http://forums.civfanatics.com/downloads....le&id=7531
The nag feature is not in yet, but he says it might be in this weekend.
Darrell
The nag feature is not in yet, but he says it might be in this weekend.
Darrell
|
PBEM4 - Email and Technical Issues
|
Cyneheard Wrote:Sorry, forgot that we'd decided to keep the game going while waiting for T-Hawk. Save sent to Whosit. We're waiting for Mist at the moment (more precisely Mist's computer), trying to brute-force crack amnesiac Krill's admin password.
Which at the moment is at :
(a-z, A-Z, 0-9) bruteforce passwords 5 letters or shorter - checked and rejected 6 letter password - over 6% progress 7 letter password - around 2% 8 letter password - under 1% (a-z, A-Z, 0-9) birthday attack 6 letter password - pass with distance 30000 finished, distance 29889 in progress (a-z, 0-9) bruteforce started, at password length 6 Epiphany In queue. Since my computer won't handle any more parallel attempts and that's basically everything I can throw at it. So keep playing and don't mind me, it *will* take time. But I'm still odds-on to finish before you're done with the game  If the password is under 10 letters 
Mist Wrote:Which at the moment is at : Mist, This seems like it would be a really parallel task. Is there any way you can make some sort of "holding@home-esque" distributed computing thing? 
You're right. It would be an excellent material for @home style processing, but I'm working within my limitations here. I'm an amateur coder with no prior experience in distributed computing. The closest I can get to that model is compile a bunch of small C++ apps with sub-ranges to check and upload them via dropbox for anyone interested to run.
But that would smack of desperation I still hope Krill set something shorter than 8. Once/if length 6 gets included and Epiphany throws a blank ( it's not like I can write anything that beats 21 days in a dedicated cluster anyhow ) I'll get back to this and raise a general call to arms. Actually the best/fastest way to do this would be to get a hold of a set of Rainbow Tables with the software to use them. But at 24GB+ for the range we need it's well outside capabilities of my connection and free HDD space. sunrise089 Wrote:Mist, Just reading this I wish I could make available some of the parallel computing power I use at work for this. 2048 CPUs? mh
"You have been struck down!" - Tales of Dwarf Fortress
--- "moby_harmless seeks thee not. It is thou, thou, that madly seekest him!"
On another angle entirely:
Mist, you said you found where in the save the MD5 hash is stored. Is it possible to change that value to the MD5 hash of a known string and store it back into the save? Will that run afoul of any checksum or anything? I made an attempt along these lines last night. In the DLL, I found where the admin password (hashed) is read from the save into a variable. I replaced that variable by setting it to a known string (hashed). That got Pitboss past the admin password check, but then it threw an error "You are trying to load a corrupt save". I couldn't figure out what was going on there - best guess is some additional password or checksum done in the Pitboss executable rather than the DLL.
If you get desperate there's always the cloud...
http://aws.amazon.com/elasticmapreduce/pricing/ 36 cents a day per computer at work.
I have to run.
Quote:Mist, you said you found where in the save the MD5 hash is stored. Is it possible to change that value to the MD5 hash of a known string and store it back into the save? Will that run afoul of any checksum or anything?When I tried that, my Civ reacted with a splendid CtD. There are 4 md5 hashes in the save header that are not storing passwords and another one at end of file. So far I can't figure out how, and what for are they used - probably options, assets check and file integrity but I had no luck figuring out details . No progress from this angle. If anyone's interested and has other ideas, hash we need to break is this one: 873af6344848d162ddfa8801525496fa |
