HomeSearchPrevious dayNext day

00:05:12qwertyasdfuiopghjkl (qwertyasdfuiopghjkl) joins
00:46:48TheTechRobo quits [Remote host closed the connection]
00:47:27TheTechRobo (TheTechRobo) joins
01:56:27Larsenv (Larsenv) joins
01:58:49TheTechRobo quits [Remote host closed the connection]
01:59:29TheTechRobo (TheTechRobo) joins
02:02:50TheTechRobo quits [Remote host closed the connection]
02:03:23TheTechRobo (TheTechRobo) joins
03:08:18TheTechRobo quits [Remote host closed the connection]
03:08:52TheTechRobo (TheTechRobo) joins
03:57:24<@JAA>Today, #imgone made me look at how the 'random order brute' code works, and it appears to simply bruteforce random codes, using MD5 to generate random numbers from the sequence number: https://github.com/ArchiveTeam/terroroftinytown/blob/80ffc526a8b3fd188e6f73fab7b425af61f45d28/terroroftinytown/services/rand.py
03:57:53<@JAA>This means repeated codes and no guarantee of complete coverage, unless I completely misunderstand something here. :-|
04:39:40<klg>I think this is the https://github.com/ArchiveTeam/terroroftinytown/issues/10; the code you linked at least uses md5 in 'counter mode' which is better than iterated md5 mentioned in #10
04:41:30<@JAA>Ah, indeed.
04:41:55qwertyasdfuiopghjkl quits [Client Quit]
04:43:24qwertyasdfuiopghjkl (qwertyasdfuiopghjkl) joins
04:48:18<klg>I think the major problem here is that all the off-the-shelf pseudorandom permutations are designed for powers of two while power of something like 26 or 36 is needed; but here in rand.py, it looks like simply 8 bytes of output are used, so it should be easy to replace md5 with something like 64-bit block cipher with fixed key to make it prp
04:52:32<@JAA>I'm not entirely sure I'd go down the encryption route. That would only work correctly if we want to scan the entire keyspace of one block (or multiple blocks), I think.
04:53:00<@JAA>Since encrypting a single block should be a guaranteed 1-to-1 mapping.
04:56:47<klg>yes, one block only
05:00:07<@JAA>So all we need is a weird block cipher that operates on 6 base36-characters etc. :-P
05:00:47<@JAA>A bit of base conversion and bit shuffling should do the trick.
05:00:49<klg>but transform_sequence_num simply uses md5 as mapping from 64-bit integers and 64-bit integers so there's already assumption you want whole 64-bit space; replacgin md5 with a cipher would simply remove repetitions and make it eaasy to track coverage
05:01:47<@JAA>Yeah, true, it'd be better than the current code for sure.
05:02:21<@JAA>And it truncates the value further after the base conversion.
05:08:04qwertyasdfuiopghjkl quits [Client Quit]
05:21:10<klg>related concept is that of format-preserving encryption ftr, but implementing any of those would probably be an overkill
05:22:27qwertyasdfuiopghjkl (qwertyasdfuiopghjkl) joins
05:29:38qw3rty_ quits [Ping timeout: 252 seconds]
05:54:46Somebody2 quits [Ping timeout: 252 seconds]
05:55:09Somebody2 joins
05:58:27qwertyasdfuiopghjkl quits [Client Quit]
06:02:22qwertyasdfuiopghjkl (qwertyasdfuiopghjkl) joins
06:04:40Somebody2 quits [Ping timeout: 252 seconds]
06:08:07Somebody2 joins
06:16:23Somebody2 quits [Ping timeout: 252 seconds]
06:20:03Somebody2 joins
06:39:30Somebody2 quits [Ping timeout: 252 seconds]
06:42:57Somebody2 joins
06:55:28Icyelut|2 quits [Quit: bye]
07:02:53Icyelut (Icyelut) joins
07:02:54qw3rty joins
08:06:15qwertyasdfuiopghjkl quits [Remote host closed the connection]
08:08:18benjinsm joins
08:09:08Nickwasused joins
08:09:48benjins2_ joins
08:11:32benjins2 quits [Ping timeout: 252 seconds]
08:11:32benjins quits [Ping timeout: 252 seconds]
08:16:40@rewby quits [Ping timeout: 252 seconds]
09:01:49rewby (rewby) joins
09:01:49@ChanServ sets mode: +o rewby
09:26:58Nickwasused quits [Client Quit]
09:28:00Nickwasused joins
09:40:41Nickwasused quits [Client Quit]
10:10:44tbc1887 quits [Read error: Connection reset by peer]
14:11:29BearFortress quits [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
15:11:48TheTechRobo quits [Remote host closed the connection]
15:12:28TheTechRobo (TheTechRobo) joins
15:20:53TheTechRobo quits [Ping timeout: 252 seconds]
17:04:10benharri leaves [ping me anyways i guess]
17:32:53qwertyasdfuiopghjkl (qwertyasdfuiopghjkl) joins
17:43:10T31M quits [Quit: ZNC - https://znc.in]
17:43:30T31M joins
18:06:15qwertyasdfuiopghjkl60 (qwertyasdfuiopghjkl) joins
18:06:42qwertyasdfuiopghjkl60 is now known as qwertyasdfuiopghjkl_
18:08:04tech234a quits [Quit: Connection closed for inactivity]
18:08:36qwertyasdfuiopghjkl quits [Ping timeout: 265 seconds]
18:13:37qwertyasdfuiopghjkl_ is now known as qwertyasdfuiopghjkl
18:21:21tech234a (tech234a) joins
18:27:20qwertyasdfuiopghjkl quits [Client Quit]
18:29:41qwertyasdfuiopghjkl (qwertyasdfuiopghjkl) joins
18:41:38Icyelut quits [Ping timeout: 252 seconds]
18:59:17qwertyasdfuiopghjkl quits [Client Quit]
19:01:06qwertyasdfuiopghjkl (qwertyasdfuiopghjkl) joins
19:14:21qwertyasdfuiopghjkl quits [Client Quit]
19:15:49qwertyasdfuiopghjkl (qwertyasdfuiopghjkl) joins
20:22:30michaelblob quits [Read error: Connection reset by peer]
20:22:31TheTechRobo (TheTechRobo) joins
20:37:49<TheTechRobo>I thought this project increases a number and converts it to base whatever?
20:40:00<@JAA>It does for sequential scanning. This is about the random order ones.
20:42:14<TheTechRobo>JAA: Ah
21:39:35BearFortress joins
22:19:17threedeeitguy quits [Quit: The Lounge - https://thelounge.chat]
22:32:57threedeeitguy joins

HomeSearchPrevious dayNext day