host via x file and then host sends it back via m file, I was under impression that fleet name has no uses for host and that it only matters for player (like "Send x Pop to y").

Weird.

Report message to a moderator

XyliGUN wrote on Thu, 08 July 2010 13:36

the problem here is that even for one ID (i.e. the same message) there are can be different arguments structure (i.e. 1 or 2 bytes for example)

Weird. Exactly what Stars! messages can have variable arguments? Battle reports? MT meetings?

P.S: varargs is a common trick for c-like languages, so no big surprise here. The "parser" uses the message template itself to fetch the correct number of parameters.
P.P.S: the real problem would be, I presume, when some of these arguments are used to alter the way the template is used, and somehow code the number of remaining arguments after them.

[Updated on: Fri, 09 July 2010 13:39]

Report message to a moderator

Report message to a moderator

m.a@stars wrote on Fri, 09 July 2010 21:37

Weird. Exactly what Stars! messages can have variable arguments? Battle reports? MT meetings?

For example it's "You have built {0} factories on {1}." it could be 6 or 7 bytes long. Here is an examples of Server Messages block that correspond to that message:
3600 FD00 0F FD
3600 FD00 15 FD
3600 FD00 18 FD
3604 3901 02 3901
3604 3901 10 3901
3604 3901 14 3901
3604 3901 1A 3901
3604 0401 04 0401

First byte I beleive is message id (at least it's true for all turns that I've checked). 3 and 4 is planet id, 5 is factories number. 6 and 7 is again planet id (what for? i don't know). Sometimes like in first 3 examples it's not 7, but 6 bytes long - there second byte could be a hint (kind of format id), but it's not works for other messages.

Foe example here is a differet messages with "format" = 2:

A902 0C01 0C01 - "Your home planet is {0}. Your people are ready to leave the nest and explore the universe. Good luck." - again why does planet id stored twice? I don't know. But it's true for 10 turns that I've checked out.

2F02 0580 3801 05 - "Your starbase at {0} has built new {1} ship." here 3 and 4 bytes is unknown, 5 and 6 is planet id, last one is fleet id. But it's definitely not enough to store any fleet id, since it could be larger then 255.

5302 0086 00 06 FF - "{0} has orders to colonize {1}, but you have failed to bring along any colonists. The colonize order has been canceled."

m.a@stars wrote on Fri, 09 July 2010 21:37

P.S: varargs is a common trick for c-like languages, so no big surprise here. The "parser" uses the message template itself to fetch the correct number of parameters.

That's could be a true, I'll check it a bit later.

m.a@stars wrote on Fri, 09 July 2010 21:37

P.P.S: the real problem would be, I presume, when some of these arguments are used to alter the way the template is used, and somehow code the number of remaining arguments after them.

So far I beleive this is exactly how it is, but I can be wrong there.

Report message to a moderator

Report message to a moderator

[Updated on: Sat, 10 July 2010 14:19]

Report message to a moderator

Report message to a moderator

Don't they usually just compare hashes to see if they're identical cause they are often non-reversible?

And yes, that is where the stars serial is. I recall messing with it once. But I just wanted to test something and so I erased it to see what would happen. And if you do it asks for a serial.

[Updated on: Sat, 10 July 2010 15:32]

Report message to a moderator

Captain Maim wrote on Sat, 10 July 2010 22:18

From what I've seen in trying to understand the formatting of the headers and such, stars tends to explicitly state the byte length of things. I mean I'm sure you know that it uses 2 bytes to define a block, a block ID and the block length.

Yep.

Captain Maim wrote on Sat, 10 July 2010 22:18

Also I think in the "planet (x) has built (y) ships" as I recall, the additional value you may have forgotten the 4 bytes are probably the ship counts. Cause I'm pretty sure it can say how many of what kind you've made.

Yes, it can, but there is special message (i.e. different message id), which says planet (x) has built (y) new (z).

Captain Maim wrote on Sat, 10 July 2010 22:18

In reading the files, do you ever alter them to observe the results? I find doing that tends to make deciphering them much faster. (Tweak, observe, record, repeat.)

This will be my next step, so far I have a tool that reads Stars! file and shows it as a tree containg blocks with data extracted from them (with their meaning).

Captain Maim wrote on Sat, 10 July 2010 22:18

I was thinking that the server messages would be useful to an AI. I recall talking to PaulC about what you quoted there. It was his intention to do that so that no one needed the encryption and it would thus be of limited circulation. Course as you said, if anyone wants it they can get it, cause I guess we're not as tightly gripped about that sorta thing these days.

Yeah, AI based on the events processing would be a good idea. As for encryption as I said before I'm going to keep Stars! Password Protection feature.

Captain Maim wrote on Sat, 10 July 2010 22:18

Heck, if this evolved into a whole new (and shinier) stars client that still used the existing server. That would just rock.

Will see.

Report message to a moderator

Captain Maim wrote on Sat, 10 July 2010 23:26

Don't they usually just compare hashes to see if they're identical cause they are often non-reversible?

Yep, and since hash algorithm is known, I'll verify password by comparing its' hash with one stored in the m file.

Captain Maim wrote on Sat, 10 July 2010 23:26

And yes, that is where the stars serial is. I recall messing with it once. But I just wanted to test something and so I erased it to see what would happen. And if you do it asks for a serial.

OK, can anybody give me a hint on how can that string stored in GlobalSettigs be decoded to the real 15 bytes hash (btw it looks like it always 30 characters long, which is matched to 30 hex digits)? Any ideas?

Report message to a moderator

Report message to a moderator

[Updated on: Sun, 11 July 2010 04:34]

Report message to a moderator

Report message to a moderator

Report message to a moderator

wasp wrote on Sun, 11 July 2010 10:44

Here's a table of the message format strings and expected argument counts, straight from the executable: http://orthanc.fixme.fi/stars/stars-msg.txt

Cool!

That will hopefully help understand how the game itself works, and why there's sometimes erroneous info reported (such as who battled who and who destroyed an AR's orbital)

Many of them look like obsolete remainders of earlier versions, though. Transfer of colonists?

Quote:

The argument count doesn't include the extra arg which is the target of the Goto button.

Interesting. Where's that arg, then?

Report message to a moderator