#qi-hardware IRC log for Monday, 2012-04-09

DocScrutinizerroh: ping00:24
rohDocScrutinizer: pong02:04
DocScrutinizerroh: hi! :-)02:08
DocScrutinizerroh: any news or ETA guessing for the big move of doc.om.org et al? seems quite some people are missing svn and a few others02:09
rohnope... sean sadly doesnt react on mail so its delaying02:10
rohbut i think i will happen sometime soon02:10
rohpropably while regular work-hours02:11
DocScrutinizerany chances you kick svn and doc vserver? until the move02:12
rohthe vm is up, but its filtered on the outside02:12
DocScrutinizero.O ??02:13
rohhad some 'guest' using it as spam-relay quite a while ago02:13
DocScrutinizeryeah, munin seemed to suggest they are all running02:13
rohkicked him out, but the block needs (signed?) paperwork02:13
DocScrutinizeroh shit02:13
rohwell.. it will be resolved with the move, afaik02:14
DocScrutinizercould somebody (me?) contact hetzner and ask what's the deal?02:14
DocScrutinizerre block02:14
rohin the current legal think i think sean would need to sign, but since we change all that it will be much easier.02:15
rohlegal thing of course02:15
rohrelay. is ostermontag. nothing will happen in germany anyhow.02:16
rohirgh. relax02:16
DocScrutinizerno hurry, no hassle. Just asking02:16
DocScrutinizeresp since admintrac is down02:17
DocScrutinizerand I published nagging me via IRC is the fallback method then02:18
rohhetzer said they replaced the hw..02:19
rohwell.. it seems agni is completely down and i dont know why.02:25
DocScrutinizerduh, again?02:25
DocScrutinizernot since long02:25
rohi cant reach it.02:25
rohanyhow.. will drive to the lab now. and dont think much will happen soon02:26
kyakxiangfu: hi05:15
kyakxiangfu: a question: do we still need everything from nanonote-files/data/qi_lb60/files/etc/uci-defaults?05:15
kyakit's setting up some links which was a workaround in previous releases, but now it could be unnecessary/dangerous05:16
kyakxiangfu: another question: nananote-base-files only provides kernel-6x11-font. Maybe we should move it to script-files and get rid of base-files at all? It seems strange that we provide both scripts/data with script-files and kernel-6x11-font looks like forgotten in base-files :)05:18
kyakxiangfu: and the last question; is there a reason why you decided not to send all openwrt-xburst patches to openwrt upstream?05:19
kyakmy understanding is that if we use these patches for release, they are all ready for openwrt upstream. I'm not saying they are ready for kernel upstream05:22
xiangfukyak, Hi.05:51
xiangfujust back from lunch. 05:51
xiangfu1.(nanonote-files/data/qi_lb60/files/etc/uci-defaults)  I think we still need this one: 01-optimization-for-ben-nanonote. the other two can remove.05:56
xiangfu2. (kernel-6x11-font) totally agree. 05:57
xiangfuthe plan is 'scripts' for the files that nanonote needs and don't over-write any other openwrt base files.06:02
xiangfuand hte 'data/qi_lb60/files' is for the files that over-write the openwrt base files.06:02
xiangfuI needs clean up them.06:02
xiangfuI think the namd of 'script-files' is not fit for those files any more.06:04
xiangfuI will move all files to base-files. just like the openwrt names06:05
kyakthis is a good idea06:09
kyakbtw, do we need the posinstall section in nanonote-files Makefile?06:10
kyaki think the idea was to overwrite openwrt base-files, but this doesn't always work (depends on installation order).. since we set up a symlink like feeds/local_qipackages/nanonote-files/data/qi_lb60/files/ in openwrt-xburst, we coudl remove the postinstall 06:11
xiangfuI think we needs keep the nanonote-data-files/postinst. it should be works fine on overwrite06:11
xiangfukyak, my plan is remove the 'CONFIG_PACKAGE_nanonote-base-files=y' in config.full_system.06:12
xiangfubut people can still opkg install this package.06:12
xiangfuthe opkg install should works fine with postinstall.06:13
kyakyeah, totally agree then06:13
xiangfukyak, the build order. yes correct.06:13
xiangfuI think we needs move 'gmen2x' to base-files 06:14
kyakxiangfu: maybe we should do like that: put all gmenu2x icons in nanonote-gmenu2x-icons package, OR - just get rid of all nanonote-whatever-package and just keep everything (except for example files) in nanonote-base-files06:17
kyakto make things simple06:17
xiangfu(nanonote-gmenu2x-icons) great. 06:17
kyakxiangfu: btw, gmenu2x-settings, can't exit on "S" -\06:18
kyakhave to kill gmenu2x :)06:18
kyakwe kind of paid too small attention to gmenu2x, and it floated away in direction of dingoo.. too many regressions on Ben06:19
xiangfulet me commit first. I will split the gmenu2x-icons in next commit. 06:23
kyakit's great if we reorganize nanonote-files in a smart way. I'm starting to get lost in all those files :)06:24
xiangfukyak,(gmenu2x) maybe we switch to David's 'alfilesel' someday :-)06:24
kyakyeah! it's simple and just works06:24
qi-bot[commit] Xiangfu: cgminer update to 2.3.2, update the BIT_ENDIAN patch (master) http://qi-hw.com/p/openwrt-packages/9d550b806:25
qi-bot[commit] Xiangfu: nanonote-files: cleanup the file structure. update version to 201204 (master) http://qi-hw.com/p/openwrt-packages/223a15806:25
xiangfukyak, yes. gmen2x is big. complicate. 06:25
kyakxiangfu: so basically it's like that now: 1) nanonote-base-files is to keep all files (both scripts and data) except for gmenu2x icons and /etc. 2) Those files in /etc are provided by the "files" symlink or nanonote-data-files on "opkg install" 3) nanonote-gmenu2x-icons holds gmenu2x icons06:30
xiangfukyak, yes.06:30
kyak4) and there is nanonote-example-files holding music and whatever06:30
xiangfuyes. 06:31
qi-bot[commit] Xiangfu: naonote-files: split the gmenu2x icons to one package (master) http://qi-hw.com/p/openwrt-packages/65baf0106:31
qi-bot[commit] Xiangfu: nanonote-files: config.full_system include gmenu2x-icons (master) http://qi-hw.com/p/openwrt-packages/625f21506:31
kyakok, this is great.. perhaps i would suggest one more change to make it more clear: move nanonote-files/data/qi_lb60/files to nanonote-files/files (so that other directories in nanonote-files/data/qi_lb60 don't confuse)06:32
kyakthen by looking at root of nanonote-files/ we can't directly say which directory goes where06:33
kyak*we can06:34
xiangfukyak, yes. 06:37
kyaki wonder if git will handle it when i pull all the changes :)06:37
kyakneed to recreate the symlink in openwrt-xburst then06:38
qi-bot[commit] Xiangfu: nanonote-files: cleanup the file structure. move data/qi_lb60/files to package root (master) http://qi-hw.com/p/openwrt-packages/f74c59306:40
xiangfukyak, also needs update the buildhost compile script file.06:40
xiangfumore clear is better.06:40
xiangfuI will add a 'wpan-ipv4.ogg' for video sample 06:42
kyaki like David's sample video with bunny, not sure about the size though06:43
kyakprobably too much06:43
qi-bot[commit] Xiangfu: nanonote-files: compile-openwrt-xburst.sh update the files symlink (master) http://qi-hw.com/p/openwrt-packages/b2d577a06:43
kyakxiangfu: thanks a lot!06:45
xiangfukyak, example video. 20MB should be ok. since people should install to /data/ 06:47
kyakxiangfu: that is, if /data is actually mounted :)06:50
kyakand probably opkg will fail downloading ipk that big anyway06:50
kyakdue to RAM06:51
xiangfuok. yes. correct. 06:54
kyakxiangfu: what about the openwrt upstream patches?06:54
xiangfuwe are use /usr/share/opkg for download hte package. but same situation .06:54
xiangfukyak, the patches I didn't send out are 3 parts 1. WPAN patches. 2. KS7010 patches. 3. the Nanonote patches (logo, keyboard)06:55
kyaki see! any good reason for that?06:56
xiangfu1. (wpan) I needs split it to atUSB and atBen. for atUSB needs to send to target/linux/generic for all targets.06:56
xiangfu2. not sure if we need that any more. keep in our repo first. 06:57
xiangfu3. since those patches is for Ben Nanonote not fit for 'xburst' target. 06:57
kyakah, ok; sure06:58
xiangfuit will influence  other xburst target. 06:59
kyakyeah, i understand now06:59
xiangfunext task split WPAN first. atUSB atBen.06:59
xiangfuif we have atUSB upstream that will be great. Ben Nanonote can connect to openwrt router(with usb host) 07:00
kyakyes, this would be great07:00
kyaki will go an upgrade my router to the one with usb host right away :)07:00
kyakxiangfu: do you know if atUSB is compilable is module?07:01
kyaki remember it could only be built monolitically07:02
xiangfunot sure that. 07:03
xiangfuI haven't try atUSB yet. :(07:04
xiangfuonly tried 2 atBen talk to each other.07:04
kyaki did try atusb, and it is really disturbing that i had to compile it in07:06
xiangfuwill pay attention on this built-in.07:38
xiangfuthis is cool: https://home.comcast.net/~sdwalker/uscan/index.html10:40
xiangfuI have asked sdwalker add openwrt-package 10:40
LunaVoraxHello everyone!14:44
wolfspraulFallenou: hey, nice to see you in Qi Hardware too! ;-)14:46
wolfspraulI thought you were an exclusive milkymister14:46
Fallenouhehe I'm trying to be more social14:46
FallenouI have just read your email14:47
Fallenouabout Ya NanoNote being based on Milkymist SoC14:47
Fallenouwhat a crazy (and awesome) idea :)14:47
wolfspraulwell, why do we work on an SoC if we then don't use it *everywhere*?14:48
wolfspraulit's like the Linux kernel, my friends for many years knew me as someone who would first install Linux on *every* computing device he got. OF COURSE!? huh?14:49
wolfspraulso let's see. I am learning about IC design and SoC. total beginner.14:49
wolfspraulbut I think we will use a free IC design as the central digital signalling hub for everything, sure. yes. what else? :-)14:50
wolfspraulis there any legacy stuff left out there?14:50
Fallenouthat makes totally sense to me 14:50
wolfspraulI knew it14:50
wolfspraulso then14:50
Fallenouwhy using some proprietary chip if there is a functional competitive alternative which is open source :)14:50
wolfspraulyes, a few pieces of legacy here and there, we clean them up, and then into the future14:51
Fallenouindeed it's work14:51
Fallenoubut I think you don't fear working =)14:51
wolfspraulthere's just enormous power available to us once we move beyond the x86/arm black box / iron curtain14:51
wolfspraulsince I only do what I love, how could I fear anything?14:51
wolfspraulno, it feels right14:52
wolfspraulmilkymist one today is still too basic, many things it could do don't work today14:52
Fallenouyes a lot of work to do14:53
Fallenouyou would need a few dozen engineers 14:53
Fallenouevery open source project does not have the Linux work force14:53
wolfspraulwell I don't like to look at this just by how much one could work14:54
wolfspraulthe reason there are many linux devs is because linux is valuable to many companies and lots of the devs make their living that way14:55
wolfspraulkeep in mind that we made a usable and fun product around the milkymist soc today14:55
wolfspraulit's not like "we need another xyz"14:55
wolfspraulno we need nothing14:55
wolfspraulwe just need m1 and some good music14:55
wolfspraulthe rest is non-technical :-)14:55
Fallenousure 14:57
FallenouI was just thinking about work force because there is a road map and a lot to do (usb, mmu, linux etc)14:57
wolfspraulevery creator is faced with shortage of resources14:58
wolfspraulfrom people with access to vast resources - how many create something lasting?14:58
wolfspraulso let's see. you have limited resources, and became passionate about the mmu, which a lot of people would love to have. is it enough to create something? we shall find out :-)14:59
wolfspraulbut whatever the result is, it's worth trying and the most important is that you enjoy being on that path! irrespective of the result14:59
wolfspraulby the time we just buy what we need things are boring15:00
wolfspraulFallenou: you mentioned some roadblocks recently? did sleeping over them help in any way?15:01
wolfspraulwriting this up could be helpful in itself, since there are a number of beginners/lurkers around15:01
Fallenouoh a sum up of the state of the mmu dev ?15:02
FallenouI'm planning on sending a sum up email with a call for advices and ideas, not done yet, I tried a few more things yesterday15:03
wolfspraulyou mentioned something you got stuck on15:03
wolfspraulah ok15:03
Fallenouyes I'm still stuck :(15:03
wolfspraulI see15:03
FallenouI asked a few fpga engineers , they didn't help much about my problem15:03
Fallenoubut they did point a few things in Milkymist design15:04
Fallenounot sure if it's a real issue but I could talk to lekernel about it15:04
Fallenouabout meta stability, asynchronous reset and DCM initialization15:04
wolfspraultrust yourself15:04
FallenouI lost a few hours trying to get my hands on ModelSim , a verilog simulator15:05
wolfspraulthe most those people can often help you with is to be angry at their uselessness and finally realize yourself what's going on :-)15:05
Fallenouright now I am doing simulations using ISim (ISE stuff)15:05
FallenouI wanted to try something else, since ISim does not help much here15:05
Fallenouthe MMU is working in ISim, so there must be something wrong with the simulation being too optimistic somewhere :)15:06
Fallenou17:11 < wolfspraul> the most those people can often help you with is to be angry at their uselessness and finally realize yourself what's going on :-) < ahah yes, that's the good moments, when you eventually figure out what is going wrong and fix it yourself :)15:06
qi-botThe build has FAILED: http://fidelio.qi-hardware.com/~xiangfu/building/Nanonote/Ben/openwrt-xburst.minimal-20120408-1732 21:55
--- Tue Apr 10 201200:00

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