IRC logs of #shogun for Tuesday, 2016-05-24

--- Log opened Tue May 24 00:00:09 2016
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has quit [Quit: Leaving.]		00:03
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has joined #shogun		00:04
-!- mode/#shogun [+o HeikoS] by ChanServ		00:04
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has quit [Quit: Leaving.]		00:14
-!- Saurabh7 [Saurabh7@gateway/shell/panicbnc/x-wnnuuobwbmuhgbpn] has quit [Ping timeout: 260 seconds]		02:42
-!- Saurabh7 [Saurabh7@gateway/shell/panicbnc/x-snpnavgctsbffalf] has joined #shogun		03:05
-!- Saurabh7 [Saurabh7@gateway/shell/panicbnc/x-snpnavgctsbffalf] has quit [Ping timeout: 260 seconds]		05:12
-!- Saurabh7 [Saurabh7@gateway/shell/panicbnc/x-qkwwseokbysimqpf] has joined #shogun		05:34
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:51
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:51
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:51
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:51
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:51
lisitsyn	cleaning your kboard?	06:51
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:51
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:51
lisitsyn	:D	06:51
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:51
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:52
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:52
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:52
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:52
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:52
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:52
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:52
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:52
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:52
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:52
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:53
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:53
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:53
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:53
lisitsyn	uhm interesting wiking-mode	06:53
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:53
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:53
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:53
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:53
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:53
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:53
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:54
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:54
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:54
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:54
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:54
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:54
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:54
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:54
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:54
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:54
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:55
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:55
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:55
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:55
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:55
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:55
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:55
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:55
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:55
-!- arianepaola [~ariane@unaffiliated/arianepaola] has left #shogun []		06:55
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:56
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:56
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:56
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:56
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:56
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:56
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:56
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:56
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:56
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:56
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:57
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:57
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:57
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:57
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:57
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:57
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:57
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:57
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:57
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:57
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:58
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:58
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:58
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:58
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:58
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:58
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:58
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:58
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:58
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:58
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:59
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:59
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:59
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:59
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:59
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:59
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:59
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:59
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:59
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	06:59
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	07:00
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	07:00
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	07:00
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	07:00
@wiking	``````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	07:00
@wiking	`````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````````	07:00
@wiking	mmm	07:00
@wiking	error :)	07:00
shogun-buildbot	build #243 of deb1 - libshogun - PR is complete: Failure [failed configure] Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun%20-%20PR/builds/243 blamelist: OXPHOS	07:32
@wiking	lisitsyn: :) fell asleep while coding :P	07:36
lisitsyn	wiking: lol	07:37
-!- arianepaola [~ariane@unaffiliated/arianepaola] has joined #shogun		09:29
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has joined #shogun		10:04
-!- mode/#shogun [+o HeikoS] by ChanServ		10:04
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has quit [Quit: Leaving.]		10:59
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has joined #shogun		11:02
-!- mode/#shogun [+o HeikoS] by ChanServ		11:02
-!- sanuj [~sanuj@117.203.1.250] has joined #shogun		11:21
sanuj	HeikoS, yo	11:22
shogun-buildbot	build #244 of deb1 - libshogun - PR is complete: Success [build successful] Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun%20-%20PR/builds/244	11:22
@HeikoS	sanuj: hihi	11:22
sanuj	i saw the cookbook errors	11:22
@HeikoS	how are things?	11:22
sanuj	didn't built it in csharp and r on my pc	11:22
sanuj	good	11:22
@HeikoS	sanuj: the error comes from the fact that a double is passed to a method that accespts int	11:23
@HeikoS	and since java and csharp are compiled, the compiler moans about that	11:23
sanuj	yes i got that	11:23
sanuj	do i need to change json	11:23
@HeikoS	not sure	11:23
sanuj	or my cookbook page	11:23
sanuj	okay	11:23
sanuj	i'll check	11:23
@HeikoS	what we need is to define this number as an integer	11:23
sanuj	yes	11:23
@HeikoS	Real component_num = 1	11:24
@HeikoS	sould be Int	11:24
@HeikoS	but I am not sure whether the cookbook accepts that :)	11:24
sanuj	i tried that	11:24
@HeikoS	if not	11:24
@HeikoS	need to adjust the json	11:24
@HeikoS	and maybe the grammar	11:24
sanuj	cookbook doesn't build i i keep Int instead of Real	11:24
sanuj	gives error	11:25
sanuj	shall i tell you the error?	11:25
@HeikoS	no no	11:25
sanuj	okay	11:25
sanuj	:P	11:25
@HeikoS	I know why :)	11:25
sanuj	do we have a tutorial on how examples are generated	11:25
sanuj	somewhere in the wiki?	11:26
sanuj	and	11:26
@HeikoS	the readme	11:26
sanuj	HeikoS, need your feedback on this	11:26
sanuj	https://docs.google.com/document/d/1ZVV9KsHxpoj2G3HYYp-8MNjDpHVL0RGvj5hD0r03lgg/edit?usp=sharing	11:26
@HeikoS	in the meta examples dir	11:26
sanuj	me and sergey wrote a timeline	11:26
sanuj	after midterm is not very specific....might change	11:27
@HeikoS	thats good I guess	11:27
@HeikoS	seems tight	11:27
sanuj	oh you saw that?	11:27
sanuj	cool	11:27
@HeikoS	yeah	11:27
@HeikoS	fine with it	11:27
sanuj	cool	11:27
@HeikoS	sanuj: mmg	11:28
@HeikoS	whats the error when you put Int?	11:28
@HeikoS	looking at it now, it should work	11:29
sanuj	error is this	11:29
sanuj	[ 0%] Generating examples from meta-language	11:29
sanuj	Traceback (most recent call last):	11:29
sanuj	File "/home/sanuj/Projects/shogun/examples/meta/generator/generate.py", line 116, in <module>	11:29
sanuj	includedTargets=args.targets, storeVars=storeVars)	11:29
sanuj	File "/home/sanuj/Projects/shogun/examples/meta/generator/generate.py", line 56, in translateExamples	11:29
sanuj	storeVars=storeVars if target['FileExtension']==".cpp" else False)	11:29
sanuj	File "/home/sanuj/Projects/shogun/examples/meta/generator/translate.py", line 363, in translate	11:29
sanuj	return translator.translateProgram(ast["Program"], programName, tags, storeVars)	11:29
sanuj	File "/home/sanuj/Projects/shogun/examples/meta/generator/translate.py", line 54, in translateProgram	11:29
sanuj	dependencies=self.dependenciesString(),	11:29
sanuj	File "/home/sanuj/Projects/shogun/examples/meta/generator/translate.py", line 119, in dependenciesString	11:29
sanuj	allClassDependencies = template.substitute(classlist=self.seperatedClassDependencies("AllClasses"))	11:29
sanuj	File "/home/sanuj/Projects/shogun/examples/meta/generator/translate.py", line 157, in seperatedClassDependencies	11:29
@HeikoS	sanuj: argh	11:29
sanuj	csdependencies += elementTemplate.substitute(element=x, include=self.getIncludePathForClass(x))	11:29
@HeikoS	dont paste here ;)	11:29
sanuj	File "/home/sanuj/Projects/shogun/examples/meta/generator/translate.py", line 179, in getIncludePathForClass	11:29
@HeikoS	spam	11:29
sanuj	raise Exception('Failed to obtain include path for %s' % (' or '.join(variants)))	11:29
@HeikoS	send me a gist	11:29
sanuj	Exception: Failed to obtain include path for int32_t or Cint32_t or int32_t or Cint32_t	11:29
sanuj	make[3]: *** [examples/meta/CMakeFiles/meta_examples] Error 1	11:29
sanuj	make[2]: *** [examples/meta/CMakeFiles/meta_examples.dir/all] Error 2	11:29
sanuj	make[1]: *** [doc/cookbook/CMakeFiles/cookbook.dir/rule] Error 2	11:29
sanuj	make: *** [cookbook] Error 2	11:30
sanuj	haha	11:30
sanuj	won't happen again	11:30
sanuj	do you want a gist	11:30
sanuj	this is the whole error ^^^^	11:30
sanuj	:P	11:30
@HeikoS	yes	11:30
@HeikoS	gist please	11:30
sanuj	okay	11:30
sanuj	HeikoS, https://gist.github.com/sanuj/3fbf04519f9bb7e4c8dfb196af50e8d9	11:31
sanuj	HeikoS, can't we make simple vectors and matrices in cookbook?	11:33
sanuj	like RealVector a([1,2,3])	11:33
lisitsyn	hey there	11:33
sanuj	hi lisitsyn	11:33
sanuj	HeikoS, are you fixing the error with Int right now?	11:36
@HeikoS	sanuj: looking at it	11:40
@HeikoS	sanuj: nope	11:40
sanuj	okay	11:40
@HeikoS	sanuj: if you want it, you need to add it :)	11:40
@HeikoS	possible easily	11:40
sanuj	i see	11:41
sanuj	okay	11:41
sanuj	HeikoS, `Expectation Maximization(EM) algorithm <https://en.wikipedia.org/wiki/Expectation%E2%80%93maximization_algorithm>`_ is used to train ......	11:41
sanuj	are these hyper links fine in the cookbook ^^	11:41
@HeikoS	sanuj: yes you can do that	11:47
@HeikoS	or just put them at the references below	11:47
sanuj	okay	11:47
sanuj	cool	11:47
@HeikoS	sanuj: btw can you debug the example generation error yourself?	11:47
sanuj	i will give it a try	11:47
-!- lambday [8028b10a@gateway/web/freenode/ip.128.40.177.10] has joined #shogun		11:47
-!- mode/#shogun [+o lambday] by ChanServ		11:47
@HeikoS	It comes from the fact that it is trying to find the include path for the in variable	11:47
@HeikoS	but it shouldnt do that	11:47
@HeikoS	just like for Real	11:47
@HeikoS	where it doesnt do that, and it doesnt fail	11:48
@HeikoS	need to find that check and extend it to Int	11:48
@HeikoS	lambday: jojo	11:48
@lambday	HeikoS: hey	11:48
@HeikoS	lambday: all good?	11:48
sanuj	aren't you guys sitting together this time? :P	11:48
@lambday	HeikoS: yeah..	11:48
@HeikoS	sanuj: technically yes	11:49
@HeikoS	haha	11:49
@HeikoS	but I work from remote quite a bit	11:49
@lambday	sanuj: I've stone his LAN cable so he's wfh	11:50
@HeikoS	so still at home, meeings since morning	11:50
@lambday	stolen*	11:50
@HeikoS	haha	11:50
sanuj	oh i see	11:50
sanuj	lambday, how's london :)	11:50
@lambday	sanuj: pretty ;)	11:50
@HeikoS	lambday: will finish something and then come in	11:50
@HeikoS	see you soon	11:50
@lambday	HeikoS: alright.. see you	11:50
@HeikoS	sanuj: let me know how things go with the meta example fix	11:51
sanuj	HeikoS, yes	11:51
-!- HeikoS [~heiko@host-92-0-162-192.as43234.net] has quit [Quit: Leaving.]		11:54
@lambday	sanuj: how's it going?	11:55
sanuj	lambday, made some prototypes	11:55
sanuj	lots to do	11:55
sanuj	these passport people are frying my head :P	11:55
@lambday	sanuj: hehe what happened?	11:55
sanuj	it's been 5 months but i didn't get it	11:56
sanuj	did police verification twice	11:56
sanuj	from home and college	11:56
@lambday	sanuj: ah I've heard that things in guwahati is quite slow :(	11:56
sanuj	oh	11:56
@lambday	you should have done it from bangalore.. I got mine within 10 days of applying	11:56
@lambday	that too normal, not tatkal	11:56
sanuj	oh reallly	11:56
sanuj	wow	11:57
sanuj	haha	11:57
sanuj	i hope i get it soon	11:57
sanuj	want to apply for research interns soon	11:57
@lambday	yeah don't worry.. sometimes they take time	11:57
sanuj	okay :)	11:57
@lambday	cool!.. decided not to join samsung?	11:57
sanuj	i never wanted to	11:57
sanuj	but first i need to get an intern	11:57
@lambday	sanuj: try here ;)	11:58
sanuj	talked to a few guys working there in samsung delhi office	11:58
sanuj	oh	11:58
sanuj	haha	11:58
sanuj	i'll	11:58
sanuj	;)	11:58
sanuj	what are you working on?	11:58
@lambday	sanuj: statistical testing framework... same things that heiko worked on..	11:59
sanuj	heiko for his phd?	11:59
@lambday	he did most of these things in his Masters.. I'm just learning now..	12:00
sanuj	all my friends are going into deep learning research	12:00
@lambday	:D	12:00
sanuj	oh	12:00
sanuj	cool	12:00
@lambday	hehe.. yeah ... Arthur said that deep learning was like in 2nd place in terms of the percentage of accepted papers in NIPS..	12:00
@lambday	1st was GP :P	12:00
sanuj	i see	12:01
@lambday	doesn't matter though	12:01
sanuj	lambday, you will apply for PhD next year?	12:01
@lambday	I haven't decided my topic yet.. let's see	12:01
sanuj	yeah	12:01
@lambday	sanuj: haha tough question.. I want to..	12:01
sanuj	phds are shorter in london i have heard	12:02
sanuj	4 years	12:02
@lambday	4 years	12:02
@lambday	yes	12:02
sanuj	lambday, did you visit google deepmind :P	12:02
@lambday	sanuj: not yet.. but we'll visit .. hopefully	12:03
sanuj	cool	12:03
@lambday	but fish and chips man.. awesome food :D	12:04
sanuj	haha	12:04
sanuj	i'm a vegetarian	12:04
@lambday	that's what london is famous for.. so I heard..	12:05
@lambday	haha I know ;)	12:05
sanuj	but can try for sure	12:05
sanuj	:P	12:05
@lambday	really! that's great then	12:05
sanuj	and it rains a lot?	12:05
sanuj	compared to bombay?	12:05
@lambday	in bombay it only rains for 3 months.. here it rains always I've heard..	12:05
@lambday	but not heavy rainfall like bombay	12:06
@lambday	today it's sunny yayy :D	12:06
sanuj	okay	12:06
sanuj	hah	12:06
sanuj	i did an intern in iit-b	12:06
@lambday	awesome! with whom?	12:06
sanuj	was stuck due to rain for 3 days	12:06
sanuj	oh it was first year	12:06
@lambday	haha I can imagine!	12:06
sanuj	not a prof	12:06
sanuj	ITSP	12:07
sanuj	some student organized project workshop	12:07
@lambday	I see	12:07
sanuj	we made a home automation system controlled by facebook	12:07
sanuj	:P	12:07
@lambday	haha sounds awesome!	12:07
sanuj	lambday, where are you staying?	12:09
sanuj	sharing with someone?	12:09
sanuj	and how is UCL	12:09
@lambday	sanuj: I stay pretty far.. near east ham.. staying in a paying guest with a gujarati family.. that's the cheapest option I could find.. plus I have a friend who's staying there..	12:10
@lambday	london is quite costly	12:11
sanuj	yeah london is costly	12:11
@lambday	UCL is awesome.. distributed campus.. not like iits there	12:11
sanuj	then how do you travel?	12:11
@lambday	tube	12:11
sanuj	cool	12:11
sanuj	looking forward to coming there sometime :P	12:12
@lambday	haha yeah you'll love it..	12:12
@lambday	it's an ancient city.. and looks like a prettier version of mumbai/camcutta	12:12
@lambday	calcutta*	12:13
sanuj	ohkay	12:13
@lambday	people are very polite (and indirect) :D	12:13
sanuj	haha	12:14
sanuj	lambday, weren't you getting married? :D	12:14
@lambday	sanuj: yeah.. next Jan	12:14
sanuj	i see	12:15
sanuj	2017	12:15
@lambday	yep	12:15
@lambday	anyway.. I'll let you get back to work.. and so will I (see? learned politeness already)..	12:16
@lambday	haha	12:16
@lambday	see ya	12:16
sanuj	haha	12:16
sanuj	yeah	12:16
sanuj	take care	12:16
@lambday	you too	12:16
sanuj	:)	12:17
arianepaola	hello :-)	13:19
-!- c4goldsw [c1a99ae1@gateway/web/cgi-irc/kiwiirc.com/ip.193.169.154.225] has joined #shogun		13:26
c4goldsw	Hey all, is the meeting today being held in the IRC?	13:28
-!- c4goldsw [c1a99ae1@gateway/web/cgi-irc/kiwiirc.com/ip.193.169.154.225] has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client]		13:53
-!- HeikoS [~heiko@nat-229-174.internal.eduroam.ucl.ac.uk] has joined #shogun		13:54
-!- mode/#shogun [+o HeikoS] by ChanServ		13:54
-!- leagoetz [~leagoetz@nat-231-184.internal.eduroam.ucl.ac.uk] has joined #shogun		14:17
-!- c4goldsw [c1a99ae1@gateway/web/cgi-irc/kiwiirc.com/ip.193.169.154.225] has joined #shogun		14:22
c4goldsw	HeikoS Hey, is the meeting today going to be held in the IRC?	14:22
-!- sanuj [~sanuj@117.203.1.250] has quit [Ping timeout: 276 seconds]		14:53
-!- c4goldsw [c1a99ae1@gateway/web/cgi-irc/kiwiirc.com/ip.193.169.154.225] has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client]		14:56
-!- Saurabh7_ [~Saurabh7@117.215.53.11] has joined #shogun		15:11
-!- sanuj [~sanuj@117.203.1.250] has joined #shogun		15:15
@HeikoS	yes	15:35
-!- besser82 [~besser82@fedora/besser82] has joined #shogun		15:40
-!- mode/#shogun [+o besser82] by ChanServ		15:40
@HeikoS	wiking: where is bender?	15:41
sanuj	i'm about to fix the meta examples error	15:41
sanuj	took a lot of time to track down	15:41
sanuj	but now i understand how meta examples are generated	15:42
sanuj	HeikoS, Real also finds the include path	15:42
sanuj	it's shogun/lib/common.h	15:43
@HeikoS	sanuj: great stuff	15:45
@HeikoS	why doesnt that work for the int?	15:45
sanuj	because for Int there is no path	15:45
@HeikoS	the path comes from ctags I guess	15:45
@HeikoS	is int not defined in common.h?	15:45
sanuj	let me see	15:46
sanuj	it is defined	15:46
@HeikoS	so how did you fix it?	15:46
sanuj	i'm going to fix it....wait	15:47
sanuj	:P	15:47
@HeikoS	kk	15:47
leagoetz	hi everyone	15:57
leagoetz	I see it's getting nicely busy in IRC :)	15:57
leagoetz	let's maybe wait 5 more mins, or is everyone here?	15:58
Saurabh7_	hi	15:58
Saurabh7_	oh its in 5 min	15:58
Saurabh7_	wasnt it 3 pm ?	15:59
sanuj	HeikoS, i dunno how ctags work :/	15:59
@HeikoS	sanuj: lisitsyn is your man	15:59
leagoetz	yep, 3pm	15:59
@HeikoS	3pm London time	15:59
@HeikoS	which is now	15:59
@HeikoS	leagoetz: maybe send an email to everyone saying that meeting is now	15:59
Saurabh7_	oh i see	15:59
leagoetz	sorry about the time confusion guys, my doodle was a bare bones version	16:00
leagoetz	yes will do	16:00
Saurabh7_	i was checking utc its says 2	16:00
-!- OXPHOS [9d8b131c@gateway/web/freenode/ip.157.139.19.28] has joined #shogun		16:00
-!- c4goldsw [c1a99ae1@gateway/web/cgi-irc/kiwiirc.com/ip.193.169.154.225] has joined #shogun		16:00
sanuj	Saurabh7_, it was 3 pm UTC	16:01
sanuj	the meeting i mean	16:01
sanuj	?	16:01
sanuj	anyways i'm here	16:01
Saurabh7_	yeah	16:01
Saurabh7_	ye me too :)	16:01
@HeikoS	lets see	16:01
leagoetz	given that EVERYONE seems to be here it doesn't really matter, does it?	16:01
leagoetz	:)	16:01
sanuj	nope	16:01
lisitsyn	hey	16:01
sanuj	lisitsyn, ping	16:01
sanuj	oh hi	16:01
lisitsyn	yeah just in time	16:01
@HeikoS	wiking, arianepaola, lisitsyn, OXPHOS are you there?	16:01
@HeikoS	lisitsyn: hi	16:01
c4goldsw	Hey everyone.	16:01
sanuj	lisitsyn, i need your help with c tags	16:01
@HeikoS	rcurtin: you around by any chance?	16:02
OXPHOS	yep	16:02
@HeikoS	c4goldsw: hi there	16:02
@HeikoS	you want to join?	16:02
c4goldsw	That's why I'm here :)	16:02
@HeikoS	cool	16:02
@HeikoS	wiking is the last missing	16:02
@HeikoS	but his timezone is tricky	16:02
@HeikoS	leagoetz: did he say he can make it on the doodle?	16:02
leagoetz	kk do you think we should start without him	16:02
leagoetz	he said he could make it	16:02
leagoetz	but maybe he thinks it's in an hour :~	16:03
@HeikoS	well his account is online, so he can read logs	16:03
@HeikoS	lets start	16:03
leagoetz	OK	16:03
leagoetz	Hi everyone and thanks for joining this kick-off meeting. it’ll be a short meeting. Students and mentors can	16:03
leagoetz	take the opportunity to hang out afterwards, discuss, etc	16:03
leagoetz	GSoC 2016 has officially begun and we’re very excited!	16:03
leagoetz	This meeting will be a short reminder of what needs to be done during the summer, what we expect of you and to start the discussion.	16:04
leagoetz	btw, I will put up the minutes of this meeting on the Wiki for future reference (special request by sanju: ^^)	16:04
sanuj	sanuj*	16:04
leagoetz	yes, sorry	16:04
leagoetz	(and do ask questions any time if you have them).	16:04
@HeikoS	"the minutes"?	16:04
leagoetz	or make comments	16:04
leagoetz	and correct me	16:04
leagoetz	yes, that's what it's called	16:04
leagoetz	list of things that have been said	16:04
@HeikoS	I see	16:04
leagoetz	1. short reminder of our Expectations of GSoC during the next weeks:	16:05
@HeikoS	also we can put the chat log link	16:05
leagoetz	(was my plan)	16:05
leagoetz	you’ll be members of our Shogun community, so don't just code alone in the dark:	16:05
leagoetz	instead hang out in IRC	16:05
leagoetz	share the experience with the other students :)	16:05
leagoetz	keep in touch with your mentor, and ..	16:05
leagoetz	the other shogun members	16:05
leagoetz	As part of this, please send a very brief email to the mailing list each Monday	16:05
leagoetz	saying what you’ve been up to in the previous week and what are your plans for the coming week.	16:05
leagoetz	Everyone will be curious to see how it’s going.	16:05
leagoetz	Remember: GSoC is full-time - we expect you to fully concentrate on Shogun this summer. This way, you will also get out the most	16:06
leagoetz	We expect you to always be visible in IRC during GSoC, and try to talk to your mentor at least 2-3 times a week.	16:06
leagoetz	Finally, and very important: we expect everyone to enjoy this experience -- we did so in the past :)	16:06
leagoetz	ok done with monologue part 1. any questions?	16:06
leagoetz	:)	16:06
@HeikoS	maybe a comment:	16:07
@HeikoS	usually, if everyone hangs out in IRC, everyone has more fun	16:07
@HeikoS	experience tells	16:07
sanuj	yes	16:07
@HeikoS	especially if lisitsyn is around	16:07
leagoetz	yes	16:07
leagoetz	haha	16:07
sanuj	haha	16:07
lisitsyn	sorry	16:07
leagoetz	ok moving on	16:07
OXPHOS	lol	16:07
lisitsyn	:D	16:07
@lambday	lisitsyn: don't be :P	16:07
@HeikoS	sorry? :D	16:07
leagoetz	2. Plans for the first weeks	16:08
lisitsyn	HeikoS: that's my default answer to any highlight of my nickname	16:08
lisitsyn	I thought something is wrong :D	16:08
Saurabh7_	haha	16:08
leagoetz	Can each student please describe briefly what you want to do in the first 2 weeks.	16:08
leagoetz	briefly, yet concretely. Just let us know what you are currently doing,	16:08
leagoetz	what the issues are, and when you plan to finish it.	16:08
leagoetz	Sanuj, do you want to start?	16:08
sanuj	okay	16:08
sanuj	In my first week, main aim is to get shogun support tags and parameter framework	16:09
sanuj	so that OXPHOS can work on serialization with the new framework	16:09
sanuj	and cookbooks etc	16:09
sanuj	done	16:09
leagoetz	Ok, then OXPHOS:	16:09
OXPHOS	I plan to work on the prototype of serialization - with models in SGVector and SGObject.	16:10
OXPHOS	Not sure whether there's a conflict with sanuj but doubt I'll be that fast	16:10
leagoetz	in any case you two should keep talking	16:11
OXPHOS	and cookbooks. and maybe linalg design. done.	16:11
OXPHOS	sure	16:11
@HeikoS	about cookbooks: we expect to see all sorts of issues there as this is the first time we really stress test the system scaling up. sanuj already found a bug, and there will be more. Consider fixing them as part of your GSoC experience ;)	16:11
leagoetz	arianepaola:	16:11
leagoetz	next up	16:11
leagoetz	hm, are you here arianepaola?	16:12
leagoetz	Ok, let's have Saurabh7: then	16:12
Saurabh7_	Ok right now I am working on making crossvalidation run in parallel, i.e thread safe features and subsets	16:13
Saurabh7_	wiht that next week i can dive into an algorithm like random foresta nd use the thread safety there too	16:13
Saurabh7_	and cookbooks ocfc	16:13
Saurabh7_	done	16:13
leagoetz	happy to hear that you're all thinking about the cookbook, because my next point was ...	16:14
leagoetz	3. Cookbook patches.	16:14
leagoetz	As mentioned often enough ;P, we expect you to contribute 2-3 patches to our cookbook.	16:14
leagoetz	(which after your GSoC contributions will be awesome I think btw)	16:14
leagoetz	disclaimer	16:14
leagoetz	As some of you have experienced already, there are some problems with this at the moment (we are stress-testing things).	16:14
leagoetz	But please send the PRs nevertheless.	16:14
leagoetz	and finally ...	16:15
@HeikoS	I suggest to do these in a sep. feature branch each (on your side), so that we can manage multiple of them at once	16:15
@HeikoS	(it might take time to resolve some of the problems)	16:15
@HeikoS	And please:	16:15
@HeikoS	chose the most important algorithms first	16:15
@HeikoS	before adding cookbooks on single kernels or similar	16:15
leagoetz	oh and make sure you're not two of you working on the same	16:16
@HeikoS	shall we maybe have a list or so	16:16
@HeikoS	of work in progress	16:16
leagoetz	but if you're interacting should be no problem	16:16
@HeikoS	google doc	16:16
leagoetz	yes	16:16
@HeikoS	or just leave it to you guys to sort that out?	16:16
sanuj	yes that would be helpful	16:16
@HeikoS	maybe you can ask each other	16:16
leagoetz	or just write that in your Monday email the one's you plan to tackle	16:16
@HeikoS	leagoetz: +1	16:16
sanuj	cool	16:16
leagoetz	yes, sort it out in IRC, then write which one's you're resposible for in the coming week	16:16
leagoetz	Ok great	16:17
leagoetz	I think Heiko and Viktor want to say something about the release before the meeting is finished	16:17
leagoetz	HeikoS:	16:17
leagoetz	?	16:17
@HeikoS	yeah	16:17
leagoetz	wiking:	16:17
leagoetz	?	16:17
@HeikoS	as wiking is not here	16:17
leagoetz	oh right	16:17
@HeikoS	so there is the list of issues on the release:	16:17
@HeikoS	let me get the link	16:18
@HeikoS	https://github.com/shogun-toolbox/shogun/issues?q=is%3Aopen+is%3Aissue+milestone%3A%22Shogun+4.2.0+-+Takechi+Hanpeita%22	16:18
@HeikoS	some of you already fixed some bugs, warnings etc	16:18
@HeikoS	thanks a lot for that	16:18
@HeikoS	some are still open, and need work	16:18
@HeikoS	if you know how to solve something, do it	16:18
@HeikoS	very much appreciated	16:18
-!- sonne\|work [~sonnenbu@ip5b414dfc.dynamic.kabel-deutschland.de] has joined #shogun		16:19
@HeikoS	arianepaola: in particular you could help e.g. with the eigen3 problem	16:19
@HeikoS	apart from that we are almost ready to release	16:19
leagoetz	wohooo	16:19
@HeikoS	after the release	16:19
@HeikoS	you can use c++11 in your codes	16:19
@HeikoS	which will be interesting for sanuj and OXPHOS	16:19
sonne\|work	hey everyone...	16:19
@HeikoS	done	16:19
@HeikoS	sonne\|work: hi!	16:19
leagoetz	ok, does anyone have any questions?	16:20
leagoetz	(if something comes to you later, send me an email or find me in IRC)	16:20
c4goldsw	A quick one: what are cookbooks exactly?	16:20
c4goldsw	Examples of how to use certain parts of features of Shogun?	16:20
@HeikoS	http://shogun.ml/cookbook/latest/	16:20
@HeikoS	c4goldsw: ^	16:20
c4goldsw	Thanks.	16:21
leagoetz	more questions?	16:21
OXPHOS	What exactly is the time in UTC now?	16:21
@HeikoS	c4goldsw: it is three things: examples, integration tests, and documentation for all modular targets	16:21
@HeikoS	http://lmgtfy.com/?q=current+time+UTC	16:21
@HeikoS	;)	16:21
leagoetz	2:20	16:21
leagoetz	I messed up	16:21
leagoetz	oh well	16:21
@HeikoS	haha	16:22
OXPHOS	Aha yes I googled and confused	16:22
@HeikoS	well almost everyone made it	16:22
OXPHOS	thought the world break down	16:22
@HeikoS	wiking and arianepaola can talk with each other I guess	16:22
leagoetz	yes, I'll write to them	16:22
leagoetz	OK	16:22
leagoetz	quick summary and then the meeting is over:	16:22
leagoetz	1. make sure you meet the expectations:	16:22
leagoetz	full-time,	16:22
leagoetz	hang out in IRC &	16:22
leagoetz	keep in touch with your mentor,	16:22
leagoetz	update us on your progress,	16:23
leagoetz	contribute to cookbook	16:23
leagoetz	2. your plan for the first 2 weeks …	16:23
leagoetz	… do it :)	16:23
leagoetz	3. don’t forget the cookbook patches	16:23
leagoetz	4. have a great time!	16:23
leagoetz	That's it	16:23
leagoetz	Take care everyone and let me know if you have any questions/issues/comments!	16:23
@HeikoS	thanks leagoetz!	16:23
sanuj	great	16:23
OXPHOS	thanks!	16:23
lisitsyn	thanks	16:23
@HeikoS	lambday and OXPHOS: maybe the two of you can discuss a bit the linalg design?	16:24
@lambday	HeikoS: yeah..	16:24
@lambday	OXPHOS: you have a min?	16:24
OXPHOS	lambday: sure. I have something to show. lemme dump it in the gist?	16:24
@HeikoS	sanuj: so hows the patch going?	16:24
@lambday	OXPHOS: alright..	16:24
OXPHOS	lambday: or you wanna say sth. first?	16:25
@HeikoS	Saurabh7_: any news on the xvalidation?	16:25
@lambday	OXPHOS: let me check your gist	16:25
@lambday	maybe you have something better ;)	16:25
sanuj	OXPHOS, i will update you about my tags/parameter prototype later today...didn't want to confuse you earlier as it was not complete :)	16:25
sanuj	HeikoS, lisitsyn about the meta example fix	16:26
OXPHOS	lambday: https://gist.github.com/OXPHOS/16c7a4d31cb4f5553fd8e7bc70ca5114	16:26
OXPHOS	sanuj: thx!	16:26
Saurabh7_	HeikoS, working on it, but found that things a re a bit different for this locked training we have for say kernel machines	16:26
@HeikoS	Saurabh7_: that is if kernel matrix is precomputed	16:27
@HeikoS	Saurabh7_: only want to do that once in xvalidation	16:27
@HeikoS	so there	16:27
Saurabh7_	HeikoS, hm	16:27
@HeikoS	it is not just the features which get a different subset every time	16:28
@HeikoS	but also the kernel	16:28
sanuj	HeikoS, lisitsyn there is nothing to include for int32_t from tags so it breaks	16:28
Saurabh7_	yeah kernel have row as well as column subsets	16:28
@HeikoS	But I think, the first version should ignore the locked case	16:28
Saurabh7_	yup i am ignoring it for now	16:28
Saurabh7_	for now	16:28
lisitsyn	sanuj: ??	16:28
sanuj	lisitsyn, have you seen this build failure https://travis-ci.org/shogun-toolbox/shogun/builds/132213314	16:28
lisitsyn	didn't get it	16:28
Saurabh7_	HeikoS, so i will try with kernel svm and relevant features and unlcoked training ?	16:29
Saurabh7_	to get it going	16:29
lisitsyn	sanuj: the errors look unrelated?	16:29
-!- HeikoS [~heiko@nat-229-174.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.]		16:29
sanuj	lisitsyn, i found the reason for the error	16:30
-!- leagoetz [~leagoetz@nat-231-184.internal.eduroam.ucl.ac.uk] has quit []		16:30
sanuj	lisitsyn, let me link you to the code	16:30
lisitsyn	ok	16:30
@lambday	OXPHOS: so the idea here is, to have a "dot" method in the class itself, and then in linalg we use that one	16:30
@lambday	OXPHOS: did I get this right?	16:30
-!- HeikoS [~heiko@nat-229-174.internal.eduroam.ucl.ac.uk] has joined #shogun		16:31
-!- mode/#shogun [+o HeikoS] by ChanServ		16:31
OXPHOS	lambday: aha it doesn't have to be dot only. All methods can be dumped in. If this is what you meant	16:31
sanuj	lisitsyn, https://github.com/shogun-toolbox/shogun/blob/develop/examples/meta/generator/translate.py#L179	16:31
@lambday	OXPHOS: yeah that was my point.. but it kinda defeats the purpose of having the methods in linalg then, isn't it?	16:32
@lambday	OXPHOS: I think what ideally we want is to just have linalg::foo(....) and get rid of all the class methods..	16:33
sanuj	lisitsyn, the error in the build was due to using Real instead of Int as Java and csharp have problem with type conversion	16:33
@lambday	OXPHOS: does it sound reasonable?	16:33
OXPHOS	lambday: you mean like linalg::foo(dot, a, b)?	16:33
sanuj	lisitsyn, when i tried using Int the cookbook gave this error: https://gist.github.com/sanuj/3fbf04519f9bb7e4c8dfb196af50e8d9	16:34
@lambday	OXPHOS: more like, linalg::dot(a, b).. and similarly, other methods	16:34
lisitsyn	sanuj: oh ok	16:34
lisitsyn	so it's ctags issue?	16:34
sanuj	lisitsyn, yep	16:34
OXPHOS	lambday: I think this is what I showed. dot is in linalg namespace. Or I was wrong?	16:35
sanuj	lisitsyn, in case of Real float64_t includes shogun/lib/common.h	16:35
sanuj	lisitsyn, but in case of Int i.e. int32_t there is nothing to include so throws the exception	16:36
@lambday	OXPHOS: yeah.. that's fine.. but what I meant is that, this solution requires that we have a dot method in the vector class itself as a member method	16:36
@lambday	that's something we want to avoid	16:36
OXPHOS	lambday: aa sry I get it	16:36
@lambday	OXPHOS: no worries..	16:37
@lambday	so..	16:37
sanuj	lisitsyn, i dunno how ctags work	16:37
OXPHOS	it is in base class. not exactly vector (but go ahead)	16:37
lisitsyn	sanuj: ok got it	16:37
@lambday	let's think about how we're going to use the API	16:37
@lambday	OXPHOS: yeah but in the subclasses also you must have to overload the virtual dot methods to make it work	16:37
sanuj	lisitsyn, what's the fix :)	16:37
@lambday	or did I get this wrong?	16:38
@lambday	override*	16:38
lisitsyn	sanuj: let me check	16:38
OXPHOS	lambday: yes. but the subclasses are like CPU_end and GPU_end? I don't feel like the calculation will fall back to SGVector..But I might be wrong	16:39
@HeikoS	Saurabh7_: yes do that first	16:39
@lambday	OXPHOS: https://gist.github.com/OXPHOS/16c7a4d31cb4f5553fd8e7bc70ca5114#file-dot-cpp-L9 this call here will always call the Base<T>::dot, isn't it	16:40
@lambday	because we're not using the pointers	16:40
-!- c4goldsw [c1a99ae1@gateway/web/cgi-irc/kiwiirc.com/ip.193.169.154.225] has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client]		16:41
OXPHOS	lambday: yes	16:41
@lambday	OXPHOS: okay.. so maybe I am not getting it properly.. so let me check if I get this correctly...	16:41
@lambday	you have a Base.. and then you extend this to create classes CPU vec and GPU vec	16:41
@lambday	and these override the dot interface (similarly other methods)	16:41
@lambday	and then you give your SGvector to the factory... get an instance of CPU and GPU vectors based on the available backend..	16:42
@lambday	and then want to call dot on those	16:42
@lambday	that's the idea	16:42
@lambday	am I right>?	16:42
OXPHOS	yes. send SGVector in. then get sth out for the further calculation	16:43
@lambday	OXPHOS: yeah..	16:43
@lambday	OXPHOS: so, to make this work, I think we need the factory to return a ptr so we are able to make the virtual call	16:43
@lambday	otherwise, it always calls the base::dot	16:44
lisitsyn	sanuj: lets just add that to common.h	16:44
lisitsyn	should be ok	16:44
OXPHOS	lambday: hmm. we can do it right?	16:45
@lambday	OXPHOS: yeah.. but let's think a bit.. the linalg interface in this case always works with SGvector	16:45
@lambday	OXPHOS: am I right?	16:46
OXPHOS	lambday: yes	16:46
-!- besser82 [~besser82@fedora/besser82] has quit [Ping timeout: 258 seconds]		16:46
OXPHOS	lambday: you're thinking having a separate GPUVector?	16:46
@lambday	OXPHOS: so, in each linalg::dot call, we MUST call the factory to get the appropriate vector type	16:47
@lambday	nono	16:47
OXPHOS	yes	16:47
@lambday	so in that case, we end up doing cpu <--> gpu transfer in each linalg::dot call, where we simply could have used	16:47
@lambday	auto a = factory(v)	16:47
@lambday	auto b = factory(u)	16:47
@lambday	a->dot(b)	16:48
@lambday	so it's not very good, no?	16:48
OXPHOS	very clumsy : /	16:48
@lambday	OXPHOS: yeah	16:48
@lambday	OXPHOS: so let's think for a sec about how one would actually use this	16:49
@lambday	let's write a small use-case	16:49
@lambday	say, this is the code for mycoolalgo	16:49
@lambday	inside that, I do	16:49
@lambday	auto a = factory(u) // u is my input vector..	16:49
@lambday	so any cpu <--> gpu thing happens here	16:49
@lambday	then I want to use just this one..	16:50
@lambday	linalg::foobar(a)	16:50
@lambday	linalg::barbar(a)...	16:50
@lambday	so on	16:50
@lambday	and then finally.. i want to transfer the result back to cpu	16:50
@lambday	so the linalg interfaces in this case should directly work with the Base<T> ptr type..	16:51
sanuj	lisitsyn, what do you want to add to common.h?	16:51
lisitsyn	sanuj: the missing types	16:51
lisitsyn	int64_t int32_t	16:51
@lambday	OXPHOS: does it sound reasonable?	16:51
sanuj	its there	16:51
lisitsyn	nope	16:52
lisitsyn	float32_t is defined	16:52
lisitsyn	but not int32_t	16:52
OXPHOS	lambday: sounds good. and I feel like it is similar to what I have..?	16:52
sanuj	there is typedef int32_t index_t;	16:52
@lambday	OXPHOS: yeah, except, it's all runtime now	16:52
@lambday	so we always work with base<T>* now	16:53
@lambday	instead of stack objects	16:53
OXPHOS	lambday: bascially ask the users to do the factory() step?	16:53
sanuj	lisitsyn, ^	16:53
@lambday	OXPHOS: nono.. this is what will happen inside shogun code..	16:54
OXPHOS	lambday: I guess I never get the 'runtime' and 'compile' difference	16:54
@lambday	factory thing will be sort of a plugin thing.. that (based on the installation) will return things..	16:55
lisitsyn	sanuj: it doesn't define int32_t	16:55
@lambday	OXPHOS: hehe don't worry..	16:55
@lambday	OXPHOS: what is important for now, is that we keep the linalg interface ready that works with Base ptrs	16:56
-!- besser82 [~besser82@fedora/besser82] has joined #shogun		16:56
-!- mode/#shogun [+o besser82] by ChanServ		16:56
@lambday	we'll worry about the factory later onwards	16:56
sanuj	lisitsyn, okay, how do i define int32_t...there are typedefs of others like float64_t etc	16:56
lisitsyn	sanuj: typedef int int32_t;	16:56
@lambday	but imagine that by some magic, you have an instance of a GPUVector ptr/CPUVector ptr..	16:57
sanuj	lisitsyn, there is already typedef int machine_int_t;	16:57
@lambday	now, we need to make our dot work with that	16:57
lisitsyn	sanuj: what has it to do with int32_t? :D	16:57
sanuj	lisitsyn, 2 typedefs with same name are fine?	16:57
@lambday	OXPHOS: if you have a min, then you can have a look at the LinearOperator interface..	16:58
OXPHOS	lambday: step1: transfer (done in factory) step2: make linalg work with transferred datatype (Base*)?	16:58
OXPHOS	or reverse	16:58
sanuj	lisitsyn, will you be around after an hour	16:58
sanuj	need to go have dinner	16:58
@lambday	OXPHOS: yes that's how we'd use it.. but we can develop step2 first and make step1 work later	16:58
@lambday	all we need is the classes	16:59
@lambday	OXPHOS: but is it clear now that if we don't use Base*, then it's no more runtime decided?	17:00
@lambday	because we'll be using runtime polymorphism to make things work.. which, without pointers, won't really work I'm afraid	17:01
OXPHOS	lambday: !!! finally..so dot(Base* a, Base* b), but not a.dot(b)?	17:01
@lambday	OXPHOS: I'm afraid so :)	17:01
OXPHOS	lambday: yes	17:01
@lambday	OXPHOS: great!	17:02
OXPHOS	lambday: So I'll 1. code base and subclasses 2. refactor dot	17:03
@HeikoS	OXPHOS: probably good to summarise things in an email, send it back to lambday, and cc Viktor	17:03
@lambday	OXPHOS: maybe first draft the use-case.. and walk through the details?	17:03
@lambday	OXPHOS: how the classes should look.. etc	17:04
OXPHOS	HeikoS lambday: sure	17:04
@lambday	OXPHOS: we don't want that we put so much effort in coding up something and then it's not really a good idea	17:04
@lambday	OXPHOS: so, inside the linalg dot, can we make it work, without having to have a dot method in the class itself?	17:05
@lambday	let's think!	17:05
@lambday	(the tricky part is th first method :D .. once a single method is added, rest is simple :)	17:05
arianepaola	hmmm, just saw it is the other UTC UTC :-(	17:05
@HeikoS	arianepaola: hi there	17:06
@HeikoS	no worries	17:06
@HeikoS	you can read th elogs	17:06
@HeikoS	know where they are?	17:06
@HeikoS	arianepaola: I guess what we need from you is some status for the coming two weeks, what will you work on, how long will it take? etc	17:07
@HeikoS	apart from that, helping with the release issues would be useful	17:07
@HeikoS	and of course a cookbook page	17:07
arianepaola	thank for the update HeikoS	17:07
OXPHOS	lambday: very violent: Base now is just a datatype, either Eigen or GPU. Can we just ask dot to take up either Eigen or GPU parameters?	17:08
@lambday	OXPHOS: please feel free to ask about anything that's not clear.. don't worry if what i am saying doesn't make much sense	17:08
@HeikoS	arianepaola: is that all good?	17:08
OXPHOS	lambday: yea thanks so much	17:08
@HeikoS	arianepaola: the other thing: try to be in IRC during your work hours	17:08
@HeikoS	so that we can always catch up when we have a minute	17:08
@HeikoS	and its more fun with the others	17:08
@lambday	OXPHOS: what do you mean? Base is out native type, right?	17:09
@HeikoS	arianepaola: so what will you look at first?	17:09
OXPHOS	lambday: ideally base* = factory(SGVector). In factory we decide which backend(CPU, GPU) we want to use	17:09
@lambday	OXPHOS: exactly!	17:10
OXPHOS	lambday: so the base* will be the data either on CPU or GPU	17:10
@lambday	OXPHOS: exactly!	17:10
@lambday	OXPHOS: let's say, it will be either CPUVector<T> <-- Base<T> or GPUVector<T> <--- Base<T>	17:11
arianepaola	HeikoS: I am working on the Debian packages	17:11
OXPHOS	so dot(base * a, b) is either dot(EigenVector a, b) or dot (GPUVector a, b)	17:11
OXPHOS	yes	17:11
arianepaola	looking at the LDATest and FLDATest that are failing	17:11
@lambday	yeah.. but we have the pointers..	17:11
@HeikoS	arianepaola: lets talk pm	17:11
@HeikoS	arianepaola: sent you a message	17:11
@lambday	OXPHOS: so in both cases, the call will go to dot(base* a, base* b)	17:12
OXPHOS	lambday: so with pointers we cannot overload(?) the dot	17:12
@lambday	OXPHOS: nope.. overload is not an option here	17:12
@lambday	OXPHOS: we want it to be decided in the runtime	17:12
@lambday	OXPHOS: here is an idea	17:13
@lambday	typecast :)	17:13
OXPHOS	didn't get it : (	17:14
@lambday	OXPHOS: let's say, we do something like this in the dot method..	17:14
@lambday	auto casted_a = dynamic_cast<CPUVector<T>>(a);	17:15
@lambday	if "a" is actually an underlying CPUVector type, then we get a valid ptr in casted_a	17:15
@lambday	which is of CPUVector ptr type	17:15
@lambday	if it is not of CPUVector type (i.e. GPUvector maybe), then we get nullptr	17:16
@lambday	so let's assume that casted_a is not null	17:16
@lambday	so now we know how to locate the actual data ptr inside it.. we can just call eigen	17:16
@lambday	using those ptrs and make the dot work	17:16
@lambday	that's the if (casted_a != nullptr) part	17:17
OXPHOS	fancy	17:17
@lambday	otherwise, we try to cast it to GPUVector.. check if that's not null.. and then call viennacl to make that work	17:17
@lambday	that's the else part	17:17
@lambday	OXPHOS: that's one way to do it	17:18
@lambday	OXPHOS: maybe you have better ideas?	17:18
@lambday	I see 2 points.. (1) this doesn't require that we have the "dot" methods inside the classes itself.. dunno whether that's a good idea or bad (2) it will be a bit slow but that's the price we pay for being runtime-decided	17:19
@lambday	lisitsyn: yo	17:19
@lambday	lisitsyn: don't say worry :D	17:19
@lambday	sorry*	17:19
@lambday	lisitsyn: you had an idea to bypass dynamic casting, right?	17:19
OXPHOS	lambday: this one looks so cool.. or can we just have a flag in subclasses? like CPUVector.onGPU() = false	17:19
OXPHOS	const	17:20
@lambday	OXPHOS: yeah that too.. but you have to cast it anyway to access the underlying data structures	17:20
@lambday	with just base ptr, you don't have access to that	17:20
OXPHOS	lambday: ahh right	17:20
lisitsyn	lambday: wazup?	17:21
@lambday	lisitsyn: yoyo	17:21
lisitsyn	what do you mean by bypassing dyn casting?	17:21
@lambday	lisitsyn: so the problem we have here is that we have a base ptr, and the interface we design works with that..	17:22
@lambday	but inside the method we want to do something very specific	17:22
@lambday	so (1) either we have to cast or (2) write virtual calls to do that stuff	17:22
@lambday	lisitsyn: which one would be faster?	17:22
lisitsyn	is it about linalg?	17:22
@lambday	lisitsyn: yeah.. we're designing it all runtime now	17:23
lisitsyn	I wouldn't care about virtual calls in linalg then	17:23
lisitsyn	:)	17:23
@lambday	lisitsyn: dynamic casting is slow?	17:23
lisitsyn	I mean your virtual call is going to start	17:23
lisitsyn	e.g. multiplying 10000 numbers	17:24
lisitsyn	why bother about its runtime :D	17:24
lisitsyn	lambday: dynamic casting should be slower than virtual calls	17:24
@lambday	lisitsyn: I see	17:24
@lambday	that's what I wanted to know	17:24
lisitsyn	okie	17:24
@lambday	in that case.. the best idea would be to simply override methods..	17:25
@lambday	as class members	17:25
@lambday	so OXPHOS	17:25
@lambday	the use case then becomes:	17:25
@lambday	Base<T>* vec = Factory(a);	17:25
@lambday	Base<T>* vec2 = Factory(b);	17:25
@lambday	vec->dot(2)	17:25
@lambday	instead of using non-member linalg methods, this would be faster I think	17:26
lisitsyn	please no pointers	17:26
lisitsyn	:D	17:26
lisitsyn	please	17:26
@lambday	lisitsyn: I see what you mean.. but without ptrs you can't make it really runtime dependent :D	17:26
OXPHOS	so dot will be in Base class?	17:27
@lambday	lisitsyn: other-wise it's compile-time.. fast, true, but two different binaries	17:27
lisitsyn	lambday: no just use some wrappers	17:27
@lambday	OXPHOS: yes.. and so is all the methods	17:27
lisitsyn	I mean no raw pointers in the code	17:27
@lambday	lisitsyn: oh yeah sure that's no biggie	17:27
@lambday	she will use unique-ptrs :D	17:28
OXPHOS	np :D	17:28
lisitsyn	yes	17:28
lisitsyn	call it Shared	17:28
lisitsyn	:)	17:28
@lambday	lisitsyn: why shared :( that's too much :P	17:29
@lambday	I don't like to share stuffs... I eat everything :D	17:29
lisitsyn	Unique!	17:29
@lambday	Unique is good.. lightweight :D	17:29
@lambday	OXPHOS: did you get a picture now?	17:30
@lambday	OXPHOS: just check LinearOperator interface once	17:30
@lambday	OXPHOS: basically that's how this stuff would work	17:30
@lambday	and the GPUMatrix will actually be a part of LinearOperator interface itself	17:30
@lambday	BUT	17:31
@lambday	maybe write a mail with the design and check it once with wiking	17:31
OXPHOS	lambday: sure. I think so..just basically what I have for now except the pointer part?	17:31
@lambday	OXPHOS: check that we have this CDenseMatrixOperator... imagine that we have a similar CGPUMatrixOperator or so..	17:31
OXPHOS	but still the same vector will be factory-ed multiple times	17:32
@lambday	OXPHOS: no.. we won't have those linalg::dot(...) calls anymore at all!	17:32
@lambday	in code, this directly calls the methods..	17:32
@lambday	OXPHOS: just check the above use-case i wrote..	17:33
@lambday	basically that's how we'd use this in shogun code	17:33
@lambday	wherever we need to use dynamic linalg backend	17:33
@lambday	OXPHOS: does it make sense?	17:34
OXPHOS	lambday: I have the same thing in my gist..?	17:34
@lambday	we already have the dynamic polymorphism going on.. why would we need it to go through another method call?	17:35
@lambday	OXPHOS: yes.. except the top namespace linalg {.... } part.. that won't be there anymore	17:35
@lambday	OXPHOS: https://gist.github.com/OXPHOS/16c7a4d31cb4f5553fd8e7bc70ca5114#file-dot-cpp-L5 this should go	17:35
arianepaola	HeikoS: The nightly Fedora rpm is marked for 4.2.0 https://github.com/shogun-toolbox/shogun/issues/3131	17:36
@HeikoS	arianepaola: yep, thats what I meant with helping for release	17:36
@HeikoS	arianepaola: the pip install is as well	17:36
OXPHOS	lambday: ahh so everything is base. sorry to be torturing you	17:36
arianepaola	HeikoS: Integrating the snippet that I posted should build it on buildbot. The only issue is to get HEAD from git instead of a tag	17:36
@lambday	OXPHOS: don't worry about that.. it's important that we get this done right :D	17:37
@lambday	OXPHOS: what do you mean that everything is base?	17:37
@HeikoS	arianepaola: maybe that is a good thing to start with then?	17:37
@lambday	in base, we'll have all the interfaces... for dot and all kinds of other shits..	17:37
@lambday	then subclasses override those methods	17:37
@lambday	and then while using it, we query the factory to give us a type...	17:37
arianepaola	HeikoS: Tried to ask besser82 about it	17:38
@lambday	we make the call..	17:38
arianepaola	besser82: ping	17:38
@HeikoS	arianepaola: it would be good to discuss this with wiking as well	17:38
@lambday	OXPHOS: if you check LinearOperator interface, there also we do similar thing.. we have a "apply" method...	17:39
@lambday	that is pure virtual..	17:39
arianepaola	HeikoS: wiking mentioned to ping besser82 about the command to download the git HEAD from rpmbuild	17:39
@lambday	OXPHOS: then in subclasses (CDenseMatrixOperator and CSparseMatrixOperator) we implement those..	17:39
@HeikoS	arianepaola: ok	17:39
arianepaola	that's the part that is blocking	17:39
@lambday	OXPHOS: basically the same idea here!	17:39
@HeikoS	lets see what he sais	17:39
@HeikoS	besser82: ^	17:39
OXPHOS	lambday: lemme see	17:40
@HeikoS	arianepaola: doesnt google or the fedora IRC should asnwer that?	17:40
arianepaola	not google, but have to try fedora IRC	17:41
OXPHOS	lambday: I got the `apply` part. So for the factory - it should be OUTSIDE base class - right?	17:41
@HeikoS	lisitsyn: jo	17:48
lisitsyn	HeikoS: sup	17:49
@HeikoS	what if they static cast	17:49
@HeikoS	in linalg::dot(a,b)	17:49
@HeikoS	check the flag	17:49
lisitsyn	static?	17:49
@HeikoS	a.lives_on_gpu()	17:49
@HeikoS	and then static cast it to subclass	17:49
lisitsyn	I am sorry I wasn't following - too much for me	17:49
lisitsyn	:D	17:49
lisitsyn	what's happening?	17:50
@HeikoS	lisitsyn: I dont know	17:50
@HeikoS	I thought you knew	17:50
lisitsyn	meh	17:50
lisitsyn	then	17:50
lisitsyn	what problem are you trying to solve?	17:50
@lambday	lisitsyn: let me clarify a bit	17:50
@lambday	OXPHOS: just give me a sec	17:50
@lambday	lisitsyn: so, remember the dynamic cast vs virtual calls thing I asked you about?	17:51
@lambday	:P	17:51
lisitsyn	yes	17:51
@lambday	lisitsyn: so HeikoS says that, maybe we can avoid dynamic casting altogether if we rely on flags and static cast	17:51
@lambday	lisitsyn: would that one be faster than virtual calls?	17:51
lisitsyn	just make it dynamic	17:52
lisitsyn	you're going to run a heavy operation	17:52
@HeikoS	lisitsyn: is that always true?	17:52
@HeikoS	what about low dim dot product?	17:52
lisitsyn	like 3-4-5?	17:52
@HeikoS	like the issue Saurabh7_ discovered in kmeans	17:52
@HeikoS	where in low dimenions, the overhead of pass-by-value of SGVector had a severe impact	17:53
lisitsyn	yes but that's copy?	17:53
@HeikoS	of SGVector	17:53
lisitsyn	virtual call is much faster	17:53
@HeikoS	not of data	17:53
lisitsyn	of sgvector?	17:53
lisitsyn	any reference?	17:53
@HeikoS	no	17:53
@HeikoS	:D	17:53
@HeikoS	just saying	17:53
@HeikoS	dot products can be cheap	17:53
@HeikoS	so not sure about the dynamic cast	17:54
lisitsyn	no dynamic cast but virtual call	17:54
@HeikoS	sure	17:54
@HeikoS	but cant we check a flag and do a static cast?	17:54
lisitsyn	static cast is free	17:55
@lambday	lisitsyn: I think the idea is to avoid having member methods..	17:55
@HeikoS	yes, thats what I meant	17:55
lisitsyn	what flag?	17:55
@HeikoS	a.lives_on_gpu()	17:55
@HeikoS	if (a.is_on_gpu()) static_cast_a_as_gpu_vec	17:55
@HeikoS	;	17:56
lisitsyn	why to do that?	17:56
@HeikoS	to access the gpu data structure inside linalg::foo	17:57
@HeikoS	lisitsyn: see why?	17:58
lisitsyn	ok	17:58
@HeikoS	so we dont have to do	17:58
@HeikoS	a.dot(b)	17:58
@HeikoS	but can do	17:58
@HeikoS	linalg::dot(amb)	17:59
@HeikoS	dot(a,b)	17:59
lisitsyn	HeikoS: well in this case this seems reasonable enough	17:59
lisitsyn	but I am not a fan of avoiding dynamic stuff at all	17:59
@HeikoS	then we can do what wiking suggested, but wihtout having ViennaCL as a hard dependency	18:01
@lambday	lisitsyn: https://gist.github.com/OXPHOS/a9b79a995b75a0972811f634e80ff632	18:01
@HeikoS	see down there	18:01
@lambday	lisitsyn: just have a look at wiking's comment	18:01
arianepaola	Maybe this can help: http://www.stroustrup.com/fast_dynamic_casting.pdf - "We describe a scheme for implementing dynamic casts suitable for systems where the performance and predictability of performance is essential."	18:02
arianepaola	but it's from 2005	18:02
-!- leagoetz [~leagoetz@nat-231-184.internal.eduroam.ucl.ac.uk] has joined #shogun		18:03
@HeikoS	arianepaola: thanks, checking :)	18:03
@lambday	OXPHOS: actually, what wiking wrote there is pretty doable :)	18:04
@lambday	wiking: with a few changes here and there	18:04
leagoetz	arianepaola: just saw your email, thanks!	18:04
arianepaola	leagoetz: :-)	18:04
@lambday	arianepaola: thanks..	18:04
OXPHOS	lambday: I had question with a.onGPU() part	18:05
@lambday	OXPHOS: yes please ask	18:05
OXPHOS	if it is a SGVector, I can't see a way to avoid ifdef in SGVector	18:06
@lambday	OXPHOS: how about we make it work with the base?	18:07
@lambday	OXPHOS: let me make a draft	18:07
OXPHOS	lambday: was trying to work sth out with wiking and HeikoS 's comments. But turned out to be what I have now...	18:07
arianepaola	found a speed comparison (2010), tested with g++, clang, MSVC http://tinodidriksen.com/2010/04/14/cpp-dynamic-cast-performance/	18:09
OXPHOS	arianepaola: thanks!	18:10
-!- sanuj [~sanuj@117.203.1.250] has quit [Ping timeout: 260 seconds]		18:10
arianepaola	OXPHOS: one more https://ciaranm.wordpress.com/2010/05/24/runtime-type-checking-in-c-without-rtti/	18:10
arianepaola	OXPHOS: :-)	18:10
@HeikoS	arianepaola: haha lots of stuff going on there	18:11
@HeikoS	but you know, I think this one is not really the problem	18:11
@HeikoS	as we can static cast things	18:12
@HeikoS	I think it is also more clean	18:12
@HeikoS	simple flag, if-then-else and cast, done	18:12
-!- sanuj [~sanuj@117.203.1.250] has joined #shogun		18:13
@HeikoS	we will only ever have two classes	18:13
@HeikoS	not 100s	18:13
arianepaola	:-)	18:13
@HeikoS	OXPHOS: lambday is drafting something here next to me	18:14
OXPHOS	HeikoS: please say thank you for me. And thank you XD	18:15
@HeikoS	will say so :)	18:15
@lambday	OXPHOS: something like https://gist.github.com/lambday/ffc5ac685650775772e8eda0994257c5	18:18
sanuj	HeikoS, the int stuff in common.h is like	18:19
arianepaola	does anyone know what the latest clang version is that is used on the buildbot instances or that compiles shogun git without any major problems?	18:19
sanuj	HeikoS, typedef int machine_int_t;	18:19
@HeikoS	arianepaola: you can always check the buildbot logs	18:19
arianepaola	by any chance it is 3.8?	18:19
@HeikoS	I dont know by heart	18:20
sanuj	and typedef int32_t index_t;	18:20
@HeikoS	sanuj: thats not enough?	18:20
sanuj	okay	18:20
sanuj	sergey told me the same	18:20
sanuj	he told me to do typedef int int32_t	18:20
arianepaola	it's 3.8 :-)	18:20
sanuj	HeikoS, but 2 typedefs with same alias won't work	18:21
@HeikoS	sanuj: chain them maybe?	18:21
sanuj	what's that	18:22
OXPHOS	lambday: thank you. very clear.	18:22
@lambday	OXPHOS: the details are not really there.. but maybe it is doable with just two different vector classes	18:22
@lambday	OXPHOS: we avoid dynamic casting altogether	18:23
@HeikoS	sanuj: define the one you want, then then use it to define the other one	18:23
arianepaola	Is there anything particular about the ppa:george-edison55/cmake-3.x (having cmake 3.4.1) used in the Dockerfile. We could use the official ones from http://llvm.org/apt/	18:24
OXPHOS	lambday: So the user has to do the SGVector->vector transfer? Or can we actually put the vector class in SGVector?	18:24
@HeikoS	arianepaola: wiking know that, maybe drop an email	18:24
@HeikoS	no idea, I think it was just the first he found	18:24
@lambday	OXPHOS: yeah the user calls the factory..	18:25
OXPHOS	If every single SGVector owns a vector, 1.users can directly call dot(SGVector) 2. Factory will be called only once for one SGVector instance	18:27
OXPHOS	lambday ^	18:27
arianepaola	:-) sure, I just need to "Vagrant Docker" the things a little bit (https://github.com/arianepaola/shogun/tree/feature/vagrant_docker/configs + the fixed Dockerfile from #3200)	18:28
@lambday	OXPHOS: maybe we won't make any changes in the SGVector class at all..	18:28
@lambday	OXPHOS: so, the factory will take a SGVector, and will return a CPUVector instance or GPUVector instance..	18:29
@lambday	that's dynamically decided..	18:29
@lambday	and then using that instance, we call our linalg::dot(...)	18:29
@lambday	that's how it will be used	18:29
arianepaola	is there a particular swig version dependency? or will both 2.x and 3.x work equally well?	18:29
@lambday	so as long as we don't let go of the Vector instance that we got from the factory, the data will still be on GPU	18:30
OXPHOS	lambday: I got it. Just think it'll be more concise in SGVector	18:30
@lambday	we'll only release that one, when we're done with it	18:30
@lambday	OXPHOS: well, SGvector is being used in many places.. and we have used SGVector.vector and SGvector.vlen in many places.. plus it already has a working map to modular types.. (e.g. numpy arrays)..	18:31
@lambday	if we touch SGvector, then we may mess up many things	18:31
@lambday	this linalg stuff is more of internal.. we won't need to provide any modular interfaces for these..	18:32
@lambday	that means, no one will call this method from python when using shogun	18:32
@lambday	this is just our shogun developer's use	18:32
OXPHOS	lambday: ahh got it. you're absolutely right	18:32
@lambday	OXPHOS: think you can make a small program like this work? a running gist?	18:33
sanuj	lisitsyn, ping	18:33
OXPHOS	lambday: yes	18:33
@lambday	OXPHOS: don't worry about the factory and the details now.. just a working implementation of the gist I sent you would be nice	18:33
@lambday	we can test it by creating the appropriate vectors by hand for now	18:33
OXPHOS	lambday: cool! thanks so much	18:34
@lambday	OXPHOS: no worries...	18:34
@lambday	OXPHOS: let me know if you have further questions :)	18:34
sanuj	lambday, you know ctags?	18:34
@lambday	OXPHOS: looking forward to see the gist :)	18:34
@lambday	sanuj: nope!	18:34
sanuj	okay	18:34
OXPHOS	@lambday: copy that!	18:34
@lambday	OXPHOS: over and out :D	18:35
@HeikoS	arianepaola: I think both should work, 2 and 3	18:36
@HeikoS	there are some cases that dont work, but these are blocked from cmake	18:37
sanuj	HeikoS, have some questions	18:38
@HeikoS	sanuj: shoot	18:38
@HeikoS	haha gsoc madness in IRC :)	18:38
sanuj	look at common.h	18:38
sanuj	what's machine_int_t	18:38
arianepaola	thanks HeikoS	18:38
@HeikoS	sanuj: that is used for enums in swig	18:39
sanuj	okay	18:39
@HeikoS	CT_CLASSIFIER	18:39
@HeikoS	K_GAUSSIAN	18:39
@HeikoS	etc	18:39
sanuj	HeikoS, how do i add int32_t	18:39
@HeikoS	ever saw those?	18:39
sanuj	in python?	18:39
@HeikoS	sanuj: yes from modular	18:39
sanuj	yeah	18:39
@HeikoS	sanuj: btw	18:40
@HeikoS	just an idea	18:40
@HeikoS	check targets/cpp.json	18:40
@HeikoS	in the meta example dir	18:40
sanuj	yeah	18:40
sanuj	it's open	18:40
@HeikoS	there you can "select" to what c++ type your "Int" in the meta examples is translated	18:40
@HeikoS	can't you just choose something that works in there?	18:40
@HeikoS	i.e. that is defined in common?	18:40
sanuj	there is index_t and machine_int_t	18:41
sanuj	HeikoS, shall i choose machine_int_t	18:41
sanuj	and i'll have to do this for other jsons also	18:42
sanuj	java and csharp	18:42
@HeikoS	might be	18:42
@HeikoS	no maybe lets stay with int32_t	18:42
sanuj	HeikoS, how do i add it in common.h	18:42
@HeikoS	typedef int int32_t	18:42
@HeikoS	doesnt work?	18:42
@HeikoS	and then typedef int32_t machine_int_t	18:43
@HeikoS	?	18:43
sanuj	can we have 2 typedef with same alias name?	18:43
@HeikoS	its not that	18:44
sanuj	there is already typedef int machine_int_t and typedef int32_t index_t present in common.h	18:44
@HeikoS	you just define the one you want	18:44
sanuj	oh	18:44
@HeikoS	and then change the other definition	18:44
@HeikoS	and base it on the one you want	18:44
@HeikoS	that was what I meant with "chaning"	18:44
@HeikoS	sanuj: does that work?	18:46
@HeikoS	just getting some water, back in 5 mins	18:46
sanuj	i'm checking	18:46
sanuj	oh man sorry	18:48
sanuj	i had lost it for a while :P	18:48
sanuj	i was thinking that in "typedef a b" a is alias	18:49
sanuj	sigh	18:49
@HeikoS	haha	18:55
@HeikoS	to deep in the code	18:55
@HeikoS	does it work now then?	18:55
sanuj	building it	18:55
-!- Saurabh7_ [~Saurabh7@117.215.53.11] has quit [Quit: Leaving]		18:57
sanuj	HeikoS, not working	19:03
sanuj	same error	19:03
lisitsyn	back	19:04
@HeikoS	sanuj: what did you exactly do now?	19:04
sanuj	typedef int int32_t;	19:05
sanuj	typedef int32_t machine_int_t;	19:05
@HeikoS	mmh	19:08
@HeikoS	weird	19:08
@HeikoS	re-run ctag?	19:09
@HeikoS	renmove the tag file and re-run I mean	19:09
@HeikoS	maybe it wasnt updated	19:09
@HeikoS	sanuj: I gotta leave now, have to continue tomorrow	19:09
@HeikoS	bye	19:09
-!- HeikoS [~heiko@nat-229-174.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.]		19:09
lisitsyn	yes please try that	19:10
lisitsyn	could help	19:10
-!- leagoetz [~leagoetz@nat-231-184.internal.eduroam.ucl.ac.uk] has quit [Remote host closed the connection]		19:11
sanuj	yeah	19:11
sanuj	doing	19:11
@lambday	OXPHOS: there?	19:21
sanuj	lisitsyn, it worked with a fresh build	19:39
arianepaola	sanuj: fresh builds = fresh baked bread :-) always tasty	19:52
sanuj	arianepaola, haha	19:53
rcurtin	HeikoS: I am here now, but you are not :(	20:01
-!- lambday [8028b10a@gateway/web/freenode/ip.128.40.177.10] has quit [Ping timeout: 250 seconds]		20:13
-!- sanuj [~sanuj@117.203.1.250] has quit [Quit: Leaving]		20:24
arianepaola	until tomorrow everyone	22:03
lisitsyn	see ya	22:09
-!- besser82 [~besser82@fedora/besser82] has quit [Ping timeout: 260 seconds]		22:49
-!- besser82 [~besser82@fedora/besser82] has joined #shogun		23:51
-!- mode/#shogun [+o besser82] by ChanServ		23:51
--- Log closed Wed May 25 00:00:10 2016

Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!