**** BEGIN LOGGING AT Fri May 28 03:00:14 2010 May 28 06:06:58 goodmorning May 28 06:08:29 tsjsieb: hello May 28 07:22:47 03Roger Monk  07org.openembedded.dev * r2b363197c0 10openembedded.git/recipes/u-boot/ (u-boot.inc u-boot_git.bb): (log message trimmed) May 28 07:22:47 u-boot: Install UBOOT_BINARY in /boot, add package for -fw-utils May 28 07:22:47 * Drop binary in target /boot folder to allow compatible set of May 28 07:22:47 boot files to be installed in target image alongside kernel, etc May 28 07:22:47 * Create new package for the firmware utils May 28 07:22:48 * Bump PR May 28 07:22:49 Acked-by: Denys Dmytriyenko May 28 07:22:50 03Chase Maupin  07org.openembedded.dev * r8c348ea2a4 10openembedded.git/recipes/ti/ (3 files in 3 dirs): (log message trimmed) May 28 07:22:50 matrix-gui-e_svn: Changed executable name May 28 07:22:51 * Upgraded to the latest source revision May 28 07:22:51 * Changed the executable name in the recipe and scripts May 28 07:22:52 to allow it to co-exist with the X11 version to be added May 28 07:22:52 later. May 28 07:22:53 Acked-by: Denys Dmytriyenko May 28 07:22:56 03Chase Maupin  07org.openembedded.dev * r4a63e4ba9e 10openembedded.git/recipes/ti/matrix-gui-e_svn.bb: May 28 07:22:56 matrix-gui-e_svn: add init script to runlevel 3 May 28 07:22:56 * only add the matrix-gui-e init script to run level 3. May 28 07:22:56 Acked-by: Denys Dmytriyenko May 28 07:22:56 Signed-off-by: Chase Maupin May 28 07:23:06 morning May 28 07:23:08 03Chase Maupin  07org.openembedded.dev * rc528aa3b9d 10openembedded.git/recipes/ti/ (4 files in 3 dirs): (log message trimmed) May 28 07:23:08 matrix-gui_svn: X11 version of matrix GUI May 28 07:23:09 * Added recipe to build matrix GUI application with Qt X11 May 28 07:23:09 libraries. May 28 07:23:10 * Changed init script to not place an entry in the /etc/rc5.d May 28 07:23:10 directory. Init script is still used for starting the May 28 07:23:11 matrix_gui application. May 28 07:23:11 03Roger Monk  07org.openembedded.dev * rb4aa462e12 10openembedded.git/recipes/angstrom/ (2 files in 2 dirs): May 28 07:23:12 (17 lines omitted) May 28 07:30:44 good morning May 28 08:07:21 khem, u here? May 28 08:08:28 I made an image and "nm" is installed in "/usr/arm-angstrom-linux-gnueabi/bin/nm" May 28 10:01:13 morning May 28 10:54:39 hi May 28 10:55:12 I use a gentoo chroot because my laptop is dead and I have (while bitbaking a new recipe: scummvm 1.1.1) : May 28 10:55:13 The working tree was created by an older version of quilt. Please run 'quilt upgrade'. May 28 10:58:10 it seem that PaulePanter got the same issue May 28 11:01:10 your system has newer quilt then OE one? May 28 11:03:12 bitbake quilt-native ? May 28 11:03:24 hi GNUtoo, hrw, others May 28 11:03:36 ah ok May 28 11:03:40 thanks a lot hrw May 28 11:03:45 I'll use oe's quilt May 28 11:09:01 hi everyone May 28 11:10:11 chouimat, hi May 28 11:10:12 mickey|office, ping May 28 11:10:40 * chouimat lost or someone stole his n900 ... May 28 11:10:47 mickey|office, I have seen you mentioning somebody ran fso on a htcpassion / nexus one May 28 11:11:29 chouimat, sucks.. I remember when I _thought_ I lost my htc kovsky May 28 11:12:59 dcordes: nexus is htc passion? I thought that htc desire and nexus one are same hw May 28 11:13:01 dcordes: I will go back to the pub tonight, maybe they saw it, I heard beeps last night but they weren't long enough to help localise it ... it doesn't show on my dhcp server. the worse is might have to order a new one May 28 11:13:27 hrw, yep the naming is true pita May 28 11:13:39 hrw, too many names for each devices May 28 11:15:25 dcordes: yes, some folks made phonecalls with FSO2 on nexus May 28 11:15:26 hrw, passion and bravo (desire) are both qsd8250 based. I need somebody with these devices to try OE builds May 28 11:15:36 mickey|office, did they base on OE builds ? May 28 11:15:56 no, probably rather debian May 28 11:18:03 chances are I'll get a nexus with the task to put OE and FSO2 on it May 28 11:18:07 then I'd add it to OE May 28 11:19:18 mickey|office, you know my oebug right. passion (nexus) has same chip as leo May 28 11:24:05 what does ${datadir} point to by default? May 28 11:24:14 dcordes: i see, so that means i will have problems with OE on nexus? May 28 11:28:20 mickey|office, unless the observed problems on htcleo are kernel related you will run into same thing on passion as both are qsd8250 May 28 11:29:18 mickey|office, I think they are not. kensan booted a minimal-image initrd I compiled for leo on passion and he said it rebooted May 28 11:29:36 mickey|office, unfortunately he could not provide any details yet May 28 11:32:32 mickey|office, but I'm on it with khem's help hope we can find the answer soon May 28 11:33:46 ok, cool May 28 11:35:27 mickey|office, and it would be nice to see you get some snapdragon device. together with Kensan's passion and NetRipper's and my leo it will make us 4 May 28 11:35:43 ok time to go to work, last day for this contract ... will try to finish everything and not think about the lost n900 like yesterday ... not sure the next customer will appreciate I bill then twice for a n900 ... May 28 11:36:21 chouimat, good luck getting the job done and n900 back :) May 28 11:37:25 dcordes: how fast snapdragon is? can you compare it with omap3? May 28 11:37:38 dcordes: it wasn't in the case yesterday morning, and since the case as a plastic cover that goes over the keyboard I don't see how it can get out of it by itself May 28 11:38:11 dcordes: I will probably find it after I clean the huge mess that is called my appartment ... May 28 11:38:41 hrw, I compiled mplayer with neon support and it's very fast. I can _almost_ decode 720P video realtime May 28 11:38:57 hrw, without making any use of the very fast dsp.. May 28 11:39:19 nice May 28 11:40:39 hrw, htc leo (hd2) has 1GHz clocked cpu and 512MB ram May 28 11:46:03 * dcordes holds breath May 28 11:46:12 crap May 28 11:46:53 khem, build with the posted patch succeeded but still init crash. May 28 11:57:00 mickey|office, do you know where those fso2 on nexus folks can be found? May 28 11:57:07 gm May 28 12:02:43 !oebug 5435 May 28 12:02:44 * * Bug 5435, Status: UNCONFIRMED, Created: 2010-05-19 14:26 May 28 12:02:45 * * lukas(AT)htc-linux.org: machine htcleo: gcc version etc. May 28 12:02:46 * * http://bugs.openembedded.org/show_bug.cgi?id=5435 May 28 12:03:22 dcordes: #openmoko-cdevel, deubeleu May 28 12:06:47 there must be some detail, something the snapdragon core doesn't understand May 28 13:28:55 hi... any reason not to http://paste.debian.net/75182/ May 28 13:30:21 i mean, it's pretty unlikely that the BUILD_ARCH or BUILD_OS changes while we're running, i guess. Live-migration of userspace processes across arches is NIH, AFAIK ;) May 28 14:45:15 If you have a SRC_URI with a file://abc.patch type reference, at what stage is the patch applied? Does the patch need to have it's md5 hash in a file along with it? May 28 14:47:40 whittenburg1: I don't think it needs a checksum May 28 14:52:52 My package builds ok, but it seems to be ignoring the patch file, so I was trying to find the log for which step it should get executed and see why it's ignored. May 28 14:53:12 I thought maybe a lack of md5 file caused it to be skipped. May 28 14:57:17 whittenburg1: checksum is fatal and would be reported loudly. did you put the ;patch=1 on the end of that file url? or replace 1 with whatever the level is May 28 14:57:41 ah. no. May 28 14:57:58 SRC_URI = "svn://gforge.ti.com/svn/gstreamer_ti/branches/BRANCH_DDOMPE/gstreamer May 28 14:57:58 _ti/ti_build;module=ticodecplugin;proto=https;user=anonymous;pswd='' \ May 28 14:57:58 file://${GST_TI_RC_SCRIPT} \ May 28 14:57:58 " May 28 14:57:58 SRC_URI_append_omap3 = " \ May 28 14:57:59 file://gstreamer-ti-ddompe-add-omapfb.patch \ May 28 14:58:07 That's probably it then. May 28 15:40:42 17:40 hrw@home:openembedded$ git pull May 28 15:40:43 Permission denied (publickey). May 28 15:40:48 someone? May 28 15:43:07 works for me :) May 28 15:44:54 ah.. this is git proto May 28 15:45:34 ok for ssh too May 28 15:45:37 Jay7: are you the one working on tosa kernel? May 28 15:46:10 undrwater: no, I just building angstrom for tosa May 28 15:46:16 and do some testing May 28 15:46:23 I'm not a kernel developer :( May 28 15:47:25 undrwater: lumag is tosa kernel guru :) May 28 15:48:16 Jay7: okay...was curious how it was getting along...and if a donation of a tosa would help any May 28 15:48:38 lumag still awaiting for tosa :) May 28 15:48:48 * Jay7 scratches XorA May 28 15:48:53 awaiting? May 28 15:49:08 he would like one to do further devel? May 28 15:49:11 XorA should send tosa for lumag and collie for me May 28 15:49:25 Jay7: i can send you a collie May 28 15:49:53 and can send lumag a tosa ;) May 28 15:49:56 hm.. that might be good idea May 28 15:50:15 well, let's wait for any XorA's answer :) May 28 15:50:19 hrw, maybe some ssh-agent issues? May 28 15:50:19 k May 28 15:51:22 $ ssh git.openembedded.org May 28 15:51:23 ssh: Permission denied (publickey). May 28 15:51:25 hehe.. May 28 15:53:02 GNUtoo: will catch someone later May 28 15:53:33 ok May 28 15:53:35 for me it works May 28 15:53:37 SSH_AUTH_SOCK="" git pull May 28 15:53:51 5ff088e..4a63e4b org.openembedded.dev -> origin/org.openembedded.dev May 28 15:53:52 etc... May 28 15:54:05 seems git falling back to git May 28 15:54:12 .. proto May 28 15:54:16 ah? May 28 15:55:01 17:53 hrw@home:openembedded$ SSH_AUTH_SOCK="" git pull May 28 15:55:01 Permission denied (publickey). May 28 15:55:21 anyway need to go now May 28 15:55:32 I use SSH_AUTH_SOCK="" because of ssh-agent issues,not shure if you should use it May 28 15:55:33 ok May 28 15:56:12 have a nice weekend everybody May 28 15:56:18 ah, no.. it operating over ssh May 28 15:56:21 strange May 28 16:00:46 morning May 28 16:01:28 03Chase Maupin  07org.openembedded.dev * rc39f6ea2ee 10openembedded.git/docs/usermanual/chapters/getting_oe.xml: May 28 16:01:28 docs/getting_oe: Fix documentation errors May 28 16:01:28 * Fixed up typos and other errors in the documentation. May 28 16:01:28 Acked-by: Koen Kooi May 28 16:01:28 Signed-off-by: Chase Maupin May 28 16:01:29 Signed-off-by: Denys Dmytriyenko May 28 16:01:29 03Chase Maupin  07org.openembedded.dev * r2f165113a3 10openembedded.git/docs/usermanual/chapters/recipes.xml: May 28 16:01:30 docs/recipes: Fix documentation errors May 28 16:01:31 * Fixed up typos and other errors in the documentation. May 28 16:01:31 Acked-by: Koen Kooi May 28 16:01:32 Signed-off-by: Chase Maupin May 28 16:01:32 Signed-off-by: Denys Dmytriyenko May 28 16:01:33 03Chase Maupin  07org.openembedded.dev * rc3cd6731d7 10openembedded.git/docs/usermanual/chapters/common_use_cases.xml: May 28 16:01:33 docs/common_use_cases: Fix documentation errors May 28 16:01:34 * Fixed up typos and other errors in the documentation. May 28 16:01:34 Acked-by: Koen Kooi May 28 16:01:46 Signed-off-by: Chase Maupin May 28 16:01:46 Signed-off-by: Denys Dmytriyenko May 28 16:01:46 03Chase Maupin  07org.openembedded.dev * r42d3b44313 10openembedded.git/docs/usermanual/chapters/metadata.xml: May 28 16:01:46 docs/metadata: Fix documentation errors May 28 16:01:47 * Fixed up typos and other errors in the documentation. May 28 16:01:47 Acked-by: Koen Kooi May 28 16:01:48 Signed-off-by: Chase Maupin May 28 16:01:48 Signed-off-by: Denys Dmytriyenko May 28 16:02:40 * kergoth_ grumbles, forgot to update the docs per the patch.bbclass changes.. should do that today May 28 16:03:45 kergoth_: gm May 28 16:05:25 good morning May 28 16:05:44 so I'm giving another presentation on OE at the Open Source Bridge conference next week, and I'm updating my slides May 28 16:06:01 I'd like to list some hardware vendors that ship their boards with OE/Poky support May 28 16:06:18 so far I know gumstix, Bug Labs do this May 28 16:06:30 any other known vendors? May 28 16:07:41 dreambox May 28 16:07:47 palm pre May 28 16:07:54 webos is oe-based afaik May 28 16:08:41 do they distribute the webos buildsystem, or only the individual sources for minimal gpl compliance? May 28 16:08:43 zenlinuxPDX: did you mean official vendor support in OE/Poky? or just hardware support in OE? May 28 16:08:53 kergoth_: hrw should know more :) May 28 16:09:00 ah May 28 16:09:02 I know this from his blog :) May 28 16:09:05 hehe May 28 16:10:24 I'm looking for board vendors that entrepreneurs and businesses would use to build products with May 28 16:10:39 so BSPs based on OE/Poky, more or less May 28 16:11:20 and requiring vendor support of that, so you don't want something unofficially supported, rather than vendor shipped? May 28 16:12:09 right, the point of this particular slide is to highlight hardware companies that understand the advantages of OE and are not using their own custom Linux distro based on butchering Debian or RedHat or whatever May 28 16:12:15 Jay7: erk, sorry, life is chaos May 28 16:12:20 * kergoth_ nods May 28 16:12:22 hmm May 28 16:12:34 Does SheevaPlug ship with OE these days? May 28 16:12:37 XorA: np, just asking for current status :) May 28 16:12:47 zenlinuxPDX: although our website is not up to date, we begin to ship our system on modules with OE last year and I'm working at mainlining the things May 28 16:12:53 doesn't sheevaplug still ship with ubuntu? May 28 16:12:59 ericben: nice, who do you work for? May 28 16:13:10 http://www.eukrea.com/ May 28 16:13:37 ppl: copy this info to OE site please :) May 28 16:13:41 ericben, great! I will share the news May 28 16:13:42 about vendor support May 28 16:14:21 Here's the short abstract of my talk: http://opensourcebridge.org/sessions/390 May 28 16:14:33 It will just be a quick intro on how to get started with OE and why it kicks ass :) May 28 16:14:44 roadmap is to push our i.MX25/27/35/51 boards's support to OE by this summer May 28 16:17:15 zenlinuxPDX: This one ships with OE built stuff: http://www.avnet-memec.eu/products/newsletter/marvell/marvell-issue-43february-2010-launch-of-warpcomm-marvell-mpu-dev-kit.html May 28 16:18:59 florian, thanks May 28 16:20:17 zenlinuxPDX: some Karo modules are also mainlined, some Ronetix ones also by Kaeil May 28 16:20:45 Atmel is using OE for it's eval board (and also buildroot I think) May 28 16:21:14 ericben: Do you have a module comparable to the TX51? May 28 16:22:13 florian: it is work in progress and prototypes should be on my desk by the end of june (pin to pin compatible with i.MX25 & i.MX35 modules) May 28 16:22:25 florian: CAD was finished today for this one ! May 28 16:22:26 ericben: The Ka-Ro ones do not use OE officially even if we have support for some of them now. May 28 16:22:51 so as a follow-up question, if you were to name three boards that someone could buy and use OE on easily, which would you recommend? May 28 16:23:18 In this case I believe the SheevaPlug can run OE, and I might want to mention that one since I think it's a pretty affordable and is becoming a popular platform. May 28 16:24:24 ericben, hi I will read all your mail,I already tried one of your driver but it failed to compile because of undefined function, I fixed some but now I need to finish that fixing work May 28 16:24:41 undrwater: can you check shipping cost from US to Russia by USPS? May 28 16:24:45 zenlinuxPDX: aprat from gumstix I don't know the other ones ! May 28 16:25:28 also igep (omap3 based) is using OE officialy May 28 16:25:31 zenlinuxPDX, beagleboard like boards? May 28 16:25:42 oh right, the beagleboard - that's a good one May 28 16:25:44 s/igep/igepv2 ? May 28 16:25:59 GNUtoo: yes this one May 28 16:26:02 ok May 28 16:26:35 Hello! May 28 16:26:56 "Every hardware device vendor seems to have its own embedded Linux distribution and way of developing for it." - that's no longer true... May 28 16:26:59 about sheevaplugs, there are new ones like guru plug no? May 28 16:27:10 Is there any suggestion for an embedded linux device which processor to use if the memory usage efficiency and low power consumption are the most important viewpoints May 28 16:27:11 ? May 28 16:27:20 we use Blackfin DSP right now for image processing May 28 16:27:28 with our own operating system May 28 16:27:37 but we would like to change it for embedded linux May 28 16:27:44 znlinuxPDX: easy to use boards : 1st one would be beagle of course, 2nd Gumstix, 3rd maybe Bug Labs' products May 28 16:27:48 djszapi|work, blackfin is uClinux right? May 28 16:28:07 Blackfin is a DSP processor May 28 16:28:11 I know well buglabs board May 28 16:28:13 Digital Signal Processor May 28 16:28:21 from Analog Devices vendor May 28 16:28:21 ok May 28 16:28:29 and there is a uclinux support for that yeah May 28 16:28:35 so 3 OMAP3 based boards, TI made a good job with Beagle board :-D May 28 16:28:36 It's used in a low cost asterisk appliance May 28 16:28:37 but Blackfin processors aren't ok in the future for us May 28 16:28:41 to slow memory usage May 28 16:28:44 133 MHz May 28 16:28:48 ok May 28 16:28:56 Jay7: is usps the only shipper from us to russia? May 28 16:28:57 SDRAM May 28 16:29:02 denix, perhaps we disagree on this, but from my experience and what I've seen, the embedded Linux "marketplace" is incredibly fragmented. It's hard to develop products in a standardized way with all the idiosyncrasies you have to deal with every time you migrate to a new hw platform. May 28 16:29:04 and we would like DDR RAM May 28 16:29:05 djszapi|work, There is a plan to port "linux" on c64x May 28 16:29:13 and minimum 128K cache usage May 28 16:29:14 beagleboard paved the way, but now almost all TI's platforms use OE - DaVinci, OMAP3, OMAPL1, Sitara etc. May 28 16:29:39 undrwater: we have lower custom tariffs for USPS May 28 16:29:47 Jay7: ok May 28 16:29:56 DHL and FedEx have some limits May 28 16:29:56 zenlinuxPDX: what I meant is it used to be much worse! :) it's getting better these days... May 28 16:30:13 this processor uses 3 W power consumption May 28 16:30:16 denix: yes on chip makers's point of view TI & Atmel seems to be good supporters of OE May 28 16:30:28 so it would be fine to find a better one in this regard May 28 16:30:31 zenlinuxPDX, do you need infos on the buglabs's board? May 28 16:30:38 and so that is the point I want to make, and show that OE offers one possible way of getting some standardization May 28 16:31:03 GNUtoo, I think the web site is pretty good. Do you work for Bug Labs? May 28 16:31:14 zenlinuxPDX, no but I've 2 bugbase May 28 16:31:15 denix, ah. :) May 28 16:31:22 one without wifi and one with wifi May 28 16:31:23 GNUtoo, I will ping you if I have questions about it May 28 16:31:26 ok May 28 16:31:29 nice May 28 16:31:30 thanks May 28 16:31:33 GNUtoo: are you from England ? May 28 16:31:36 djszapi|work: look at beagleboard - uses 4W, has powerful ARM + DSP May 28 16:31:47 beagleboard, okay. May 28 16:31:47 djszapi|work, no I'm french and live in italy May 28 16:31:51 let me check it, thanks May 28 16:32:27 djszapi|work, be carefull with beagleboard,they included no accessories last time I checked,not even the power supply May 28 16:32:36 denix: I look for just a processor May 28 16:32:38 not a complete board May 28 16:32:43 we have hardware designers. May 28 16:32:51 they will design the board according to our requirements. May 28 16:32:56 some other beagle-like board are more expensive but include some accesories like power supply May 28 16:32:58 ok May 28 16:33:02 djszapi|work: beagleboard uses OMAP3 (3530 or 3730) May 28 16:33:28 does it have floating point emulation ? May 28 16:33:32 djszapi|work, so you're interested in FLOSS hardware? May 28 16:33:38 djszapi|work: there is also hawkboard, which uses OMAPL138 - weaker ARM, but stronger DSP (fp) May 28 16:33:42 TI is a rather expensive company :-) May 28 16:33:58 hawkboard has also a sata connector no? May 28 16:34:06 GNUtoo: yep May 28 16:34:10 nice May 28 16:34:27 GNUtoo: what does it mean floss ? May 28 16:34:34 djszapi|work: beagle - ~$150 per board, hawk - <$100 per board May 28 16:34:36 djszapi|work, free libre and open source May 28 16:34:46 yes. May 28 16:34:52 like blackfin &&uclinux May 28 16:35:03 http://blackfin.uclinux.org May 28 16:35:08 I need for a free solution. May 28 16:35:09 djszapi|work, basically if I say free it's misleading because free means free as in cost and free as in freedom too May 28 16:35:10 djszapi|work: TI makes great products but there lead times are actually incredilby long May 28 16:35:22 denix: I would not like a full board, really. May 28 16:35:23 Jay7: looks like $14 US...but that's a small box...maybe power cable won't fit May 28 16:35:28 I am not kidding :-) May 28 16:35:45 djszapi|work: I'm not pushing you boards, I'm giving you examples May 28 16:35:56 ericben: what do you mean under lead times ? May 28 16:35:59 djszapi|work, what about gumstix? but only expansions board have schematics etc... May 28 16:36:21 schematics doesn't matter May 28 16:36:24 the time between your order and when you get the chips, sometimes over 26 weeks actually ! May 28 16:36:38 our hardware designers are very good and famous in their profession. May 28 16:36:38 Jay7: $44 US for the medium flat rate box :P May 28 16:36:50 undrwater: hehe.. May 28 16:36:50 gumstix, let me check it May 28 16:36:51 djszapi|work: gee, go talk to sales, if you need qutes on volume... :) May 28 16:37:00 ericben: 26 weeks May 28 16:37:01 lawl May 28 16:37:10 ericben: probably :) May 28 16:38:17 ericben: unless you are as big as nokia or moto... :) May 28 16:38:19 undrwater: well, we will try again with XorA :) May 28 16:39:13 1000 euro / processor is rather expensive May 28 16:39:23 intel atom motherboard costs 150-200 euro May 28 16:39:30 and it's a full atom, processor May 28 16:39:37 with desktop PC capability May 28 16:39:40 512K cache May 28 16:39:50 1,1 GHZ May 28 16:39:56 and 1,6GHZ May 28 16:40:07 is that the price for the blackfin? :) May 28 16:40:39 djszapi|work : for example, for OMAP3530 it's 14 weeks on the website, but for tps65023 (power chip we are using with omap3530) it's 20 weeks on the website and much more in real life ;-) May 28 16:40:40 20 euro / blackfin May 28 16:40:55 btw, I'm looking for some chinese board developers May 28 16:40:56 this doesn't count ericben May 28 16:41:06 It's a long distance development May 28 16:41:11 at least it will be. May 28 16:41:17 we have a lot do as usual :-) May 28 16:41:24 I can do other thing that time. May 28 16:41:59 ericben: yeah, power/peripheral chips like TPS are in huge demand for some reason... :) May 28 16:42:07 but anyway other manufacturer actually also have long leadtimes but TI's one beat them May 28 16:42:14 hm, omap processors are used for mobile phones May 28 16:42:23 .not sure it's the best for image, video signal processing May 28 16:42:33 like cameras, fingerprint scanners, so forth May 28 16:42:34 DaVinci May 28 16:42:57 DJWillis: are you here? May 28 16:43:02 that's a specialized family for video/image May 28 16:43:05 denix: yes, TPS products are great, but having to wait over 20weeks for them is not great :-) May 28 16:43:11 starts from $5/chip May 28 16:43:14 Jay7: does lumag make it here? May 28 16:43:14 we have 9-10 MPixel sensors May 28 16:43:24 we must send them to the PC. May 28 16:43:34 so we need for quick memory usage processors. May 28 16:44:32 do you know of a good opensource license for hardware (including schematics & all CAD files) ? May 28 16:44:44 eagle May 28 16:44:51 for schematic &board designing May 28 16:45:09 not a software, a license for the files that make the design May 28 16:45:20 good question May 28 16:45:24 undrwater: can't say now.. May 28 16:45:33 I don't know :) May 28 16:45:41 ericban: I don't understand the question May 28 16:45:59 as far as I know in electronics patents are used,not copyright May 28 16:46:36 djszapi|work: if I have a board and want to give the sources of this board : which license is the best for these sources May 28 16:46:45 you can use creative commons for the design... May 28 16:46:53 ericben: some sort of creatvie commons? May 28 16:46:55 ericben: ah okay. May 28 16:46:56 dunno sorry. May 28 16:47:03 djszapi|work, basically for software there are licenses like the GPL,he wonder how to make the design of the hardware open source/free software May 28 16:47:20 why is eagle not okay ? May 28 16:47:23 I don't understand it May 28 16:47:26 in fact I have a design we cancelled and I would like to publish it so that the work we did on it is not totally list May 28 16:47:44 djszapi|work: do you know what license is? May 28 16:47:50 http://docs.blackfin.uclinux.org/doku.php?id=hw:jtag:gnice May 28 16:47:58 this is an open source ICE for instance. May 28 16:48:28 I know what's a license May 28 16:48:34 but I don't understand the question May 28 16:48:40 why does it cause any problem ? May 28 16:48:41 djszapi|work: I though you were talking of the software eagle ;-) May 28 16:48:59 which is a CAO soft ! May 28 16:49:10 sorry for the confusion May 28 16:49:14 :-) May 28 16:49:16 sorry, you lost me here. May 28 16:49:40 I'll try to explain May 28 16:49:43 where do you find this eagle license ? May 28 16:49:58 ahhh May 28 16:50:01 eagle license May 28 16:50:04 me too May 28 16:50:09 I was confused May 28 16:50:39 lincense doesn't depend on any software May 28 16:50:47 you can do it open source licensed whenever you want May 28 16:50:51 hey. i have the following in a recipe: http://pastebin.com/VxB6pXBg May 28 16:50:55 I don't understand the question, really. May 28 16:51:02 what causes the issue in it ? May 28 16:51:03 and i get this: NOTE: Not creating empty archive for tesseract-ocr-2.04-r0.5 May 28 16:51:05 ok I'll try to explain better May 28 16:51:45 in copyright if you don't publish your work under a license which allows things like modification,redistribution etc....it's forbidden May 28 16:51:50 is there anything i've missed so that the tesseract-ocr package remains empty? May 28 16:51:52 phone ringing May 28 16:51:54 bbs May 28 16:52:27 GNUtoo: and ? May 28 16:52:57 technical check: anybody seeing what i'm writing? May 28 16:53:07 I've to go bbs May 28 16:53:11 sorry May 28 16:53:16 no problem May 28 16:53:19 have a nice day! May 28 16:53:27 bu GNUtoo May 28 16:53:31 CMoH: yes May 28 16:53:43 ty djszapi|work, i was ready to reconnect or such May 28 16:54:00 TI davinci processors are rather expensive May 28 16:54:31 djszapi|work: in fact I'm looking for a licence which would make the sources of the board design opensource May 28 16:54:55 what about my link ? May 28 16:55:00 It's an opensource hardware May 28 16:55:07 with free schematic + board design May 28 16:55:11 it can be a good example for you. May 28 16:55:22 denix: yes I was thinking to Creative also May 28 16:55:50 djszapi|work: ok what is the license of the files ? May 28 16:55:58 check it May 28 16:56:18 djszapi|work: yes but I don't find the sources of the CAD files May 28 16:56:22 * Licensed under the GPL-2 or later May 28 16:56:43 sorry May 28 16:56:49 no GPL is the source of the soft May 28 16:56:49 it's the license of the simple appliaction May 28 16:56:53 let me check it May 28 16:56:54 you have the schematics May 28 16:57:03 but not the sources to make evolutions on the design May 28 16:57:05 no I don't have May 28 16:57:09 but if you know better ;) May 28 16:57:46 http://blackfin.uclinux.org/gf/project/stamp/frs/?action=FrsReleaseBrowse&frs_package_id=146 May 28 16:58:00 sorry, I mean you have access to the schematics but not to their sources nor those of the files of the board May 28 16:58:11 yes I've also been there May 28 16:58:16 -- Systems Engineering -- Open Source Hard and Software Projects May 28 16:58:18 Open Source Hard. May 28 16:58:28 nope May 28 16:58:33 I have the sources ... May 28 16:59:02 you can download it freely. May 28 16:59:18 ahh just screenshots May 28 16:59:20 ok maybe, I didn't found where May 28 16:59:20 strange May 28 16:59:31 I just foudn gerbers, pdf & jpg May 28 16:59:58 no problem, that's the difference between opensource design and design where one provides schematics in pdf ;-) May 28 17:00:39 :D:D May 28 17:00:52 what's the problem with an own prop. licence ? May 28 17:00:57 and is gpl for softwares only ? May 28 17:01:15 yes GPL seems not to be OK for hardware May 28 17:01:19 from what I read May 28 17:01:50 what's the problem with own license ? May 28 17:01:57 do you look for a common one instead ? May 28 17:02:00 a known one ? May 28 17:02:34 the problem is that I'm not a lawyer and don't want to pay one for that ;-) May 28 17:02:41 :D:D May 28 17:02:50 are you a hardware designer, anyway ? May 28 17:03:20 yes May 28 17:03:37 can't you ask it at the company ? May 28 17:03:43 ericben, ? May 28 17:03:45 that's my company :-D May 28 17:03:52 yes. May 28 17:03:55 nobody ? May 28 17:04:53 in fact trying to make a opensource hardware is much more complicated than makeing a proprietary one on license poin of view :-D May 28 17:05:10 :P May 28 17:06:22 so basically I think the goal of the free and open source hardware license ericben is talking about is to be copyleft, a bit like the GPL,that is to say if someone makes a modification and redistribute it he may have to license his modification under the same free and open source license May 28 17:07:58 GNUtoo: I've realised it May 28 17:08:01 GNUtoo: exactly and if he want to make a closed product, he purchased us a license for that May 28 17:08:04 while you were away :-) May 28 17:09:08 ah ok May 28 17:09:15 but thanks anyway. May 28 17:09:21 ok May 28 17:09:47 ericben: i'd ask the arduino people May 28 17:09:58 there are also QI hardware people May 28 17:10:03 it's copyleft too May 28 17:10:05 ericben: team@arduino.cc May 28 17:10:19 QI hardware makes the nanonote May 28 17:13:25 they use creative common for the schematics (nanonote) May 28 17:14:21 GNUtoo: thanks for the info. Yes creative seems interesting as it can prevent commercial usage if we want (prevent is a big world here ;-) May 28 17:15:09 ok May 28 17:15:33 also look at opencores licensing May 28 17:16:53 http://en.wikipedia.org/wiki/Open_source_hardware#Licenses could also be interesting May 28 17:18:23 someone jog my memory please. I've added a package on my build machine. I want to install it on my target. I've got /etc/opkg/local-feed.conf set up properly on my target. I do an "opkg update" and "opkg list" on my target. My package isn't listed. I know I forgot a step on my build machine. something about rebuilding package indexes or something. What is that step? May 28 17:19:25 bitbake package-index? May 28 17:19:31 thank you! May 28 17:21:06 http://focus.ti.com/docs/prod/folders/print/tms320dm6446.html#pricingpackaging May 28 17:21:10 hm, Davinci seems fine May 28 17:21:17 but I can't see its power consumption. May 28 17:23:47 can you see it ? May 28 17:30:54 davinci processors are for video signal processing May 28 17:31:01 I need for them because of image processing right now May 28 17:31:05 not video signal processing May 28 17:31:18 so smaller FPS number, but much more bigger images. May 28 17:31:36 DaVinci family has DSP-only, ARM-only and ARM+DSP SoCs May 28 17:32:42 usually, ARM-only DaVincis are cheap, low power image processors, because of ARM9 core May 28 17:33:01 really ? May 28 17:33:24 you said you need 10MP image processing and you need DSP, so you may need something more powerful May 28 17:33:56 DM 355 and 365 yeah May 28 17:33:57 as an option - DSP-only DaVincis are quite powerful, but those don't run Linux May 28 17:34:04 I don't need DSP May 28 17:34:11 if ARM can be ok for image processing May 28 17:34:15 for 10M images May 28 17:34:23 and if it can be low-power meanwhile May 28 17:34:37 and it has quick memory usage and minimum 128K cache May 28 17:34:59 the task is linux May 28 17:35:04 so I must choose such a processor May 28 17:35:35 not sure about the power consumtion on DaVincis - they are not meant as mobile chips, so PM is not as sophisticated as on OMAPs May 28 17:36:04 PM ? May 28 17:36:08 Power Management ? May 28 17:36:11 yes May 28 17:36:26 I see May 28 17:36:41 let me check ARM only ones. May 28 17:36:59 meaning - no deep sleep, not many PM domains etc. May 28 17:37:41 no deep sleep can be problem May 28 17:38:14 btw, you can roughly compare video resolution/fps to image processing... :) May 28 17:38:50 ? May 28 17:39:03 i.e. higher the resolution, lower the fps on the same processor... and when fps is <10 then it's no longer video :) :) May 28 17:39:18 yes May 28 17:39:21 it's no longer video May 28 17:39:28 1 images / 1-2 sec May 28 17:40:04 not a moving image capturing May 28 17:40:09 e..g passport reader devices May 28 17:40:15 and fingerprint scanners what we do May 28 17:40:20 and face ident devices May 28 17:40:28 we have big images. May 28 17:40:31 that's the main point May 28 17:40:34 basically, 1080p/30 is 2MP with 30fps. you need 10MP, then fps will be roughly 6 frames/second May 28 17:40:51 ^^^^ May 28 17:41:34 quick memory usage + 128K cache min + low power consumption + embedded linux May 28 17:41:37 10MP in 1-2 sec is not very intensive at all May 28 17:41:41 those are the conditions just about :-) May 28 17:42:11 (at first glance) May 28 17:43:16 that's capturing+encoding... if you then need to process the image and run some hard-core algorithms, then you may need something more powerful... May 28 17:43:26 and FPE would be a bonus May 28 17:43:45 yes, e.g. compression before sending to the PC. May 28 17:43:51 not jpeg, but minimum huffman May 28 17:44:24 ah, if you are not trying to process them on the board, but rather sending them to PC for later processing, then it seems no brainer :) May 28 17:44:33 http://focus.ti.com/docs/prod/folders/print/tms320dm6446.html May 28 17:44:35 it seems fine May 28 17:44:41 but I don't see its power consumption May 28 17:44:46 6446 is ARM+DSP May 28 17:44:50 1-2 W is the expected by us :-) May 28 17:44:51 yes. May 28 17:45:41 I'd recommend calling sales at this point - they maybe able to get those numbers for you May 28 17:45:59 okay. May 28 17:46:29 64xx family seems fine for me, anyway. May 28 17:46:34 it's quick processor relatively. May 28 17:46:39 good cache size May 28 17:46:42 not so expensive, etc. May 28 17:46:45 embedded inux May 28 17:46:47 * linux May 28 17:47:55 but I don't need for so much FPS at all :-) May 28 17:48:46 is there any other potential solution apart from Texas Instrument Davinci cores ? May 28 17:54:56 good morning May 28 17:55:06 khem: u here? May 28 17:55:50 hi May 28 17:57:06 doenix: do these processors have FPE ? May 28 17:58:39 IgorK: whats up May 28 18:01:07 TMS320DM642 Video/Imaging Fixed-Point Digital Signal Processor May 28 18:01:43 I rebuilt the image May 28 18:02:16 nm was put to "/usr/arm-angstrom-linux-gnueabi/bin/nm" May 28 18:02:50 hmmm I am in meeting May 28 18:03:02 IgorK: I will ask more later May 28 18:03:13 ok, I'll wait 4 u here... ;-) May 28 18:20:14 re May 28 18:20:22 florian: hey :) May 28 18:20:41 wiki seems gone from ltg again :) May 28 18:24:38 Jay7x: impossible ;) May 28 18:32:02 hehe.. May 28 18:32:22 now ltg gone too :) May 28 18:37:13 hmm May 28 19:32:55 hi kergoth May 28 19:33:01 hey May 28 19:53:20 having a weird issue with fetching gtk-webcore: NOTE: Task failed: Unknown fetch Error: [Errno 2] No such file or directory: /path/to/webcore_tarball.tar.gz May 28 19:55:04 i dont even see the generated run.do_fetch.XXXX script May 28 21:16:08 hi pb_ May 28 21:23:34 doh, stupid wifi stops working every time we use the microwave May 28 21:23:39 * pb_ stabs his computer May 28 21:23:54 pb_: try to switch channel :) May 28 21:24:04 pb_: we live in a waves world... May 28 21:24:33 Jay7: not my AP, sadly May 28 21:24:40 ah.. May 28 21:24:45 heh May 28 21:25:09 pb_: wwe were talking about the -static packages and package split May 28 21:25:32 it happened that misnamed recipes like udev-static_124.bb could not correctly build May 28 21:25:58 some, just some, tasks where overwritten by main package (udev in this case) May 28 21:26:08 *were ..oops May 28 21:26:34 ant__: I will post a patch to rename -static packages to -dev-static May 28 21:27:04 khem: but how should be called e.g. kexec-tools built with klibc, statically May 28 21:27:32 to distinguish btw *libc May 28 21:28:03 ant__: oh, right, yeah. May 28 21:28:15 just create kexec-tools-klibc and only package the -static? May 28 21:28:31 I guess that's just a thing: udev-static is (in that sense) no different really to udev-doc and suchlike, and oe is already a minefield of potential naming conflicts. May 28 21:28:55 renaming the -static pattern to -dev-static would be one option, but I would be tempted to rename udev-static instead since I suspect that will have a smaller impact. May 28 21:29:42 pb_: could be but I think -dev-static is more revealing May 28 21:30:31 that's probably true, but I think the time to say that was probably in September when the -static change was originally proposed. I'm not sure that renaming all those packages after nearly nine months is a good plan. May 28 21:31:05 if you do a global rename of -static to -dev-static, anybody using those packages will be hosed. May 28 21:31:35 if you rename udev-static and suchlike, that isn't such a concern since the users of those packages are already hosed at the moment. May 28 21:32:21 what' worst, the package was not built May 28 21:32:30 the do_package task was stolen May 28 21:32:37 by udev_151 May 28 21:32:48 while building udev-static_124 May 28 21:33:02 result: empty package skipped... May 28 21:33:54 pb_: there is a second issue: udev is machine specific because of some overides. udev-static is arch specific...opinions? May 28 21:34:49 ant__: I suspect most of those overrides are bogus May 28 21:35:09 they should, probably, be DISTRO overrides not MACHINE overrides May 28 21:35:23 iirc it was for the static dev tarball creation in case of RO root at boot May 28 21:35:31 as in Zaurus May 28 21:36:50 I don't really know all that much about udev though, and nor do I have much desire to know more about it :-} May 28 21:37:26 anyway, even after renaming it I don't see the two -124-static files in /sbin of console-image May 28 21:37:37 so there is some packaging issue (or my bad...) May 28 21:38:02 the .ipk seems correct, though May 28 22:16:49 khem: I'm preparing a patch to submit to you May 28 22:29:31 khem: u back from meeting? May 28 22:46:47 how to apply a patch in a recipe depending on a local.conf setting? e.g. build u-boot with or without NFS bootargs. May 28 22:50:29 i'm thinking about using 2 recipes which just differ in applying the patch or not and in the local.conf i would choose the right one with PREFERRED_PROVIDER, but is it the right way? May 28 23:13:58 khem, have some new pointers it is related to kernel May 28 23:21:31 IgorK: yes now I am May 28 23:24:39 khem: the program was installed in "/usr/arm-angstrom-linux-gnueabi/bin" May 28 23:25:00 so, when typing "nm" I got "command not found" May 28 23:32:50 IgorK: hmmm thats strange May 28 23:34:02 matgnt: You can add this to DISTRO_FEATURES May 28 23:34:14 and use that in recipe May 28 23:34:47 another option is that you could use your machine override May 28 23:38:32 IgorK: Can you check the binutils build tree May 28 23:40:19 03Andrea Adami  07org.openembedded.dev * rc027cfceec 10openembedded.git/recipes/udev/ (7 files in 2 dirs): May 28 23:40:19 udev-static: rename recipe to udev-compat to avoid packaging clash May 28 23:40:19 Signed-off-by: Andrea Adami May 28 23:40:20 khem: now feel free to rename May 28 23:40:30 03Andrea Adami  07org.openembedded.dev * r28bc48904e 10openembedded.git/recipes/udev/ (udev-151/udev-compat-static.patch udev_151.bb): May 28 23:40:30 udev: patch init of version 151 and support older kernels May 28 23:40:30 * Some devices don't function properly when using a recent udev on an old May 28 23:40:30 * kernel, so udev will now try to fall back to an older version is present May 28 23:40:31 Signed-off-by: Koen Kooi May 28 23:45:17 ant__: now we have to rename May 28 23:45:25 other recipes too May 28 23:45:44 I left the 'static' suffix to simplify your grep :) May 28 23:45:52 btw I'm checking klibc May 28 23:45:56 why did you call it compat May 28 23:45:57 we can remove some May 28 23:46:06 it was not needed May 28 23:46:12 if we renamed the auto packages May 28 23:46:39 now it builds at least May 28 23:47:04 ipk is sane May 28 23:47:30 ...still have to check runtime personally... May 28 23:47:37 doing this now May 28 23:48:59 ant__: thats ok May 28 23:49:29 may be we should just rename few recipes as pb suggested May 28 23:50:16 I'm testing klibc >1.5.15 now that the make_headers_install thing is implemented May 28 23:50:38 ok May 28 23:50:43 it failed with kernel 2.6.26 and headers 2.6.31 May 28 23:52:01 I'm sometimes under the impression the only users of klibc are we of kexecboot project May 28 23:52:02 * khem has to pickup kids May 28 23:52:15 we )2-3) May 28 23:52:21 could be May 28 23:52:35 1.5.18 is tested with 2.6.34 May 28 23:52:38 fyi May 28 23:55:50 hm.. a known issue...asm-generic/signal-defs.h: No such file or directory May 28 23:58:00 missing patch it sems May 29 00:03:03 khem: thanks for your hint. with "machine override" you mean MACHINE_FEATURE? or the "_machine" to add machine specific parts? May 29 00:04:15 _machine May 29 00:07:55 so i would have to create two machine configs to use this option, right? May 29 00:09:02 khem: I could solve the >=2.6.31 #include vs. #include May 29 00:09:05 but May 29 00:09:27 now /include/sys/socket.h:146: error: redefinition of typedef 'sa_family_t' May 29 00:09:40 hm..I've already seen this too ... May 29 00:11:27 khem: where do u want me to look: http://pastebin.com/AhQsw4rc May 29 00:15:38 good night May 29 00:47:17 anybody knows what I need to fix here: http://pastebin.com/nXTv0GGq, lines 17, 18? **** ENDING LOGGING AT Sat May 29 02:59:57 2010