#qi-hardware IRC log for Tuesday, 2010-06-15

Textmodeand the xburst build has exhausted all available space on my build machine. :/00:04
freespacew00t, got my nanonote02:37
kyakfreespace: gratulations!02:49
freespacenow to flash it, it apparently needs flashing :P03:27
kyakthis is the first to do03:32
tuxbrain_awayfreespace:the bnn look at you with bright eyes and says with sweet low voice, "please, flash me the last distro, keep me updated and I will make you a happy geek" but beware rear this sweetness is an angry beast if your are not careful you will end compiling your own packages, testing beta uboots, and soldering wires... I know , learn from my experience :P03:38
freespacenone of those are new experiences, thankfully :)03:59
kyakhm, so much on the Internet about USB gadget g_ether driver not working with Windows RNDIS..04:01
kyakand it's still not working :) though it used to, in some kernels04:01
kyakethernet not working, wifi not working - so i can suck with my windows workstation at work04:02
tuxbrainfreespace: then you have no way to run :) welcome to the merry nanohackers world04:08
arctanxAre there any works underway to jury-rig some sort of short-range wireless on the BNN? I thought it would be cool to set up some mesh applications, such as usenet04:41
arctanxNot official works of course, I just mean projects that people are up to04:41
Textmodearctanx: i understand that many have used SDIO wireless cards successfully04:44
Textmodehas anyone looked into how dingux's network-over-USB manages to "just work", while ben's needs configuration at both ends?04:46
arctanxWow, I never even knew SDIO existed... that's kinda neat...04:46
arctanxCheers Textmode04:47
arctanxBliiiimey those things cost as much as a nano04:49
Textmodeyeah...I heard they are a bit costly04:50
qi-commitsMirko Vogt: enable kernel floating point emulation, to provide fallback support for userspace applications compiled without native fpu-support http://qi-hw.com/p/openwrt-xburst/74a065504:57
kyakTextmode: there are no problems to set up automatic network-over-USB configuration in linux that "just works"05:11
Textmodekyak: then why doesn't openwrt do so?05:11
kyakat Bens's side, everything works from the beggining, probably you just need to set a default gateway in /etc/config/network05:12
kyakto avoid you typing route add default gw .. every time05:12
kyakfrom linux side, most distros have autoconfig scripts in /etc/sysconfig/network-scripts05:12
kyakin addition to udev rules05:13
kyakso this is a question of your distro and it's documentation05:13
kyaki'm exploring now the Ethernet over USB in Windows - now, this is a real problem05:14
Textmodekyak: then why does dingux's networking feature work instantly, while oi the same machine, with the same settings, the ben can't connect?05:14
kyakTextmode: dunno, i don't have Dingoo device :) does it create the usb0 interface?05:15
TextmodeI guess. I could dig it out, I was just wondering if anyone had looked into it05:16
kristianpaulseems the flickernoise pic still needed to improve or you atleast a GUI concept05:48
calamarztuxbrain: have you succeed with keymouse conf?06:40
tuxbrainno time yet dude, I hope this night got some time to play with it06:43
calamarzI managed to have the mouse somehow moving... but still not good :/06:43
tuxbrainreally? any clue on where the problem can be?06:45
calamarznot at all... the keymouse creates /dev/input/event3, I made a symlink from mice to it, and put some keycodes in the conf... the pointer moves, debug ok, but pointer keeps going back to 0,006:55
rafakristianpaul: what did you do to avoid a bit the flickernoise?07:10
tuxbraincalamarz: can you put your config some where to download or send me it by mail david@tuxbrain.com07:39
Textmodenight all07:40
tuxbrainnight Textmode07:40
kyakany git experts around? :) i'm trying to update my local git repo, but preserving my modifications. i do 'git checkout -m' in /my/path/openwrt-xburst, and the only output is "M       target/linux/xburst/qi_lb60/config-2.6.32"08:26
kyaki guess it means that this file is modified, ok; but why isn't git checking out latest change from remote repo?08:27
larscgit pull08:28
kyakok, this wasn't in "Everyday GIT With 20 Commands Or So" :)08:29
kyakreading the manual now, thanks larsc08:30
kyakerror: Entry 'target/linux/xburst/qi_lb60/config-2.6.32' not uptodate. Cannot merge.08:34
kyaki think i should be using git merge08:35
calamarztuxbrain: sent08:51
tuxbrainthanks, I will test this night in openwrt08:56
kristianpaulrafa: oh sorry wrong channel09:17
kristianpaulrafa: this is other flickernoise ;)09:17
kristianpaulour flickering is a sofware i think09:20
rafakristianpaul: I am asking you because I do not get good FPS with mplayer or snes9x and I was guessing that perhaps there is something to improve09:22
kristianpaulwell i dint try debian yet09:23
kristianpauli heard you could get good FPS with no audio, but once it get working FPS go down09:24
freespacemplayer on openwrt?09:25
kyaklarsc: i did "git pull -s ours", i think it did what i want. but "git status" says "Your branch is ahead of 'origin/xburst' by 1 commit." what does it mean?09:39
kyakmy local changes are not committed, is it right?09:39
freespacei think it means your branch has one more set of changes than head of origin/xburst, e.g. your locl changes09:41
kyakhow does it know the number of commits ahead? is it some kind of trickery from Linux Torvalds? :)09:41
kyakfreespace: ok, thanks for confirmation09:42
freespacei hope that is right :) i am still a git n00b myself09:42
kyaki so prefer svn09:42
freespacehehe, i don't, not any more :)09:42
kyakgit remind me of the octopus they mention so much in their docs09:43
freespaceyou just have to think about it the right way09:43
freespaceit probably helps to run gitk09:43
freespaceand visually see the history of the branches09:43
kyakmm, thanks for the hint, will do09:43
freespaceor some other git-histor visualiser09:43
rafafreespace: mplayer on jlime10:28
freespaceah, cool10:28
rafafreespace: http://jlime.com/mw4/index.php/Jlime_Muffinman if you want check.. there is a video link of mplayer10:29
freespacevery interesting10:33
freespacei wans't aware of jlime :)10:33
qi-commitsXiangfu Liu: [usbboot] fix if there is space at the end of command casue error http://qi-hw.com/p/xburst-tools/c1873f110:38
qi-commitsXiangfu Liu: correct the jz4760 lepus board mem parameter: http://qi-hw.com/p/xburst-tools/327211910:38
kristianpaulrafa: the video looks great, guessing because no audio?10:47
sdschulzelarsc: ping10:55
sdschulzemake'ing target/{clean+install} didn't rebuild my modules.10:56
rafakristianpaul: no, it has audio okey10:57
rafajust that I had a headphone connected..10:57
rafakristianpaul: if you turn up the volume of the video you will hear the sound of the movie10:57
rafaat least something, very low10:58
larscsdschulze: make package/kernel/install11:00
larscwill rebuild pack the modules11:00
sdschulzelarsc: It doesn't rebuild them.  It expects them to have been built, which isn't the case.11:06
sdschulze"cp: cannot stat `/home/soeren/openwrt-xburst/build_dir/linux-xburst_qi_lb60/linux-': No such file or directory"11:06
sdschulzelarsc: Gotcha.  It doesn't build the modules when doing a "make install".11:09
sdschulzebut only when doing a "make compile"11:09
sdschulzeHow much slower is kernel FPU emulation than userspace, actually?11:52
sdschulzeMy mplayer doesn't really run well.11:52
sdschulzeI can't play a 64kbps ogg file, which was possible in OpenWRT.12:07
sdschulzehm, though http://www.linux-mips.org/wiki/Floating_point says it's not *that* slow12:08
kristianpaulsdschulze: are u in debian?12:08
sdschulzejust compiled my own kernel :)12:08
kristianpaulohh :)12:10
kristianpaulare you documenting that?12:10
kristianpaulbut using debian way or openwrt way?12:10
sdschulzecross-compiled it using the OpenWRT toolchain12:11
sdschulzeI'm planning on documenting it, but if sound doesn't actually work, it's a bit useless.12:12
sdschulzeIt's about factor 8 too slow. :(12:16
sdschulzeyes, :(12:19
sdschulzeDon't even think of playing videos.12:20
sdschulzeUnfortunately, it would require *all* libraries to be rebuilt for soft-float.12:23
kristianpaulahh i want see this gift on my nano http://upload.wikimedia.org/wikipedia/commons/4/48/Aurora_Australis.gif12:32
mthsdschulze: there is an integer (fixed point) version of the Vorbis decoder12:36
larsclibtremor. i think thats what we use by default in openwrt12:39
coat_woohoo chatting from my nanonote12:41
sdschulzemth: So that's what is used by GMU?12:53
mthI don't know, but larsc says it is used on openwrt12:55
wejpGmu uses Tremor for vorbis decoding12:56
calamarzkristianpaul: mmm maybe zgv can do it?12:56
wejpthat is the fixed-point vorbis decoder12:56
kristianpaulcalamarz: is in openwrt?12:57
calamarzkristianpaul: don't know :/13:04
sdschulzewejp: And what about video decoding?13:19
sdschulzeI heard this had been done on the NanoNote.13:20
wejpwell Gmu doesn't doo any video decoding, but video decoding on the nanonote should probably be done with the hardwae video-accelerator that is part of the SoC13:21
wejpi think there are patches from ingenic for mplayer to support that13:22
zearbut it doesn't work very well13:22
zearthe mplayer in jlime uses ingenic patches13:22
kyakFuuu! I've managed to setup Ethernet over USB in Windowz :) this is a teaser, details will follow13:22
calamarzzear: with the proper video scaling mplayer seems to do quite nicely to me13:24
sdschulzezear: Nice -- what codecs are supported?13:24
zearsdschulze, no idea, but i guess whatever comes with mplayer by default13:24
sdschulzeNo, I mean, which are accelerated?13:25
mththe IPU can help by doing YUV to RGB conversion in hardware13:27
mthbut afaik there is no IPU driver in the NN kernel13:27
mthIngenic has a driver, but it statically allocates a 4MB buffer, which is too much on a 32MB system13:28
sdschulzea driver?  Is it free software?13:29
sdschulzeI wouldn't consider it too much if it's the only way to watch videos properly.13:29
mthit's under GPL, like the rest of the kernel13:31
tuxbrainkyak:Coooooograts man, you are a hero in my point of view, willing to see that tutorial on how to defeat  winblows13:34
sdschulzemth: Do you have a working link to those patches?14:21
sdschulze(So it requires both modification in mplayer and in Linux?)14:21
tuxbrainflac is free of propietary codec?14:38
nebajothcompletely open14:38
sdschulzeSo where do I get patches?14:39
sdschulze(We really need a central resource for NanoNote stuff.  Everyone is doing his own business.)14:39
sdschulzehttp://www.pdroms.de/news/19639/ has a dead link. :(14:40
tuxbrainmpeg2 is patented ,right?14:45
tuxbrainI'm trying to figure out the configuration paramenters for scummvm for openwrt :)14:46
calamarzerr... with g_file_storage driver, the nn would appear as a usb storage to the host, right?14:47
tuxbrain(how many times I have anwered this question lars? :P) In theory yes, but it didn't work due some DMA mambojambo on kernel14:49
sdschulzekristianpaul: http://sdschulze.dnsalias.org/~soeren/content/debian-nanonote-kernel.txt15:16
sdschulzein case anyone is interested15:16
kristianpaulohh great15:17
sdschulzeNow you can play uncompressed audio in Debian...15:19
sdschulzeDo you know if any JLime guy hangs around here?15:23
Action: tuxbrain is not jlime core at all but maybe I can help15:25
sdschulzeI wonder how to port the audio/video acceleration from JLime to Debian.15:26
tuxbrainsdschulze: rafa, zear, kristoffer are your guys, if they have compiled Mplayer they should have the sources15:26
sdschulzezear: <tuxbrain> sdschulze: rafa, zear, kristoffer are your guys, if they have compiled Mplayer they should have the sources15:30
sdschulze<sdschulze> I wonder how to port the audio/video acceleration from JLime to Debian.15:31
zearif you're referring to mplayer ingenic patches, that's the right source15:32
sdschulzeok, thanks15:32
sdschulzezear: Are the Ingenic patches all in the same place, or are they distributed around in the source?15:37
zearsdschulze, i have no idea, that's the source of dingux version of mplayer, which uses the ingenic patches15:37
zearthat's all i know about it15:37
sdschulzezear: So this one doesn't use kernel IPU (as suggested by the file name)?15:40
zearipu (at least on the dingoo) reserves 4MB of ram, so the performance of all the other apps drops15:40
sdschulzeWell, you can still enable swap...15:43
sdschulzeBut if it's possible without it, even better.15:45
sdschulzeErm, yes, it *is* spread around in the source. :((15:55
zearsdschulze, ipu doesn't give much of a boost15:56
zearthe changes are almost unnoticable, at least on a dingoo15:57
sdschulzeSo video support works will, if you convert them to the right format before?15:57
sdschulzeAt least they documented that they based it on MPlayer-1.0rc2-20090218.15:58
zearno idea, i haven't done any tests with mplayer on the nanonote16:02
zearask rafa16:02
xakhhow do I format my SD card? I'm having a bit of trouble.16:42
xakhrafa: ping16:44
xakhtuxbrain: ping17:01
tuxbrainxakh: pong17:01
xakhokay, tuxbrain, I'm having trouble with installing to SD17:02
tuxbrainmmm troubles wiht sd hum?17:02
xakhI have a new 4GB microSD17:02
xakhand I'm not sure how to do this17:02
tuxbrainon linux?17:02
xakhI wrote a new msdos partition table to it, to wipe it down17:02
tuxbrainwhat's the goal, jlime?17:02
tuxbrainfdisk /dev/sdX as root where X is the letter of the device of the sd, be sure the sd is unmounted17:03
xakhit is17:04
tuxbraindelete all partitions17:04
tuxbrainletter d17:04
xakhactually, mine comes in as mmcblk017:04
tuxbrainok then fdisk /dev/mmcblk017:04
tuxbrainthen press "d"17:05
xakhit said nothing speficied17:05
tuxbrainif you only have one partition it will be deleted17:05
xakhoh, I deleted all partitions before17:06
tuxbrainok one step done :)17:06
xakhI wrote a new partition table to the default for microSD, and deleted all partitions17:06
xakhanyway, what next?17:06
tuxbrainpress p to be sure there is no partition17:06
tuxbrainp (primary)17:07
tuxbrainnumber 117:07
xakhDisk /dev/mmcblk0: 4072 MB, 4072669184 bytes17:07
xakh4 heads, 16 sectors/track, 124288 cylinders17:07
xakhUnits = cylinders of 64 * 512 = 32768 bytes17:07
xakhSector size (logical/physical): 512 bytes / 512 bytes17:07
xakhI/O size (minimum/optimal): 512 bytes / 512 bytes17:07
xakhDisk identifier: 0x0000000017:07
mthsdschulze: the IPU driver is in the Ingenic kernel tree that is mirrored on projects.qi-hardware.com17:08
tuxbrainwhats that?17:08
mththe MPlayer with IPU support is on Ingenic's FTP site17:08
mththe one that zear linked does not include IPU support but does use the MXU (SIMD) instructions17:08
mthto properly support MXU, the kernel would have to be made aware of it though17:09
tuxbrainok you don't have any partition17:09
mthcurrently only one process can use it, otherwise the registers get mixed up17:09
xakhthat's what it said, so I was showing it to you17:09
sdschulzemth: ok17:10
xakhso anyway17:10
xakhI hit n17:10
xakhthen p, then 117:10
tuxbrainnow the funny part17:11
xakhdo I hit enter after 1, or does it need a parameter?17:11
tuxbrainjust 1 enter17:11
tuxbrainfirst cilinder the default just press enter17:12
tuxbrainthis will be the kernel partiton so 8 Mb will be enough17:12
xakhso I type 8?17:13
xakhI usually use gparted, sorry17:13
tuxbrainpress +8M (enter)17:13
xakhso I got one now17:13
tuxbrainrepeat process but his time the partition number will be 2 and size will be more tricky17:14
tuxbrainwe will reserve the end of the sdcard for swap17:14
tuxbrainwe need about 64Mb of swap or you want more?17:15
xakhI actually sorta want around 12817:15
xakhsince that's close to the amount an  SDHC card can move back and forth17:15
xakhso it means that it's still all accessible randomly17:15
xakhmaking it pretty fast17:15
tuxbrainso if the card has 4072669184 bytes17:16
tuxbrainyou have 124288 cilinders17:16
tuxbrain4072669184/124288=32768 bytes per cilinder17:18
xakhjust tell me what to put, and what kind of part I need to make17:18
tuxbrainwe need 128Mb so 134217728 bites17:18
tuxbrain134217728/32768=4096 cilinders for 128Mb17:19
xakhright, so I gotta do some subtraction here17:20
xakhI'm no good at subtraction....17:20
tuxbrainwhen you been asked to size of second partition you have to (maxciliders)-409617:20
xakhso I take off the other one, for the main part, for the kernel17:21
tuxbrainno need to this will be the begin of the partition 217:22
tuxbrainit's offered by default17:22
xakhso I type (maxcylinders)-409617:22
tuxbrainin your case 124288-4096=12019217:23
xakhdo I type the - I mean?17:23
tuxbrainyeah when asked for the end type 12019217:23
xakhso I go17:24
xakhthen that?17:24
tuxbrainbegining just enter17:24
tuxbrainend 12019217:24
xakhand then I do the swap one right?17:26
tuxbrainok the the last is easy17:26
tuxbrainn p 3 enter enter17:26
tuxbrainnow lets identify the swap as swap17:26
xakhand make the other ones the right FS right?17:27
xakhgot it17:28
tuxbrainnow w17:28
tuxbrainok sd partitioned17:29
tuxbraintime to format17:29
tuxbrainmkfs.vfat /dev//dev/mmcblk0p1 (I gess)17:30
tuxbrainsorry mkfs.vfat /dev/mmcblk0p117:30
tuxbrainmkfs.ext2 /dev/mmcblk0p217:31
xakhisn't ext2 sort of deprecated?17:31
xakhI mean, there's 2 versions above it17:32
tuxbrainis the best for sd, ext3 an ext4 add journaling than stress too much sd17:32
xakhah ok17:32
tuxbraincontinous writting17:32
tuxbrainbad thing on sd17:32
tuxbrainthats also why I put swap at the end17:32
xakhdo I have to mkfs swap17:33
tuxbrainyep mkswap /dev/mmcblk0p3 and you are done17:34
tuxbrainit a little bit tricky at first but when you get the point , is more faster than any graph partition manager app17:35
tuxbrainand also scriptable17:35
xakhnow to put the OS on it17:36
tuxbrainnow put the kernel imge as "uImage" in p1, untar the rootfs in p2 and next boot use M+power17:37
tuxbrainand hopefully you will have a fine splash jlime logo17:38
tuxbrainwhat happens?17:38
xakhno no, I mean wtf was that ast bit17:38
xakhveeeery confused17:38
xakhlike, how do I do all that bit.17:39
xakhoh, and important bit, I suppose17:40
tuxbrainjust mount the sd partitions (on most Linux sistems just deatach and attach again phisically will mount dem automatically)17:40
xakhoh ok17:40
xakhso do I do this terminal?17:41
tuxbrainwhat? sorry I don't understand17:42
tuxbraindo I do?17:42
tuxbrainwe can continue on terminal if you want17:42
tuxbrainfor me is better17:42
tuxbrainI can then c&p the commands here17:43
xakhokay, just checking17:43
xakhthat works for me17:43
xakhfor reference, my jlime folder is /home/xakh/jlimenano17:43
tuxbrainok enter in the p1 partition of your sd card (whatever you system has mounted it)17:44
tuxbrainbtw what linux you are using17:44
xakhubuntu 10.0417:44
xakhI know, I know, ubuntu17:44
tuxbrainI use it to shhhhhh nobody is listening :P17:45
tuxbrainok to the poing17:45
tuxbrainthen enter on mmcblk0p117:45
xakhyou said okay on to the poing17:46
xakhanyway, so I cd to it you mean?17:46
tuxbrainhahaha ok to the point sorry :P17:46
xakhsaid I couldn't17:47
xakhI have them both mounted in nautilus17:47
tuxbrainand what is the directory17:47
xakhso I'm in there17:47
xakhbrb, someone's coming over soon, realized I need pants17:48
tuxbrainwget http://jlime.com/downloads/releases/muffinman/bootloaders/openwrt-xburst-qi_lb60-u-boot.bin17:48
tuxbraintoo much information, dude17:49
xakhI already got it last night17:49
tuxbrainthen just cp to this dir and rename it to just "uImage"17:50
tuxbrainthen on the other partition , (cd to it)17:50
tuxbrainrename it17:51
xakhI meant which one of those17:51
tuxbrainthe uImage I made a mistake17:51
tuxbrainthis link is for the uboot sorry17:51
tuxbrain5 mins man17:52
xakh mins?17:52
tuxbrainminutes :P17:52
tuxbrainI must leave a moment17:52
xakhoh ok17:52
xakhsorry, in command mode17:52
xakhI'll be back in a few minutes as well17:54
xakhoh back17:55
tuxbrainhere again17:59
tuxbrainhave the kernel in place? lets goo to rootfs18:01
xakhwell not yet18:01
xakhcan I just throw it in using nautilus?18:01
tuxbrainin vfat yes, then just rename it(F2) to just uImage18:02
xakhgot that18:02
xakhnow the next one18:02
tuxbrainok this is better do on terminal18:02
xakhlet's do this18:02
tuxbraincd to the p2 partition18:03
tuxbrainthe directory I mean18:04
xakhway super long name18:04
tuxbrainsudo  tar -xzf /watever/you/have/the/tar.gz/with/rootfs/Jlime-muffinman-image-2010.1-ben-nanonote.rootfs_READY.tar.gz18:04
tuxbrainok once you cd /media/f57c4505-ae1e-4f61-a275-71806f00042318:05
xakhok got it18:05
tuxbrainsudo tar -xzf /watever/you/have/the/tar.gz/with/rootfs/Jlime-muffinman-image-2010.1-ben-nanonote.rootfs_READY.tar.gz18:05
tuxbrainso once is finished18:06
tuxbraintype "sync" to make sure all things are written to sd18:06
tuxbrainand then just unmount and plug the sd to BNN18:07
xakhit only took like a second18:07
tuxbrainalso the sync18:07
xakhthat's good right18:07
xakhoh ok18:07
xakhhit sync18:07
xakheverything just sorta stopped18:08
xakhnow over18:08
xakhso I unmount everything?18:08
tuxbrainif it complains just close the terminal or exit the sd directory18:09
xakhokay got it18:09
xakhat login screen on jlime18:09
xakhwass da password18:10
xakhtime to startx18:10
tuxbrainwelcome to the jlime proud crowd :)18:11
xakhso, how do I get iceWM goin18:12
tuxbrainheheheh you are not friend of read howtos, isn't it?18:12
xakhI don't know where they are18:13
xakhalso like tuts from people18:13
tuxbrainin jlime site is the documentation18:13
xakhcan't find the parameters18:16
Textmodemorning all19:45
rafalarsc: I would like to ask you about power-button. It is /dev/input/event0 in openwrt.. how is it hockeed to shutdown the device?.. or kernel manages it internally?. I was guessing that some stuff on userland is checking it and does the proper stuff to turn off the device.19:58
rafalarsc: :) it seems that fnkeys binary read power button event and calls poweroff binary.. now I need to read the code of poweroff binary. I am asking because I have been able to power off my nn with 2.6.3420:17
larscrafa: thats correct20:50
rafalarsc: great. Anyway, I have been able to power off the machine. I have included acpi packages in my rootfs .. perhaps that is the problem ? (or something in userland?)20:57
rafalarsc: for kernel side I have :20:58
rafa# CONFIG_PM_DEBUG is not set20:58
rafalarsc: if you see something wrong, please let me know20:58
rafaI read something related in kernel mailing list (with problems in acpi code broken for 2.6.34rc, but I am not sure if that is something current now)21:00
larsclooks good. you won't need acpi packages i guess21:04
freespacemmm, newest firmware has that graphical thingy21:42
freespacewell, now i know what i want to do first: remove the graphical thingy :P21:42
emebjust say no to graphical thingies21:42
freespacethat's me21:43
freespaceunless actual graphics is involved21:43
emebespecially graphical thingies with sekrit navigation keys.21:43
freespaceoh yes21:43
freespacethat frustrated the hell out of me21:43
emeband don't forget the overclocking code that munges up the LCD display21:44
freespacei am not sure why the B button actuall launches applications21:44
freespaceinstead of enter21:44
freespaceb which is actually x21:44
rafaemeb: what is that overclocking code?21:58
emebrafa: not sure - just was reading some things about how there was some clock control code from the dingoo in there22:00
nebajothwell then23:10
qi-commitsXiangfu Liu: [xbboot] put all jz4760 board configure to board-jz4760.h file http://qi-hw.com/p/xburst-tools/dbaf54d23:38
--- Wed Jun 16 201000:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!