#qi-hardware IRC log for Saturday, 2010-08-07

wpwrakandres_calderon: sorry for the interruption. how do you like fped ? did it crash on you yet ? :)00:01
andres_calderonwpwrak very powerfull, is not trivial (ha sido un gusto adquirido)00:02
wpwrakandres_calderon: (autorouter) ah yes, i saw that page a while ago. looks a bit chaotic :) does it perform well in real life ?00:02
wpwrakandres_calderon: exactly :) looks very queer00:03
andres_calderonnever done crash00:03
wpwrakkewl :)00:04
andres_calderonwpwrak no se si funciona el ruteo topologico, nunca lo he usado...  El código me pareció horrible. Pero me gustaría tener algo así en Kicad.00:06
wpwrakandres_calderon: there is one project that's doesn't look too bad. http://code.google.com/p/kicadocaml/00:08
wpwraki think they also have a more conservative geometry. at least they did when i looked at it the last time, about a year ago00:09
andres_calderonwpwrak parece interesante, voy a probarlos00:10
wpwrakback then, the author said he wasn't sure if he'd continue. but it looks quite alive, which is a pleasant surprise :)00:11
andres_calderonwpwrak lo que ud. hizo con heekspython me pareció interesante. Tal vez algo así sea útil en Kicad.   http://code.google.com/p/heekspython/00:12
wpwrakwell, you can always create kicad files from your scripts :) the power of plain text00:13
wpwraki like heekspython. it has many flaws, but it's the right direction00:13
wpwrakfor kicad, scripting is a little harder. i'm actually a bit surprised that people define components with fped using the scripting language - instead if the gui00:15
wpwrakof course, it's nice to see the two coexist so nicely :)00:15
wpwrakbut i don't think there are many places besides footprint definition in kicad, where scripting would normally be important. perhaps if you need, say, an array of buttons or leds00:16
andres_calderonwpwrak el poder de kicad está en los scripts :)00:17
wpwraki like the ability to extend kicad this way, yes. build around it.00:17
wpwrakandres_calderon: so, so far you're happy with kicad for Xue ?00:20
andres_calderonI do not like the GUI, but the potential is great.00:21
andres_calderonwpwrak poor clipboard support, copy/paste problems, lack of short-cuts, etc00:23
wpwrakandres_calderon: let me guess - do you like the block operations ?00:23
andres_calderonwpwrak no mucho... porque se quedan cortas al no permitir operaciones entre esquemáticos diferentes  que hacen parte de la misma jerarquia.00:26
wpwrakthere are quite a few short-cuts. what's a bit annoying is that they don't work the same way everywhere. also, the dialogs for clarifying one's selection can be a bit superfluous. for example, if you're about to delete a track, it doesn't really matter which segment of that track you selected. yet you still have to choose.00:27
wpwrak(block) yeah, moving things around in the hierarchy is always a bit of a mystery. when i play with it long enough, it eventually works, but i'm never quite sure why it didn't work before :-(00:28
wpwrak(block) but what i meant was the shift-ctrl-mouse stuff. i find that quite ugly. do you know qucs circuit simulator ?00:29
andres_calderonwpwrak yes, there are usability issues ... but I find it tedious to try to fix them. I do not like GUI programming00:30
andres_calderonwpwrak No he usado el simulador de circuitos00:31
andres_calderonwpwrak I am very new to kicad,  you've seen recent progress in the development?00:33
wpwrakthere is constant development in kicad. some good new stuff are the net classes. they let you do things like specifying that all ground traces should be at least 8 mil while the rest can be 6 mil. or to have a really wide clearance around that 220 V track ;-)00:41
wpwrak(qucs) it's pretty cool. a bit like spice, but in a way that actually makes you want to use it :)00:42
kristianpauldont blame for this but there is a good autoruitng feature in kicad, may be like addon?00:42
kristianpaulhello wpwrak and andres :)00:43
wpwrak(qucs) the gui is very nice, in the "macdraw" style.00:43
wpwrakkristianpaul: there is an external router, http://www.freerouting.net/00:43
wpwrakkristianpaul: it's not an autorouter but a pretty good push router00:44
wpwrakthe problem with it is that it's not free as in "Free Software", and you have to be online to use it00:44
wpwrakit also used some obscure java streaming extension that can sometimes be fun to install00:45
kristianpaulwow wolfspraul link about toporouter looks promising00:45
Action: kristianpaul no java guy00:45
kristianpauloh wait andres_calderon pointed :)00:46
kristianpaulwpwrak: are you bloging or documenting your experience with heekscad somewhere00:47
kristianpauli do use heeks but not as you do :)00:47
kristianpaullike the script for the scanned parts00:47
kristianpaulnice :)00:47
wpwrakkristianpaul: (heekscad) no blog. but i try to solve problems where i find them. at least some of them :) right now (well, for the last week ... with many interruptions), i'm trying to fix selections sometimes crashing on amd6400:50
wpwrakkristianpaul: i would also like to improve the heekspython interface00:50
kristianpauli used to crashing it happens always..00:51
kristianpaulyup heekspython looks promosing, but jsu few samples on wxpython like gui00:51
kristianpaulwpwrak: i'm curios what are you usings heeks for?00:52
kristianpauli just did some basic cases and things for fun00:52
kristianpaulbut it is really easy to use i think more experienced people on cad ca really take advantage of ir00:53
wpwrakkristianpaul: right now, only for the counterweight00:53
Action: kristianpaul no cad guy :P00:53
wpwraka lot of features are missing on the cad side. it's quite powerful on cam, although severely underdocumented and many things are fragile.00:54
kristianpaulbtw well i have two floss 3D printers nothing sooo accurate but if you need print something in low resolution i think i can help, i will like :)00:54
kristianpaulyes :(00:54
kristianpauleven more if is related with the nanonote :)00:55
kristianpaulor ther copyleft projetcts00:55
wpwrak;-) i should add an extruder to my mill. then could be cool and 3d print too ;-)00:56
kristianpauloh nice00:57
kristianpaulah you use heeks cnc i guess ;)00:57
kristianpauli wonder thats what most people use00:57
wpwrak(heekscnc) yup :) see http://projects.qi-hardware.com/index.php/p/ben-counterweight/source/tree/master/README00:58
wpwrakhmm, i guess i shold mention that this is for my local setup :)00:58
kristianpaulwell i need bed, nice to meet you, hope you still hanging up here00:58
wpwrak(the "workflow" section. the rest is general)00:59
wpwrak(hang out) yeah, pretty much all the time :)00:59
kristianpaulah i dint notice :p01:00
kristianpaulgn8 then !01:00
wpwraktuxbrain_zzz: the SIE logo looks a bit like a compact 2.4 GHz antenna :)01:00
wpwrakkristianpaul: sweet dreams ! :)01:01
andres_calderonwpwrak nite, ha sido agradable hablar contigo.01:06
wpwrakandres_calderon: nice to have you here ! :) see you !01:07
kristianpaulandres_calderon: youl hang arond more often01:08
kristianpauldamm i need sleep01:09
wpwrakkristianpaul: or caffeine :)01:09
andres_calderonof course, CY01:09
kristianpaulandres_calderon: di you got this toporouter working?01:09
wolfspraulwpwrak: yes we are attacking the RF problem from _ALL_ angles now01:09
wolfsprauleven the designers were asked to help01:10
wpwrakwolfspraul: ;-)) meanwhile, i keep on looking at zigbee. i have a few atmel chips that i hope to put to good use. i love ti's documentation, though. incredibly good.01:11
wpwrak(atmel was just the first i found that looked like something i could expect to make work)01:12
Action: kristianpaul zzzZ01:13
wpwrakwolfspraul: btw, here are the regulations for you: http://www.ti.com/litv/pdf/swra06001:13
wpwrakwolfspraul: some really nice paper on antennas http://www.ti.com/litv/pdf/swra08801:13
wolfsprauldo I have to read those regulations?01:15
wolfspraulgiven that I am in China, maybe I can read them a bit later...01:16
wpwrakwolfspraul: but the killer is this antenna selectio nguide: http://www.ti.com/litv/pdf/swra161a01:17
wpwrakwolfspraul: (regs) not sure. haven't read them myself yet. i hope there are no nasty surprises in there01:18
wpwrakwolfspraul: (antenna selection) e.g., see table 3 in section 4, or section 4.1.201:20
wpwrakwolfspraul: suckish situation, but good documentation of it01:20
wolfsprauljust my personal opinion on how TI works - TI is essentially a patent litigation shop01:23
wolfspraulthey buy smaller companies with strong technology01:24
wolfspraulthen they add their 'legal' / patent litigation services on top01:24
wolfsprauloften times the actual technical innovation stops after the TI purchase01:24
wolfspraulTI just takes the cream from the top with their legal litigation power, suck out what was created technically before, but not fully monetized legally (in the western world)01:25
wolfspraulso that's the TI machine01:25
wolfspraulyou can still buy 'ti' chips (mostly from acquired smaller companies anyway), but over time this stuff will get more and more patented01:26
wolfspraulno wonder they essentially 'gave up' China :-)01:26
wpwrakwolfspraul: there probably isn't much you can get that isn't patented in one way or another01:27
wolfspraulthey just live their parasitic IP life in the western world as long as that's possible. nice for them that this social system for the legal profession was invented :-)01:27
wpwrakwolfspraul: the issue are the "enhancements" that bring you deeper in patent-land01:27
wolfspraulof course, patents are a tool to make money01:27
trebuchetdon't buy ti01:27
wpwrakwolfspraul: (or at least try to lure you there)01:27
trebuchetboycott ti01:27
trebucheti broke my TI after reading a /. story01:28
wolfspraulwpwrak: just remember, I've mentioned this before - my perspective is global on this, I try to be balanced. the patent system you are thinking about essentially only exists in the US and Europe.01:28
wolfspraulthere is an entirely new patent system being started in China now, where Chinese companies are just patenting existing western stuff in China, he he01:29
wolfspraulthe Chinese have learnt the dialectics of those words...01:29
wolfspraulthe number of Chinese patent applications skyrockets, often more or less direct translations of Western patents01:29
wolfspraulhey - WHY NOT? :-)01:29
wolfspraulit's a big market here01:29
wolfspraulso I try to stay out of all this, I see it from a manufacturing perspective01:29
wolfspraulof course I will always prefer unpatented solutions01:30
wpwrakwolfspraul: beat them stupid westeners with their own weapons ;-)01:30
wolfspraulbut the reality is that there are many taxes/customs fees/toll booths/patents/whatever humans have called them01:30
wpwraki think it's important to distinguish between technology we actually care about and what's just an underlying commodity01:30
wolfspraulif I would insist that between me (the manufacturer) and the end customer, there is absolutely no way of some 'financial hook' a 3rd party installs, then I can close my business right way01:31
wpwrake.g., if the RF chip uses some patented fab process, do we really care ?01:31
wolfspraultheoretically yes, practically no01:31
wpwrakexactly :)01:31
wpwrak(hook) yup. that's the stuff to stay clear of01:31
wpwrakfree technology, with enough room to avoid the hooks01:32
wolfspraulI go reverse, from most aggressive enforcers back01:32
wolfspraulnot theoretically finding every last instance of what I consider 'parasitic' hook01:32
wolfspraulso if we go at this reverse, from most aggressive enforcers backwards, mp3/sisvel is my #101:32
wolfspraulthen maybe the mpeg people01:33
wolfspraulafter that I still need to see, open minded01:33
trebuchetWhat is the Ya NanoNote?01:34
wpwraksounds like a good approach, at least for now01:34
wolfspraultrebuchet: when we started NanoNote, we envisioned a series of products, same form factor, with improving technology01:34
wolfspraulBen NanoNote was first, and is shipping01:34
wolfspraulYa NanoNote is the second generation01:35
wolfspraulbasicall the hardware and software hacking that is currently going on around Ben NanoNote will eventually lead to Ya01:35
trebuchetSo should I hold on to this 1200 USD for a bit01:35
wolfspraulfor 1200 USD please buy 12 Ben NanoNotes01:36
trebuchetI want to throw money at Qi for it to stay alive01:36
trebuchetIs it profitable at 99USd01:36
trebuchetor should I donate01:36
wolfspraulpuh. if you are serious, there are many good things you can do with 1200 USD.01:37
kyakyeah.. like 11 Bens and a lot of beer01:37
trebuchetHey the guys at lavabit didn't 'puh' me when I spent some cash at their small shop01:38
wpwrakkyak: 1-2 bens and LOTS of beer :)01:38
wpwrakrivers of beer, gently meandering through the green landscape, towards you ... :)01:40
wolfspraultrebuchet: and now?01:40
trebuchetDammit, some guy is telling me Qi wants to sell all their v1 nanonote stockpiles then bail because there isn't enough money to go into the next, apparently an email on 4chan :|01:40
kyakwpwrak: it's morning here, don't tempt me into beer, or i might waste my day :)01:40
trebuchetwpwrak: beer is nice but nah01:41
wolfspraultrebuchet: what can we do for you now?01:41
trebuchetwolfspraul: Thank you for the wiki link01:41
trebuchetwolfspraul: Ah, is this a developer's channel? Sorry, I was thinking it was semi-casual and user oriented too :\01:42
wolfspraulfeel like at home :-)01:43
trebuchetwolfspraul: You're kind of intimidating :<01:44
wolfspraulaha :-)01:46
wolfspraulwpwrak: reading the antenna PDF a bit, interesting indeed01:46
wolfspraulthe reference PCB antennas are all for 868/915, not 43301:47
wolfspraulthe village telco people also just had a blog post about some of their antenna work: http://www.villagetelco.org/2010/08/v1-3-antenna-testing/01:48
wolfspraultrebuchet: I have an idea for you - for your 1200 USD you can finance the ticket for kristianpaul to visit 27c3 in Berlin this year (he's from Bogota/Colombia, so it's quite a long trip, if he gets a visa at all)02:26
wolfspraulI could probably arrange a place to stay with a friend in Berlin, so accomodation should be covered02:26
trebuchetwolfspraul: I was reading mailing lists, aren't there multiple developers in colombia02:27
wolfspraulyes, correct, quite a few02:27
trebuchetwell wolfspraul, i was talking with some other and they were thinking of arranging a group buy02:27
trebuchetone of them just got his BS in electrical engineering02:27
trebuchetwolfspraul: Is that just accidental? Certain places seem to attract certain tech people, there seem to be a lot of FSF sanctioned gnu+linux distributions that have started in latin america02:28
wolfsprauldo you have a NanoNote?////////02:31
wolfspraul(sorry the '/' are from my crazy kbd driver...)02:31
trebuchetwolfspraul: Not yet. Too much on my mind(major munitions shipment). I will though.02:35
trebuchetAh, have there been any talks about 'auditing'(in whatever form) the hardware in Qi projects... for possible backdoors?02:35
wolfspraulno but I think backdoor-free will become important for us later02:37
sdschulzetuxbrain_zzz: I see. :(((04:09
sdschulzeThe pricing policy of the devboard is strange.04:12
kyakwhat's the price of devboard?04:15
sdschulzeYou pay $100 extra for having a USB host and 32 MB more RAM -- but having no keyboard and no screen.04:16
sdschulzeand no case04:16
zeari guess that's the production costs for smaller batch04:17
sdschulzeMight be, though $199 looks extremely rounded.04:22
sdschulzeAnd it certainly doesn't attract developers.04:24
sdschulzeWhat I want to do is:04:30
sdschulzeI will be moving to a student apartment quite soon and I will take my computer with me.04:30
sdschulzeAt home, I want to leave a machine that doesn't cost too much and doesn't draw too much power but can do some network services.04:31
sdschulzeand is more powerful than a WRT54G04:31
sdschulzeA NanoNote would be ideal -- if it had network.04:35
sdschulze(No, don't tell me to buy an SD NIC)04:35
sdschulze(at least not if it's WiFi)04:38
zearsdschulze, you can plug it to a pc host and forward the internet to it (i know it's useless :D)04:39
sdschulzeThe point is that I want to eliminate the PC part.04:40
sdschulzebecause I'm taking my PC with me04:40
Textmodewhat are you talking about?04:43
sdschulzeabout getting a network connection on a NanoNote without a PC04:44
Textmodeah. yeah, that would be nice.04:44
Textmodeisn't your main option an SDIO card?04:44
sdschulzeMaybe the idea I came up with yesterday wasn't even so bad.04:45
Textmode(but since it has only one slot, that deprives you of a SDcard...04:45
sdschulzeTextmode: AFAIK only for WiFi04:45
sdschulzeI'm moderately skilled at USB stuff...04:45
Textmodeproblem is that while the NN has a USB port, it only operates in slave mode, doesn't it?04:45
Textmodewhich means that most ots usb equipment isn't usable.04:46
Textmodeit should of USB-to-go, then we wouldn't be having any issues :/04:46
Textmodeof been*04:46
sdschulzeTaking a Teensy, a USB hub and a USB-Ethernet adapter should make it.04:47
sdschulzewith like a week of full-time hacking04:47
Textmodewhats the smallest usb hub you can find, and is it powered?04:48
sdschulzeMy primary issue would be: how do I get a Teensy in Europe?04:50
sdschulze$9.30 shipping, not tracked, not insured04:52
Textmodethat sucks :/04:57
sdschulzeonly because two pins are not connected on the board05:02
sdschulzeOtherwise, you could easily access the USB 1.1 hot on the SoC.05:03
sdschulzethe JZ05:13
viriclarsc: is there a memory advantage in having a kernel without modules?06:06
viriclarsc: I finally understood what patches to use to build the kernel I think :)06:07
larscviric: good :)06:30
viriclarsc: I did not understand the openwrt parts about the 'sound module'06:31
viricIt looks like it builds .ko objects, but the kernel is built without module support06:32
viric(accordingt qi_lb60/config-2.6.35)06:32
viricI just noticed qi_lb60/config-2.6.35 is not very complete06:33
viricI'm trying to understand how linux/xburst/config-2.6.35 and linux/xburst/qi_lb60/config-2.6.35 are combined, and if there are more options combined in06:35
larsctarget/linux/generic/config-2.6.35 is the base config06:35
viricand the others are catted ?06:36
viricso I can cat all of them into .config, and 'make oldconfig'?06:36
viricgeneric + xburst + qb_lb60 ?06:36
viricI'm trying that06:42
viriclarsc: I cannot find the sound module...06:58
viriclarsc: I just booted my own built kernel...07:07
viriclarsc: and I cannot switch to virtual console 2 or 3, where the inittab spawns a getty07:07
viricI have VT enabled, VT_CONSOLE too, FRAMEBUFFER_CONSOLE too, ...07:08
viricin fact even power off does not work07:10
viricthe last message I see on the kernel console is "FPU emulator disabled, make sure your toolchain was compiled with software floating point support (soft-float)"07:14
kristianpaulwolfspraul: i'm in Buga\Colombia btw :)09:29
kristianpaulbogota wsa temporary just fot the cparty09:29
wolfspraulsure I know09:42
wolfspraulI just simplified for our donation friend, who I 100% certain think will never donate anything anyway :-)09:42
wolfspraul(is he still here, forgot the nick - if so I'm sure I will get another earful... :-))09:42
wolfspraulkristianpaul: I was just thinking what I would do if someone really wanted to donate 1200 USD09:43
wolfspraulbuying NanoNotes is wasteful09:43
wolfspraulgiving away hardware is also dangerous, I really don't believe in it, whether NanoNotes, or Milkymist, or what not09:43
wolfspraulso the best would be your trip to Berlin! :-)09:43
wolfspraulactually I'm curious - are you confident you would get a visa?09:44
wolfspraulI know this drama from Nelson, I thought should be no problem but I was proven wrong. He couldn't even get a transit visa to fly via Paris!09:44
kristianpaulhmm i can try on qemu isnt?09:44
wolfspraultwo channels :-)09:45
kristianpaulwolfspraul: dont known yet i'll ask a friend (he know nelso i think) who got visa and wen to germnay, so he can tell me more what to expect and to to void problems09:46
wolfspraulI think 27c3 is not realistic09:48
wolfsprauldon't know who can cover all the costs...09:48
wolfspraulmaybe next year when we are all rich :-)09:49
kristianpaulall try my self yeah, plans a next year is better more time :)09:50
kristianpaulat least if i got visa is okay for future use :)09:50
kristianpaulwolfspraul: any way thanks for help :) you are really supportive09:52
wolfspraulkristianpaul: you have to be careful with the visa thing09:57
wolfspraulI learned a lot working with Nelson on this09:57
wolfspraulin Colombia you have to present your passport _HISTORY_, i.e. you current passport as well as your older (expired) passports09:57
wolfspraulthey can trace them back because the new one will list the serial number of the prior one09:58
wolfsprauland now - the icing on the cake - if you have a denied visa in any of your history, chances are very high you will also get trouble with new visas09:58
wolfspraulthat means - you should only apply if you are relatively certain you will pass09:58
wolfspraulif not, you get a nice denial stamp in your passport, and thanks to this 'chaining' system you cannot really escape that anymore09:58
wolfspraulif you don't show up with some of your old passports, of course your visa might get denied right there already09:59
wolfspraulwe have to change a lot of things I'm afraid, sometimes I feel this copyleft hardware thing is only the first mission :-)09:59
wolfspraulso - don't apply just for fun. a visa denial can be quite bad in Colombia.10:00
wolfspraultalk to Nelson or others who have more experience10:00
kristianpaulah i'm stucked i dont have old passports to support visa thing10:01
kristianpaulok i will10:01
kristianpauldamm i tought it was hard but no so problematic :(10:01
wolfspraulyou are young, maybe the passports when you were still a child (!) are exempt10:03
wolfspraulwho knows I am not making these rules10:03
wolfspraulI just went through a painful experience inviting Nelson to Taiwan10:03
wolfspraulin the end we made it but I learned my lesson and it didn't quite go the way I thought it should go10:03
wolfspraulangry letters on my end of course didn't help a thing either :-)10:03
lekernelkristianpaul, so you're trying to go to 27c3 eventually?10:03
lekernelbtw they released the cfp wolfspraul :)10:03
lekerneland we have until october to submit10:04
kristianpaullekernel: yeah that could be a good idea10:04
wolfspraulyes I saw it somewhere10:04
kristianpaulas i said i jsut discover that event last year10:04
wolfspraulkristianpaul: it would be a great event for you indeed10:04
wolfspraulbut need to be patient10:04
wolfspraulvisa + finances, not so easy...10:04
lekernelI've heard the CCC sometimes provides financial help10:05
lekernelbut I'm not sure how and for what10:05
zearwow, that really sucks10:05
zearthe visa trouble10:06
lekernelyou have no idea about the amount of overhead in the world :)10:08
zeari do, i need to have a visa to usa10:08
kristianpaulwhere you are located?10:08
zearthe middle of europe and an EU country10:08
lekernelpoland isn't visa exempt?10:08
zearnope, it's not10:08
lekernelyou might have just to fill in that "electronic travel authorization" crap10:09
lekernelhuh, ok10:09
zearthey are known to stop us on the airport in usa and deny the entry10:09
zeareven if you have a visa10:09
lekernelI thought all EU countries were visa exempt10:09
lekernelmaybe not the eastern ones...10:10
zearbecause they suspect us terrorists10:10
zearthat's retarded, we're europeans10:10
zearlekernel, czech is, just not poland10:10
zearafter helping usa for so many years in afghanistan and iraq, they still require us to have visas..10:10
lekernelyeah, and just before boarding they make you sign retarded declarations that you're not bringing a bomb in the plane ...10:11
wpwraklekernel: even austria isn10:11
wpwrakt entirely10:11
zearseen them10:11
zearwell, at least we can get visas to north korea, unlike americans ;)10:12
Action: kristianpaul need to find what countries me can get easy easilly :)10:13
kristianpaulwolfspraul: http://www.circuitsathome.com/mcu/programming/arduino-usb-host-mini-first-prototype10:14
kristianpaulwolfspraul: may be that can be useefull to give more outputs to the Ben10:14
kristianpaulthere are 1000~ out there so just usb client issue need to be solved some way10:15
kristianpaul**just a tought**10:15
wpwrakthe ben would have plenty of io if it used it a bit more intelligently :)10:16
wpwrakpity that there's no chance for layout changes for the next ben production run. a few "interesting" ios would be easy to recover10:17
kristianpaulah this modules is almost half nanonote is10:21
kristianpaulno good idea10:21
kristianpaulin price*10:25
bartbesis it known that gmu can simply decide to exit/crash when playing certain songs11:12
bartbes(presumably those with cover art)11:12
bartbeseven when it's disabled in the conf11:17
wejpbartbes, no that should never happen12:11
wejpif it crashes, please send me the log output of gmu. gmu is pretty verbose on stdout which you can redirect to a file. this would help a lot12:12
bartbeswejp: I'll do that12:46
bartbesbut I doubt it'll be of any use12:48
bartbesthe part I normally see in the display is not too informative12:48
bartbesoh great..12:49
bartbesI must be doing something wrong12:49
wejpusually it helps finding the reason, or at least it gives a hint where to search. if it doesn't and it crashes on the same files again, you could send me one of those files12:50
bartbesit crashes on all files with embedded cover art I have12:50
bartbessomehow the redirection goes wrong..12:50
bartbesthe file is empty12:51
bartbesand it's not on stderr, I know that for sure12:51
wejpthat is really strange, but i guess i could fix that problem (if it actually exists) if you can give me such a file12:51
bartbesmaybe it hasn't flushed yet?12:51
bartbesbtw, I tried disabling cover art in /etc/gmu/gmu.nanonote.conf is that the right place?12:51
bartbes(because it failed)12:51
wejpyou could run "sync" afterwards12:51
bartbeswhat's the newest version of gmu btw?12:52
wejpbartbes, on the nanonote, you have a user config in ~/.config/gmu which overwrites the global config in /etc/gmu12:52
wejpgmu 0.7.1 is the latest version12:52
wejpso if you run as root it is /root/.config/gmu/gmu.nanonote.conf12:53
bartbesI still run 0.7.012:53
bartbesand disabling it in the user conf helped12:53
bartbes(I was looking for .gmu before)12:53
bartbescould it be something you fixed in 0.7.1?12:54
wejpthe program info screen also tells you which config path gmu uses12:54
wejpso to be sure, check that12:54
wejpalways use the latest version and try to reproduce it with that one, i always fix bug if people report them12:54
bartbesoh I found it in .local and it worked12:54
bartbesright, is there a package available somewhere (you just happen to catch me on windows..)12:55
wejpi don't remember a bug causing a crash with embedded cover artwork though, worked fine for most people12:55
wejpyes, as always, there is a package on my website12:55
bartbeswell this is the song: http://8bc.org/items/music/Henry%20Homesweet%20%20-%20TOASTER.mp312:55
wejpok, thanks12:55
bartbesstill crashes 0.7.112:59
wejpok, i'll test with that files myself13:01
bartbes(that is just an example btw, it fails to play about half of the music I put on it)13:01
wejpthat is really strange13:01
bartbesyes, it's that bad13:01
bartbesat least I know how to disable it13:01
wejpnever heard such a thing from anyone else13:02
bartbesmaybe other people have less cover art ;)13:02
wejpmaybe, but i also have files with cover art that don't cause a crash13:02
wejpbut i'll try to figure out what's going on there13:03
bartbesmaybe it's related to cover art resolution?13:03
wejpgmu should just refuse to load cover images above a certain resolution13:03
bartbesI noticed in all crashes where I bothered to read the resolution was higher than ben's screen resolution13:03
wejpthat shouldn't cause a problem, gmu scales the images anyway13:04
wejpbut the limited memory of the ben is an issue, so gmu refuses to load large images to not run out of memory13:04
wejpat least it should13:04
wejphm, your file works fine here and gmu even loads the cover image just fine :O13:05
wejpit is 450x421 pixels in size13:05
Action: bartbes copies the file to an alternate location13:05
wejpa slice of toast ;)13:05
bartbesindeed it is :P13:06
bartbescan't get it to load here13:07
bartbescan I check for unmet dependencies?13:07
wejphmm, you need libjpeg of course, as in your file the image is of type jpeg, but you should have that one13:08
wejpdoes gmu crash with a segfault?13:08
bartbesit simply.. quits13:09
bartbesthe end looks like this13:09
bartbes"Loading cover image of type image/jpeg form song meta data (tag)...13:09
bartbes"Loader thread.13:09
bartbes"Loading image from memory...13:09
bartbes"image dimensions: 450 x 42113:10
bartbes"fileplayer: UNPAUSE AUDIO!"13:10
bartbes(the others are prefixed by coverimg:)13:10
viricuh. I don't know what to do with the uboot nand part:13:10
viric Checking 454656 bytes... Comparing 454656 bytes - FAIL at off 245, wrote 0x0, read 0x2013:10
wejphm, looks fine so far13:10
wejpreally strange :|13:10
bartbestell me about it13:10
bartbesand as far as I can tell I'm not doing anything strange13:11
wejpand the "fileplayer: UNPAUSE AUDIO!" is the last one?13:11
wejpviric, did you try erasing the whole nand?13:11
bartbeson 0.7.0 it used to end with 1 stop and 2 pause iirc13:11
viricwejp: yes. I did not try that today, but when I tried, I got the same results13:12
viric'whole nand' means, though, uboot+kernel, right?13:12
viricthe command I see takes away 4096 blocks13:12
wejpviric, are you flashing through usb? i had similar problems on one computer but it worked fine on another13:12
viricwejp: yes, through usb13:13
viricwejp: it works fine for the kernel and the rootfs. But it fails on the uboot13:13
wejpviric, then try it on another computer if possible, and do not use a usb hub13:13
viricI don't use a usb hub.13:13
viricI can try another plug...13:13
wejpdo you use a laptop computer? those sometimes use internal hubs13:13
viricyes, laptop13:13
wejphm, do you have another computer around?13:14
viricMm I can find one13:14
viricI'll install the tools there13:14
wejpbartbes, you did not accidentally enable gmu's shutdown timer? ;)13:14
bartbeshow would I do that?13:15
viricwejp: am I right that the command to 'erase the whole nand' erases only kernel+uboot?13:15
bartbesand how would it time this perfectly?13:15
bartbesand, as I said, disabling cover art works13:15
viricwejp: here explained: http://en.qi-hardware.com/wiki/Xburst-tools13:15
bartbeswejp: and it's not too big of a problem, I'm probably not going to look at the cover art anyway, but this is a serious bug I'd say13:16
viricwejp: in another usb port it looks like succeeded13:16
wejpbartbes, with Alt+T13:16
wejpviric, good :)13:16
viricwejp: mmm I want an uboot command line13:16
viricwejp: I did not expect it to work because of the usb port change :)13:17
wejpbartbes, if there is a bug, it should be fixed, so i would like to figure out what's happening there13:17
bartbessame here13:17
viricwejp: I've built uboot, and it boots. I've built the kernel, and it boots, but it halts suddenly after apparently having loaded everything13:17
viricusing 'strings' on uboot I see this kernel commandline:13:17
wejpviric, hehe, me neither, but as i had similar problems which i could solve by using a different computer, i though suggesting that could help :)13:17
bartbesand I tried it on both /data and /card (and yes, that is a memory card), so it's not the fs or fs type13:18
viricbootargs=mem=32M console=tty0 console=ttyS0,57600n8 ubi.mtd=2 rootfstype=ubifs root=ubi0:rootfs rw rootwait13:18
wejpyeah, would be really strange if it was the file system13:18
bartbesbut worth checking anyway13:18
bartbesI guess it can't be a damaged file either, with so many files affected13:18
wejpas it does not crash but just seems to exit, it must be caused by something else.13:19
viricwejp: the last lines of the kernel: "VFS: Mounted root ubifs...  ;   Freeing unused kernel memory ... ; FPU emulator disabled, make sure...."13:19
viricand it even does not shut down after this.13:19
bartbeswejp: yeah, but what? buffer overflows? or is it voluntary?13:19
wejpif the shutdown timer is enabled, you see a number (the remaining minutes) in the upper left corner or "[S]"13:19
wejpit looks like it voluntarily exits13:19
bartbesit kind of does13:20
wejpstill i don't know why13:20
wejpsorry, upper right corner that is13:20
bartbesoh right13:20
bartbesno that's not it13:20
bartbesthat would make its timing exceptionally great though13:21
bartbesloaded decoders13:21
bartbesI can trim those I guess13:21
bartbesbecause some don't even have backends that work13:22
bartbesbut it doesn't seem like that would cause this..13:22
wejpyes, you can simply delete the .so files you don't need13:22
wejpso gmu always exits with that file with cover artwork enabled and does not exit with cover artwork disabled?13:23
wejpok, so at least i have an idea where to look13:24
wejpstill this makes not much sense to me. the only thing i could imagine is that, despite the check for the image size, gmu still runs out of memory and exits because of that13:25
wejpthe image size actually is well below the size limit, though13:25
viriclarsc: the kernel boots and looks hanged... any advice? :)13:25
bartbesbrb, reboot13:26
larscviric: where does it hang?13:28
viricthe last lines of the kernel: "VFS: Mounted root ubifs...  ;   Freeing unused kernel memory ... ; FPU emulator disabled, make sure...."13:29
viricit's 2.6.3513:29
larscyou have a openwrt rootfs and a custom build kernel?13:32
viricI have a debian rootfs13:33
viriccustom build kernel, yes.13:33
viric(not with openwrt)13:33
viricand custom built uboot13:33
viric(sorry :)13:33
viricBut I took all the files and patches from openwrt13:33
viricit looks like the ubifs mounts fine13:34
larscthe openwrt patches change the default init to /etc/preinit13:34
viricAhhh nice13:34
viricI better change it to /sbin/init?13:34
viricI'll take a look13:34
viricif I could get an uboot cmdline... is it possible?13:34
larscvia serial13:35
viricI don't have a serial line..13:35
viricI can rebuild uboot too13:35
viricthe kernel cmdline is there hardcoded13:35
viricI'll check.13:35
bartbeswejp: so, any ideas yet?13:38
viricyesterday I learnt how difficult it is to cross-build perl.13:50
viricvery portable, they say!13:50
kristofferperl is very nasty to crosscompile alright13:52
viricunfortunately, some people success doing that, and then they don't fix anything :)13:52
viricbtw, am I right that I have to take the battery out to exit 'usbboot mode'?13:53
wejpbartbes, not yet, like i said, the only thing that seems possible is gmu running out of memory, so it quits to avoid a crash. if that is the case, and when it is, why there is too little memory available, i don't know yet13:54
viriclarsc: humm I built without applying the preinit patch, and it still hangs there. Wouldn't it warn, if it was not finding init?13:55
bartbeswejp: is there a way to check that?13:56
wejpyou could check how much memory you have left when running gmu13:56
wejpjust run top on another terminal and it should at least give you an idea how much memory is available13:57
wejpif you use another terminal while running gmu, you might want to disable gmu's energy saving feature which turns of the display after a few seconds. you can do that through the config file13:57
bartbeswill do14:00
bartbeswhat the hell? top looks horrible!14:02
larscviric: yes. it should give you an error message14:03
bartbeswejp: the combination of top looking like this and gmu quitting before I even have a change of switching to top makes it impossible for me to test this14:04
wejpno, just run gmu, then switch to top, it is not neccessary to play a file that makes gmu quit14:06
wejpif it quits because of low memory, you must already be short on memory before that actually happens14:06
bartbeswell ehm14:08
bartbestop kind of is unreadable14:08
bartbesI'll try looking at it via ssh14:08
bartbeswejp: ah, ssh makes it readable14:11
wejpgood :)14:11
bartbes31% mem use in standby14:11
bartbes(gmu alone)14:11
wejpok, that's just fine, but the more interessting part is, how much memory is available14:11
wejpthat is, free memory without buffers and caches14:12
bartbesMem: 21004K used, 7476K free, 0K shrd, 604K buff, 4624K cached14:12
wejpok, that looks okay14:12
wejpnot like it is runnign out of memory any minute14:12
wejpmust be something else then14:13
bartbesis it normal for it to have 4 processes (forks, I guess)14:20
viriclarsc: that debian rootfs has /etc/preinit14:21
viriclarsc: totally equal to /sbin/init14:22
wejpbartbes, those are not four processes, to pshow each thread as one process. so yes, that14:23
wejpis just fine14:23
bartbestop shows threads? seriously?14:23
viricI'll try adding cpu emulation14:24
wejpyes, gmu does not fork itself14:24
wejponly busybox's top seems to do that14:24
bartbesright busybox14:24
bartbessometimes my best friends, at other times my worst enemy14:25
bartbes(mostly because they all differ from the 'full' version)14:25
bartbesjust as I *hate* the netcat currently installed on my computer14:25
wejpyeah, but on the other hand busybox is a really great tool14:25
bartbesit provides you with a lot of stuff at low cost14:26
wejpyep, one can get a complete linux system up and runnign that consists of only the kernel and busybox :)14:26
bartbestomsrtbt came close to that14:26
wejpwhen trying to build a system from scratch without busybox, one needs lots of things14:26
larsci think the reason why there are 4 process is because of how uclibc handles threads14:29
wejpah yes, that's possible14:32
viriclarsc: I added fpu emulation, and I could switch the virtual console14:33
viriclarsc: but 'backspace' or 'enter' don't work14:33
larscyou need to load a keymap for the nanonote14:34
viricshould the init scrips do that, or I can set the kernel with a proper default at build time?14:35
larsceither you use the patch from 2.6.34 that changes the keymap inside the kernel14:35
larscor you load http://gitorious.org/mzcode/nanonotes/commit/f741532e6a81ffec5e4f0893fbc2227b84e4d50d from userspace14:36
viricahm ok14:36
viricgreat! You are of a great help14:36
viriclarsc: can you imagine what can be needing FPU emulation there?14:39
viricI think the kernel does not have any float. Right?14:39
larscnope, but userspace i guess14:40
larscopenwrt userspace is compiled with softfloat14:40
larscbut debian not14:40
viricstrange that something related to the boot involves floats.14:41
viricthat hard/soft float is not something written at ELF headers; it's simply different instructions in the text, right?14:41
viricI hope I could get rid of that debian soon :)14:42
viricsomething weird happened14:47
viricNow when I put the battery back, it boots14:47
viricinstead of waiting the power button push14:47
viricand additionally it hangs on 'starting kernel'14:47
viriclarsc: hmmm the kernel 2.6.35 with the 2.6.34 patch does not boot at all14:53
viriclarsc: hmm (4096-1024) means 1,5MiB available for the kernel?15:01
larscthere are 2MB availbale for the kernel15:03
larscpage 1024 to 204715:03
viricah ok15:04
viricWhy the 'total nand erase' erases only 4096 pages? There are more in the nand15:04
viriclarsc: the kernel I have has the size 199099815:04
viric(the uImage)15:05
viricit should be fine I understand15:05
viricsomehow I broke the ubifs...15:05
viricah, total nand erase destroys the first 2MiBs of the ubifs15:06
viriclarsc: the 2.6.35 kernel with the keymap patch does not boot at all (I took the 500-* patch from 2.6.34 of the current openwrt trunk)15:07
larscthats strange15:09
viric Finish! (len 446464 start_page 0 page_num 109) Checking 446464 bytes... no check! End at Page: 10915:11
viricwhy it says 'no check'?15:11
viricgrr. I have problems to boot 'anything' now.15:12
qi-bot[commit] David Kühling: switch to newer upstream snapshot with improved MIPS support. http://qi-hw.com/p/openwrt-packages/d50d5ae15:13
viricwejp: I'm back at the uboot flashing error! grrr. The port that worked, does not work anymore15:19
viricalways a failure at the same byte 204815:22
viricI'm using xburst-tools on x86_64. It should not be a problem I gues15:23
viricwejp: using another computer I get a failure sooner15:27
viricFAIL at off 245, wrote 0x0, read 0x2015:27
viricI'll try with an expensive usb cable15:28
viric a usb2 cable should work only better15:28
viricnow it says:15:31
viric Checking 454656 bytes... no check! End at Page: 11115:31
viricwhy it says 'no check'?!15:32
viricanyway, it worked now I think15:35
viricIt will take me lots of days to learn this nanonote :)15:37
viriclarsc: I tried again, and with the keymap patch it does not boot15:39
viriclarsc: on uboot's Starting linux, the characters get moved some spaces from left to right, and there hangs15:39
viricno linux messages at all15:40
viricso I'll go for the userland tool on keymap15:40
viriclarsc: grr It's my fault about the keymap I think16:15
viriclarsc: patched bad. I forgot the xburst patches for 2.6.35 at all when I tried the keymap patch16:16
viriclarsc: and in fact 2.6.35 already has the modifier keys patch in 2.6.35. I don't know why the backspace/enter keys did not work then16:18
larscviric: no it has not16:30
viriclarsc: I have a more concrete report. It's the right column of the matrix that does not work: p/backspace/enter/vol1/vol2/leftarrow/downarror16:31
viriclarsc: https://dev.openwrt.org/browser/trunk/target/linux/xburst/patches-2.6.35/500-modifier-keys.patch16:31
larscok. thought you were refering to the qi-kernel tree16:31
viricNow I have the keymap you gave me loaded at the init scripts16:31
viriclarsc: I did not go on with the qi-kernel tree because that lacked 'make uImage'16:32
viricI'm using mainline 2.6.35 + openwrt patches. That is not the best to do?16:32
viricI remember that one day you pointed me to use the openwrt-trunk kernel patches. Maybe I understood wrong16:35
viricI'll try to build a 2.6.34 from the openwrt trunk16:36
viriclarsc: when I said leftarrow/downarrow I meant rightarrow/downarrow16:39
viricI suspect that ctrl doesn't work too.16:39
viriclarsc: I can ssh into the nanonote though.16:40
larsccould you dump /sys/kernel/debug/gpio ?16:40
viricHm I don't have that path16:42
viricI have to mount the debugfs iirc16:42
larscmount -t debugfs none /sys/kernel/debug16:43
viricI have the 2.6.34 built. May I try it? I can put the 2.6.35 back if you want later16:44
viricin 2.6.34 the keyboard works fine.16:46
viric(well, it's
larschm, yes one gpio seems to be wrong16:51
larscgpio-121 (matrix_kbd_row      ) in  hi irq-177 edge-both16:51
larscthis should be gpio-12016:51
viricand now where will the fix go?   qi-hardware kernel 2.6.35? openwrt-trunk patches?16:52
viricI want to test the sound of the nanonote. Is there anything I should know? I'd go with alsamixer and madplay16:54
larscsound is pretty loud ;)16:55
viricthe volume keys will work?16:55
larscpatch is commited in openwrt-trunk16:55
larsci don't think so16:55
viricalsamixer will work?16:55
viricIn the openwrt I tried, it did not16:55
viricAnd in the debian, there was no sound card support16:55
larscalsamixer should work if the sound module are loaded16:56
larscor sound support is builtin16:56
viricI put it on the kernel16:56
viricgrrr I don't know in what debian package alsamixer is16:56
viricalsa-utils I hope16:57
viricoh, it looks like working!16:57
larscgreat :)16:59
viricI'll put a mp3 file there now17:00
viricIt plays!17:01
viricAlthough the minimum volume is quite loud, and the maximum louder.17:01
viricah, with *muted* master it still sounds, although very (the opposite of loud)17:02
viriclarsc: how accurate are the cpu usage counter? I get during madplay a 20% user, 80% idle, loadavg of 0.0417:03
larscno idea17:07
viricis the loudness of the speaker a software issue?17:08
viriceven headphones I cannot use :)17:08
larscnope hw17:08
larscbut could be 'fixed' by using a softvol17:09
viricaha, I understand17:09
viricor an alsa plugin?17:10
viricor a proper alsa config, I don't know much17:10
viricfor the program not to need that specially.17:10
tuxbrain_awaylarsc: on the testing software the volume has 7 from 0 to louder so I thing is just soft issue17:10
tuxbrain_away7 levels17:10
larsctesting software?17:11
tuxbrain_awayyes the software what nano comes before the OpenWrt17:12
larscnever seen it17:12
larscbut i can tell you for sure that the hw only has 4 volume levels ranging from loud to very loud17:13
tuxbrain_awayit has a simple menu to test mic, screen, speakers, keyboard ....17:13
tuxbrainI don't know if wolfspraul has the source of it17:14
viricalsamixer shows 4 levels17:17
viricI'm not having luck writing the softvol .asoundrc17:17
virictuxbrain: do you use it?17:17
viriclarsc: oth, do you know where is the code for the SIMD instructions in mplayer?17:19
viricI read it was a big file in hexcode, because xburst did not want to publish the SIMD instructions17:19
tuxbrainvirc: I have use it on my first nano, but then I reflash with OpenWrt and the next once also comes with Openwrt installed17:25
tuxbrainviric: GMU player can manage the volume to 0(silence) to 15(louder)17:26
viricCan you paste your .asoundrc somewhere?17:27
viriclarsc: on 'ps aux' I get a sum of RSS of about 3 or 4MB17:27
viriclarsc: but 'free' reports 12MB used. Is that by some kernel option I should not have used?17:28
viriclarsc: or... how could I know the concerning kernel option?17:28
viric(12MB used not counting the buffers/cache)17:28
viricthere are linux around on 8MB systems...17:39
qi-bot[commit] Werner Almesberger: Added problem statement and described more details of the process. http://qi-hw.com/p/ben-counterweight/d497d6717:42
viriclarsc: I was talking about this, FYI ftp://ftp.ingenic.cn/3sw/01linux/05apps/mplayer/17:46
kristianpaulnah back to openwrt :)20:33
tuxbrainwejp: about lyrics on gmu, it has to be a separates files with same name of the of the song file in the directory, with txt extension?20:36
wejptuxbrain, either that, or for mp3 files they can be embedded into the file, but this obviously does not apply for the nanonote22:56
wolfspraulwejp: sorry for the confused question - are you the original gmenu2x or gmu author?22:59
kristianpaulHanvon N516 e-book reader23:01
wolfspraulI saw the new GMU icon from Mattes and looked around a bit where or how this would need to be committed. Which made me realize I have no idea where upstream is :-)23:01
kristianpaulwhy is this device in openwrt-xbusrt?23:01
wolfspraulthe gmenu2x project on qi-hardware says 'temporary location for the Ben NanoNote', but there seems to be quite a bit of work going on there23:01
wolfspraulkristianpaul: it's another device with Ingenic's 4740 CPU (same die as in Ben NanoNote)23:02
wolfspraulat some point I tried to get more datasheets out of Hanvon, but failed23:02
kristianpaulbtw i'm using openwrt in my nanonote as default23:02
wolfspraulit's also one of the main devices for www.openinkpot.org23:02
kristianpauljlime for sd23:03
kristianpaulnow i need find out how implement bit banging in kernel, seems well documented23:03
--- Sun Aug 8 201000:00

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