IRC logs of #tryton for Friday, 2010-07-02 #tryton log beginning Fri Jul 2 00:00:02 CEST 2010
-!- ikks(~ikks@ has joined #tryton03:06
-!- yangoon( has joined #tryton05:18
-!- sharoon( has joined #tryton05:47
-!- mr_amit(~amit@ has joined #tryton06:30
-!- gremly(~gremly@ has joined #tryton06:52
-!- paepke( has joined #tryton07:32
-!- eLBati(~elbati@ has joined #tryton07:37
-!- Timitos(~timitos@ has joined #tryton07:49
paepkecedk, down?07:58
-!- gremly(~gremly@ has joined #tryton08:13
cedkpaepke: I check08:20
cedkpaepke: host answer to ping but nothing else08:20
cedkpaepke: I will reboot it08:20
-!- irclog( has joined #tryton08:24
paepkecedk, ok. just wanna inform you. anyway: good morning :-)08:24
cedkpaepke: back online08:25
cedksomething took all the memory and the kernel started to kill random process08:28
-!- cedk(~ced@gentoo/developer/cedk) has joined #tryton09:10
-!- eLBati(~elbati@ has joined #tryton09:53
-!- bechamel( has joined #tryton10:03
-!- paepke( has joined #tryton11:06
-!- Red15(~red15@unaffiliated/red15) has joined #tryton11:29
-!- tekknokrat( has joined #tryton12:13
-!- eLBati(~elbati@ has joined #tryton12:18
-!- eLBati(~elbati@ has joined #tryton13:13
-!- paepke( has joined #tryton13:47
-!- eLBati(~elbati@ has joined #tryton14:19
-!- woakas(~woakas@ has joined #tryton14:42
-!- paepke_( has joined #tryton15:18
-!- paepke_( has left #tryton15:54
-!- cedk(~ced@gentoo/developer/cedk) has joined #tryton17:36
-!- enlightx(~enlightx@ has joined #tryton18:18
-!- bechamel( has joined #tryton18:33
-!- enlightx(~enlightx@ has joined #tryton19:20
-!- plantian( has joined #tryton19:43
paepkei have a question regarding the ir.attachment. that collision field is for the case that two files have the same md5sum?19:46
paepkethats how i read the code. just wanna be sure19:48
-!- FWiesing( has left #tryton19:53
-!- FWiesing( has joined #tryton20:09
cedkpaepke: yes20:33
paepkecedk, are you satisfied with that functionality?20:34
paepkei'm currently looking into enhancing that module for a customer20:35
paepkecedk, have you already thought about changing the hashing algorithm to sha1? I'd like to allow a single instance storage with that.20:51
paepkecedk, i wanna ask if you would accept such a patch20:58
cedkpaepke: why changing?21:07
paepkecedk, cause sha1 has currently no exploit for generating the same hash. for single instance storage i don't wanna use that collition field.21:11
paepkewith md5 its possible to have the same md5 for different files.21:12
cedkpaepke: with sha1 also21:37
cedkpaepke: and it will be for any hash algorithm21:38
paepkecedk, but its nearly impossible. other projects like mercurial are also relying for file checks21:38
paepkecedk, i agree that its possible with any hashing algorithm.21:39
cedkpaepke: I don't see any goals with this change21:40
cedkpaepke: the hash is only used to detect same data and then store only one21:40
paepkecedk, with the current collition system we aren't able to implement a single instance storage.21:40
cedkpaepke: with any hash, you will still need to collision check21:41
paepkecedk, really?21:41
cedkpaepke: except if you accept the risk21:41
cedkpaepke: still I don't see any goals21:41
paepkecedk, really stored only once? i assumed it has a -1 on that file name?21:42
cedkpaepke: what do you call "single instance storage"?21:42
paepkecedk, to store it only once. thats what you said. but i can't read that in the code.21:42
paepkecedk, maybe i'm to dumb to find the correct snippet21:42
cedkpaepke: it is the case21:43
cedkpaepke: just try21:44
paepkecedk, i'm doing :-)21:44
paepkecedk, ok. you were again totally right. i misinterpreted how the collision occurs.21:53
paepkei shouldn't doubt the tryton code you guys wrote ;-)21:54
-!- zodman(~zodman@foresight/developer/zodman) has joined #tryton22:24

Generated by 2.11.0 by Marius Gedminas - find it at!