#qi-hardware IRC log for Wednesday, 2011-01-05

qi-bot[commit] Xiangfu Liu: configu.full_system: remove IB and Toolchain http://qi-hw.com/p/openwrt-xburst/e26ce4303:31
qi-bot[commit] Xiangfu Liu: using 'strip' instead of 'sstrip' http://qi-hw.com/p/openwrt-xburst/257e48a03:31
bartbeshow does one get commit access to some openwrt repos?04:25
bartbesmaybe I could keep a fork somewhere in the mean time04:26
bartbesbut still04:26
bartbesbtw, do you guys now a libintl?04:30
xiangfubartbes: you mean the "packages" of OpenWrt?05:10
xiangfudon't know that. need talk to OpenWrt people.05:13
xiangfuthe "Stub" libintl, libiconv really not good for nanonote.05:14
xiangfumaybe they should add one option about "Stub" libs. like: "PATENTED" option.05:16
bartbessorry, could you rephrase that?05:17
bartbesdo you mean libiconv is the replacement for libintl?05:18
xiangfubartbes: no. maybe I misunderstand you. let me talk my problem first.05:19
xiangfuthere are two libintl packages in openwrt:05:19
xiangfulibintl                  Stub header for the GNU Internationalization library05:19
xiangfulibintl-full             GNU Internationalization library05:19
xiangfuthe "Stub" libintl break  other packages compile. which make me can not compile a new openwrt image.05:20
bartbesso I need libintl-full05:21
xiangfuyes. then you need change packages "Makefile"05:23
xiangfubecause there are something like: "-I$(STAGING_DIR)/usr/lib/libiconv/include" need change to "libiconv-full"05:23
qi-bot[commit] Xiangfu Liu: new pakcage: mediatomb, open source  UPnP MediaServer http://qi-hw.com/p/openwrt-packages/47496a706:07
xiangfubartbes: http://erikugel.wordpress.com/2010/12/26/mediatomb-on-the-ben-nanonote/06:32
kyakxiangfu: hi! do you have an idea how to override packages from openwrt feeds? it seems that packages from openwrt-packages don't override packages from openwrt feeds. However, they can override packages from base system07:05
bartbeskyak: rename ;)07:05
kyaki think we are stuck at this libiconv/gettext problem.07:05
kyakbartbes: rename won't work in this case07:06
bartbesI know, but that is the solution to prioritize07:06
kyaka lot of package depend on libiconv/gettext, not on renamed versions07:06
bartbesmaybe you should talk to one of the owrt devs07:06
bartbesso they can at least tell you about the behavior of their build system07:06
xiangfuI am ask "<xMff>" iin #openwrt-devel07:07
kyakthat's why i asked xiangfu07:07
kyakxiangfu: thanks!07:07
lekernelmm1 jtag pods: http://en.qi-hardware.com/wiki/File:M1_jtagserialpod_run1_trays_pcba_to_be_tested.JPG07:07
bartbesoh, is xiangfu an owrt dev as well?07:07
lekernelnice pic! :)07:07
xiangfubartbes: me,, nanonote openwrt developer :)07:08
kyakwell, we have very experienced openwrt persons in here, including xiangfu07:08
kyaki think it's not bad asking here before07:09
bartbesbut technically it has nothing to do with the nn07:09
wpwraklekernel: two errors: 1) that should be MM1s, 2) they should be stacked on the z-axis, too :)07:09
bartbeslars-c (to prevent hilight :P) helped me before with the build system07:09
kyaktechnically openwrt devs don't care about nn and about supporting of desktop options07:10
kyaktechnically we are on our own07:10
lekernelwpwrak: well we also have to find people who buy them07:10
bartbesnot quite07:10
xiangfu<xMff> I thought about the nanonote deployments breifly when ripping out i18n, but as it does not reduce functionality compared to before I just did it07:10
bartbesthis is a problem regarding their build system07:10
xiangfuout for awhile, back in ~30 mins07:11
kyakwell, fbterm doesn't compile. i don't know about rest07:11
kyakand it doesn't compile because libiconv is ripped07:11
kyakand it also causes a lot of other problems.07:11
kyakevery single commit to libiconv for the last two months is "unstubbing the stubbed"07:12
kyakso they stubbed it07:12
kyakand then started adding some fucntionalities when they see problems with a package07:12
wolfsprauladam will send a proper mail, but ahead of that he told me 97 of 100 tested fine07:15
wolfspraulah no, 9607:15
wolfspraulbut one was just reflashed with the wrong pid 0731, that can be corrected07:15
wolfspraul3 with smt issues, he will upload pictures07:15
bartbeskyak: well, I found one of their own packages that they broke07:16
bartbesbut first I need to get libintl-full to compile07:16
wolfspraulall avoidable (theoretically), I think iqc issues again, but I'm blaming WeyJia (smt corp) less than Adam here, because they are under enormous economic pressure.07:16
wolfspraulso I think it's fine. 3 bad ones, just toss.07:16
bartbes../intl/.libs/libintl.so: undefined reference to `libiconv_set_relocation_prefix'07:20
kyakbartbes: the problem is, that libintl-full will be compiled AFTER the packages that requires libintl. And this package will most likely not build against ripped off libintl07:20
kyakyes, just like this..07:21
bartbesoh I modified that07:21
bartbesI added libintl-full to the dep list07:21
kyakyeah, you modified it for 300+ packages that require libintl?07:21
kyakyou are talking about your app only.07:21
bartbesno, for the one I'm building07:21
bartbesbut it doesn't even compile07:22
kyakand it won't07:22
bartbesI mean libintl-full07:22
kyakbecause they install it into other directory07:22
bartbesthat I can all change07:23
bartbesI don't care about reproducing it for the moment07:23
kyakok, you can keep thinking about your app only07:23
bartbesyou make me sound bad07:23
bartbesI just want to get this to work07:23
bartbesto prove it's possible07:23
bartbesI know there's a deeper problem07:24
kyakok, i'm sorry07:24
bartbesI mean, I got the entire EFL already07:24
bartbesI just want to get me some e17 on my nn07:25
kyakspeaking about that, do you have X running?07:25
bartbesthe EFL was easy, really, I just needed to change some flags to make it use fb and directfb, and don't use X07:25
kyakor e17 is running in fb?07:25
kyaki see07:25
kyakthat's great, hope it works07:26
bartbesme too07:26
bartbesit has quite a few backends07:26
bartbesbut I used fb and directfb07:26
qi-bot[commit] David Kühling: gnuplot: fix ggi terminal video mode detection http://qi-hw.com/p/openwrt-packages/21acf8e07:32
qi-bot[commit] David Kühling: gnuplot: Build and package help files for internal documentation system http://qi-hw.com/p/openwrt-packages/59ad4ea07:32
wpwrakwolfspraul: 96-97% yield for a first run sounds tolerable :)07:32
bartbeswhat are they, btw?07:33
wolfspraulsure but we want 100, right? :-)07:34
wolfspraullet's wait what the report says, and pics07:34
bartbeskyak: gettext-full failing to compile doesn't seem to be helping..07:34
wolfspraulAdam was super pissed about some of the mistakes, so there seems to be something he thinks can escalate easily.07:35
wolfspraulit's very difficult to focus the limited resources exactly on the right spots.07:35
wolfsprauleconomic resources I mean, time and money07:35
wolfspraulalong those lines, I saw a pretty interesting post on our planet this morning... http://www.funcubedongle.com/?p=34207:36
wolfspraulnasty little things... hard to predict, catch, fix all of this effectively.07:36
wpwrakcracked caps. ugly.07:39
qi-bot[commit] David Kühling: gnuplot-ggi: build without mouse support so we can leave plots via keypress http://qi-hw.com/p/openwrt-packages/8a1f0d407:46
qi-bot[commit] Xiangfu Liu: using STDOUT_FILENO, remove tcsetpgrp http://qi-hw.com/p/gmenu2x/bfe772b09:12
wolfspraulwpwrak: there's some of the pics http://en.qi-hardware.com/wiki/Special:NewFiles09:14
wolfspraulsomeone had fun with the cutter :-)09:14
wpwrak"U5 direction wrong" sounds terminal09:19
wpwrakwhich one would be the cutter ?09:19
zearoh nice, gmenu2x progress09:20
qi-bot[commit] Xiangfu Liu: tunec: remove @BROKEN, fftw3 fixed in upstream r24169 http://qi-hw.com/p/openwrt-packages/48e266a09:27
qi-bot[commit] Xiangfu Liu: gmenu2x: update http://qi-hw.com/p/openwrt-packages/134dbbb09:27
virichas anyone made a nice-looking guitar tuner?09:31
viricI only used that small console program, fftw3 based.09:31
wolfspraulI don't blame them. This is an economic challenge, not a tech challenge.09:46
wolfspraulthose who don't understand that don't understand manufacturing :-)09:46
wolfspraulmaybe you could also cut the pcb with a programmed laser cutter?09:47
wolfsprauldon't know...09:47
viricwhat is v-cut?09:47
wpwrakoh, i see. these cuts are quite ugly.09:48
wpwraka laser cutter sounds nice. minimum mechanical stress.09:48
wolfspraulah nice, laser is just starting http://en.wikipedia.org/wiki/Depanel#Main_depanel_technologies09:49
wolfspraulviric: the v-cut is a groove line in the pcb, between pcbs - basically the separation line09:49
viricstudents of electronic engineering should receive points for reading this channel.09:50
wpwrakviric: harsh reality points :)09:51
viricwolfspraul: I'm about to make a release of the rss reader09:52
wolfspraulxiangfu was a little sick, but now he is fully recovered and back at hacking :-)09:53
wpwrakwolfspraul: for protruding connectors, you could pre-cut the board such that it's already completely separated under the connector or maybe even the entire side of the connector (if there's only one). that should make depanelization quite straightforward.09:53
wpwrakah, those new year hangovers ... :)09:53
viricyes, nothing new :)09:54
wolfspraulthey didn't have (chinese) new year yet09:54
wolfspraulthat's also 10 times crazier than anything I've ever seen in the rest of the world09:54
viricah, those week end hangovers ...09:54
wolfspraulso we have to give him a week off for that...09:55
wolfspraulgood thing that hearing loss doesn't matter much in hacking09:55
wpwrak(cny) ah, i thought it was a more sedate, family-oriented event. so it's fun then. good :)09:56
wolfspraulviric: if you know electronic engineering students, point them here09:56
wolfspraulunfortunately manufacturing is still considered dirty and lowish, big mistake if you look at the billions and billions apple is piling up, for example :-)09:56
wpwrakparticularly those with RF skills :)09:56
wolfspraulwpwrak: 'sedated'? you must be kidding...09:57
viricwolfspraul: I told 'page' to come in, and he came. We work together; not students anymore :)09:57
viricWell, to be honest, I'm still counted as an EE student I think. I did not finish. hehe09:57
wpwrakwolfspraul: there's a big difference between "sedate" and "sedated" ;-))09:57
viricboth page and me are graduated (mostly) as Electronics and Telecommunications engineers09:57
viricbut at least me, I'm ashamed for my lack of practical skills.09:57
wolfspraulyeah, I just wanted to say - you don't know what a v-cut is?09:58
wolfspraulyou can send your profs a big THANK YOU card :-)09:58
viricwell, they will say that "if you had come more often, and when you come, less drunk..."09:58
viricbut hey, I did not have bad points on exams.09:59
wolfspraulviric: hey, that means you can also help with Xue schematics review :-)09:59
viricI can't say no to taking a look.10:00
wolfsprauldownload the sources, use dsv to download the datasheets, and start peeking over the schematics10:00
wolfspraulif you don't have kicad, you can just use a pdf viewer10:01
viricI have kicad sure10:01
wolfsprauloh cool!10:01
wpwrakstill not quite ready for serious review yet, no ?10:01
wolfspraulI don't know.10:02
wpwrak"Design in progress!!!" all over the place10:02
wolfspraulAndres worked a bit over the holidays, not sure how much headway he made.10:02
wolfspraulmaybe he tries to fight off too harsh feedback like that ;-)10:03
wolfspraullike the google 'beta'10:03
wpwrakyeah, it has this kind of ring to it :)10:03
wpwrakwhee, qualcomm snatches up atheros. now, which of the two is less open ? :)10:04
wolfspraulah that's news, interesting.10:06
wolfspraulmore consolidation, that's not surprising.10:06
wolfspraulthere will be lots more.10:06
viricI agree.10:06
wolfspraultoday I read Samsung will invest 30 billion EUR in 2011!10:07
wolfspraulthe numbers are getting higher and higher. unbelievable.10:07
wolfspraulthey have a fantastic business.10:07
wpwrakyou should ask them to put a modest 0.1% into qi-hw :)10:07
wolfspraulwow. that would be... 30 million EUR!10:08
wpwrakalmost nothing10:08
wolfspraulSamsung totally understands manufacturing.10:08
wolfspraulbut unfortunately, unless they see a reason, they won't do the first step towards more open. :-)10:08
wolfspraulbut that's ok. they are not against it either, definitely not. So when this secrecy nonsense stops, Samsung can just flip the switch and be open.10:09
wolfspraulbut with those huge numbers it will be hard to convince them to take the risk, and I understand that.10:10
viricI try to find the xue git project10:10
wolfspraulgit clone git://projects.qi-hardware.com:xue.git10:10
wpwrak(hard to convince) yeah, legal risks ...10:11
DocScrutinizer(towards more open) ask raster10:11
DocScrutinizerqualcom atheros merger sounds :-/ to me anyway10:14
wolfspraulDocScrutinizer: why?10:15
wolfspraulAtheros was small, had to happen. There are many many more lined up.10:15
wolfspraulthe capital and investment pressure is enormous.10:15
DocScrutinizerreducing vectors to leverage the whole thing10:15
wolfspraulI'm sure at 50% of independent IC companies, some managers are working on m&a options, so they can always pull one out if big investments become necessary.10:16
wolfspraulwhat can you do...10:16
DocScrutinizerit's more fun to compete - been written on all pinball machines10:17
wolfspraulfun, for whom?10:17
wolfspraulif you need 500 million USD to invest, and you don't have them, what's fun about that?10:17
DocScrutinizerin this particular case for us, when they do10:17
DocScrutinizerwell, OTOH if you bribed *one* dude at *one* Qual-theros, now you got more for same price :-)10:19
DocScrutinizerand we've really not seen atheros and qualcom competing in regard of openness10:20
wolfspraulnobody is competing on openness10:20
wolfspraulwell, except for the marketing people here and there :-)10:20
DocScrutinizerhmm, intel at least pretends to10:20
wolfspraulDocScrutinizer: http://www.technewsworld.com/story/Intel-Builds-Sandy-Bridge-With-a-DRM-Tollbooth-71568.html10:21
wolfspraulyou were one of the DRM fans, weren't you?10:21
DocScrutinizerYEAH - my favourite (for a target in the shooting range)10:22
wolfspraulso Intel pretends what?10:23
wolfspraul"upgradeable hardware"10:25
DocScrutinizerDUH, killswitch. Thanks for the link10:35
rohheh. yes. its time for trustable south/northbridges/soc ;)10:35
wpwrakintel consider that, in some areas, a higher than average degree of openness helps them furthering their goals. i think that's an attitude we can agree with.10:36
wpwrakwolfspraul: time to make the MM0.5. then sell the upgrade to MM1 in half a year ;-)10:36
wpwrakwolfspraul: of course, since the MM1 was already announced, that would be "MM1" and "HX+MM3000"10:38
wpwraktoday seems to be dedicated to the celebration of the Perfect Immobility of the Holy Coach Potato, at customs. my digi-key stuff is still sitting there, since 1:41 am :-( meanwhile, the bens have entered brazilian airspace10:42
ctZorgNotZerghi I was hoping someone could answer some questions about sandy bridge procs?15:02
kristianpaullarsc: great (new page) !16:28
kristianpaulIs nice know how to do this separatelly :-) (at least for me)16:29
larsckristianpaul: good to know that is useful :)17:18
wpwrak(kernel build page) great, thanks !18:00
qi-bot[commit] Lars-Peter Clausen: Add gpio chager driver http://qi-hw.com/p/qi-kernel/77c61e919:59
qi-bot[commit] Lars-Peter Clausen: MIPS: JZ4740: qi_lb60: Add gpio-charger device http://qi-hw.com/p/qi-kernel/49b2a9519:59
qi-bot[commit] Lars-Peter Clausen: Add n516 board support http://qi-hw.com/p/qi-kernel/d022d3f19:59
qi-bot[commit] Lars-Peter Clausen: Add n526 board support http://qi-hw.com/p/qi-kernel/a8cc3ae19:59
qi-bot[commit] Lars-Peter Clausen: Add N516 sound SoC board driver http://qi-hw.com/p/qi-kernel/8ed027619:59
qi-bot[commit] Lars-Peter Clausen: Add N526 sound SoC board driver http://qi-hw.com/p/qi-kernel/04bfa7b19:59
qi-bot[commit] Lars-Peter Clausen: i2c: Add i2c driver for JZ47XX SoCs http://qi-hw.com/p/qi-kernel/8f1beb319:59
qi-bot[commit] Lars-Peter Clausen: MIPS: JZ4740: Fix i2c driver name. http://qi-hw.com/p/qi-kernel/302b0ba19:59
qi-bot[commit] Lars-Peter Clausen: MIPS: JZ4740: Add id800wt board http://qi-hw.com/p/qi-kernel/02f99c319:59
qi-bot[commit] Lars-Peter Clausen: MFD: jz4740-adc: Add support for the touchscreen part of the adc unit http://qi-hw.com/p/qi-kernel/db0d9c619:59
qi-bot[commit] Lars-Peter Clausen: input: Add touchscreen driver for the JZ4740 SoC http://qi-hw.com/p/qi-kernel/dc7f0a519:59
qi-bot[commit] Lars-Peter Clausen: Add defconfig http://qi-hw.com/p/qi-kernel/a7f059a19:59
qi-bot[commit] Lars-Peter Clausen: Framebuffer notifier: Call notifier callbacks prior to blanking the screen http://qi-hw.com/p/qi-kernel/a7ce78119:59
qi-bot[commit] Lars-Peter Clausen: qi_lb60: Don't use 3-wire spi mode for the display for now http://qi-hw.com/p/qi-kernel/d39fdfb20:00
qi-bot[commit] Lars-Peter Clausen: POWER: jz4740-battery: Protect against concurrent battery readings http://qi-hw.com/p/qi-kernel/bb050fc20:00
qi-bot[commit] Lars-Peter Clausen: POWER: jz4740-battery: Check if platform_data is supplied http://qi-hw.com/p/qi-kernel/c69da0220:00
qi-bot[commit] Lars-Peter Clausen: MTD: NAND: jz4740: Make 'struct platform_driver jz_nand_driver' static http://qi-hw.com/p/qi-kernel/bbc397420:00
qi-bot[commit] Lars-Peter Clausen: MTD: NAND: jz4740: Remove custom {read,write}_page handlers http://qi-hw.com/p/qi-kernel/1c3909220:00
qi-bot[commit] Lars-Peter Clausen: MIPS: JZ4740: Set nand ecc offsets for the qi_lb60 board http://qi-hw.com/p/qi-kernel/1bba17a20:00
qi-bot[commit] Lars-Peter Clausen: mfd: Convert jz4740-adc to new irq_ methods http://qi-hw.com/p/qi-kernel/b77ba0c20:00
qi-bot[commit] Werner Almesberger: cameo: array steps can now be specifed as item size plus border http://qi-hw.com/p/cae-tools/ee3a3e621:15
qi-bot[commit] Werner Almesberger: cameo/path.c (path_connect): corrected logic when merging B->A http://qi-hw.com/p/cae-tools/4b717ca21:15
qi-bot[commit] Werner Almesberger: atusd.brd: removed short edge segment left under larger segment http://qi-hw.com/p/ben-wpan/e2c5d4221:15
qi-bot[commit] Werner Almesberger: atusd/cam: updated to use cameo http://qi-hw.com/p/ben-wpan/6ccb25221:15
--- Thu Jan 6 201100:00

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