**** BEGIN LOGGING AT Wed May 23 03:00:00 2012 May 23 03:15:35 Sorry for the repost - i hoping maybe someone hopped on who can help. I am trying to mount a USB thumb drive on my beaglebone. I am using Kernel Rev 3.2.16. If I type in dmesg | tail, it recognizes that a USB drive has been inserted, but it doesn't 'assign' it to anything (sdX). It worked with 3.2.5+, but not 3.2.14 or 3.2.16. Any advice? **** ENDING LOGGING AT Wed May 23 04:35:22 2012 **** BEGIN LOGGING AT Wed May 23 04:36:15 2012 May 23 05:15:06 jsabeaudry_: Tried #angstrom? May 23 05:53:56 <_sundar_> is it possible to use LDREX and friends from userspace? May 23 07:02:14 \o/ http://www.silabs.com/products/interface/spitoi2c/Pages/default.aspx May 23 07:18:41 test May 23 07:20:37 failed :-þ May 23 07:21:49 dm8tbr: found a new favority char? May 23 07:21:55 favorite* May 23 07:37:27 av500: are you creating creating usb<->spi/i2c dongle? May 23 07:39:58 ynezz: well, there is both direction, i2c to SPI as well May 23 07:40:10 so, I'm thinking about a long chain :) May 23 07:40:34 SPI2Ception May 23 07:41:33 av500: yu so b0r3d? May 23 07:41:58 hm, they've CP2112 HID<->i2c already May 23 07:42:20 av500: been using that for quite a while when on linux. [altgr][p] May 23 07:42:20 LetoThe2nd: of course I wont actually do it May 23 07:42:44 * LetoThe2nd sighs in relief May 23 07:42:46 LetoThe2nd: file it under "man müßte mal..." May 23 07:44:49 XorA: stop playing with that google doodle! May 23 07:53:09 av500: I'm looking at a cape with TDA19989 May 23 07:53:48 av500: better don't file it at all. May 23 08:00:35 koen: you signed an NDA? :) May 23 08:10:17 av500: no, nxp dropped the nda singing AIUI May 23 08:11:29 koen: its probably not needed since all the hdcp is baked in May 23 08:13:46 use a can opener May 23 08:14:29 av500: is gen8-gpl the latest public code for it from Rona? May 23 08:16:03 it looks it it's "one of those" chips May 23 08:16:18 lots of different drivers, none upstream May 23 08:17:59 koen: we asked nxp permission the release the code May 23 08:18:08 coz it came all with scare headers May 23 08:18:10 scary May 23 08:35:00 just read the r-pi review at the register, it seems the commonwealth really was in need for a spectrum successor ;) May 23 08:36:52 yes May 23 08:40:44 always talking about $25 but then showing web-browsers which are useless on a $25-r-pi May 23 08:43:13 koen: wrt gen8, what is the question? May 23 08:43:37 koen: bah, you cant get to the pitchbender at same time as playing the notes :-( May 23 08:43:58 koen: I shall have to setup my bassstation in moog mode May 23 08:44:20 av500: if that is the latest tda code you guys have publically available May 23 08:44:48 XorA: get a second mouse. May 23 08:45:03 koen: yes May 23 08:45:27 its gen8 and I pushed it late, so nothing done on it afterwards May 23 08:46:03 LetoThe2nd: got hardware, easier than 2 mouse support in linux :-D May 23 08:46:47 XorA: hehe, but i doubt it can beat the nerd factor of playing a google doodle moog with two mouses. May 23 08:47:49 av500: thanks, I'll compare it against the other codez on the web May 23 08:48:04 e.g. cubox May 23 08:48:33 personally, I'd work first on CEC, but I guess gerald wants to have a picture on his screen first May 23 08:48:56 koen: we have CEC now in gen9 May 23 08:49:11 even TI has something for omap4 May 23 08:49:32 surprisingly it almost works with many TVs May 23 08:55:36 av500: gen9? May 23 08:56:15 jacekowski: yes? May 23 08:56:20 what's gen9? May 23 08:56:24 gen9 bb? May 23 08:56:24 ah May 23 08:56:27 our products May 23 08:56:29 link May 23 08:56:44 http://www.archos.com/products/gen9/index.html?country=us&lang=en May 23 08:56:56 ahm, tablet May 23 08:57:07 (sorry for the advertisement) May 23 08:57:16 i thought it's something more like htpc kinda direction May 23 08:59:14 that wondermedia arm 11 from via, is that an armv6 too? May 23 08:59:22 yes May 23 08:59:30 arm11 is armv6 May 23 09:00:04 ah, thanks May 23 09:00:39 av500: omap4 has cec builtin, but this is for the bone May 23 09:01:12 koen: iknow May 23 09:01:38 how does r-pi look now May 23 09:01:47 can you get it or they still have supply problems May 23 09:01:57 they have May 23 09:02:21 I ordered a few days ago, after I registered in the queue when they started "selling" May 23 09:03:03 scheduled for christmas? May 23 09:04:00 no idea May 23 09:04:04 its not like I need it May 23 09:04:29 has anybody seen one? May 23 09:04:35 ebay has May 23 09:04:41 or is it all pictures on internet May 23 09:05:06 jacekowski: no, some people have it May 23 09:25:02 fb0->overlay0, fb1->overlay1: fb0 is on top. use sysfs to get fb0->overlay1, fb1->overlay0: fb0 is still on top. that be odd, no? May 23 09:40:22 * abogani waves all May 23 09:40:23 Could anyone pinpoint me on RTFM/Wiki/README to install Ubuntu on an BeagleBone, please? May 23 09:40:32 Thanks in advance! May 23 09:49:42 let us google that for you May 23 09:50:28 http://www.youtube.com/watch?v=T29BLCh1I54 May 23 09:50:31 there's a nice video May 23 09:55:24 niro, Thanks! May 23 09:58:39 np May 23 09:58:47 but do try and use google first. May 23 10:12:52 <_troll_> /j #pandaboard May 23 10:12:58 <_troll_> grr May 23 10:13:11 noob May 23 10:13:58 <_troll_> android + shitty net connection does that May 23 10:14:14 is there a minimal kernel image I can flash to test my recently acquired beaglebone May 23 10:14:40 _troll_, do you not like android? May 23 10:15:57 <_troll_> I didn't say that May 23 10:17:32 Sorry, I'm only asking because I'm thinking of moving from iphone to galaxy s 3. May 23 10:17:41 Just looking for an opinion really. May 23 10:17:48 niro, he doesnt like net connections ;) May 23 10:17:57 he didnt say android is shitty :) May 23 10:18:08 lol. May 23 10:18:17 "shitty net connection" is an app? May 23 10:18:52 <_troll_> no, it's a conference feature May 23 10:19:49 phh: http://kotelett.no/ac100/phh/Android2.2/libflashplayer.so May 23 10:19:55 av500: ? May 23 10:20:16 just scrolled by here... May 23 10:22:35 k4rtik, Define flash? Do you mean dump on to the SD card? May 23 10:28:34 exosyst: right May 23 10:28:58 I have a bare sd, can't download a big image using my net connection right now May 23 10:33:04 k4rtik, do you have a rootfs? May 23 10:38:44 niro: nope, I currently have starterware - a TI non-OS suite, and with it a demo app, which I can't get to work May 23 10:39:26 niro: plus I have sources of rowboat-android project May 23 10:42:29 What's your connection speed? May 23 10:43:01 connection speed is good - 2Mbps but my data plan is about to expire May 23 10:43:15 as in reach the limit May 23 10:44:51 I would try going to narcissus, I just created a barebones rootfs, that was around 3mb. May 23 10:45:04 Then you should be able to get a precompiled kernel somewhere I guess. May 23 10:46:14 A uImage is about 2MB. So you should be able to get it running in around 5 mb. May 23 10:46:43 niro: great, any links where to get these? May 23 10:47:35 http://www.angstrom-distribution.org/builder/ May 23 10:47:38 That's for the rootfs May 23 10:51:17 niro: bone not there in device list, can i go for beagleboard for rootfs? May 23 10:59:56 http://downloads.angstrom-distribution.org/demo/beaglebone/ May 23 11:00:08 This is going to be easier. May 23 11:00:20 The rootfs is 12mb. May 23 11:00:57 I have to nip out. May 23 11:01:16 niro: ok thanks for your help :) May 23 11:01:30 np. hope it works out. May 23 11:19:53 hi May 23 11:20:38 where can i find the sources to the 3.2.x kernel that runs on the beaglebone with angstrom? May 23 11:21:38 niro: thanks a lot man, I got a basic shell running, hurrah! May 23 11:21:45 gotta go now May 23 11:25:41 sneak: http://www.angstrom-distribution.org/building-angstrom May 23 11:31:42 i don't have a linux box, i just want the kernel sources May 23 11:32:35 koen: i'd like to build just the kernel on the beaglebone, i don't want all of angstrom. is there a place i can download the sources from? May 23 11:57:52 sneak: http://www.angstrom-distribution.org/building-angstrom May 23 11:58:04 sneak: so tell bitbake to only build the kernel May 23 12:04:36 why can't i just download the kernel sources somewhere? May 23 12:04:37 why do i have to jump through all of these bitbake hoops? May 23 12:04:38 this seems like intentional obfuscation May 23 12:05:43 sneak: see it as an initiation rite... May 23 12:06:34 "why can't i just download" is the question of those not understanding the need for a fine mixture of patchsets. May 23 12:07:08 it also doesn't work. May 23 12:07:14 http://pastebin.com/pHzNFT4d May 23 12:08:12 so update your git May 23 12:08:16 get a newer git May 23 12:08:26 did you miss the last line May 23 12:08:27 get a nevergit? May 23 12:08:47 it says 1.7.6 or newer. i have 1.7.7. May 23 12:08:55 * LetoThe2nd would suggest to compile on the dev box anyways, be it only for time reasons. May 23 12:09:07 i don't have any linux dev boxes, i can't stand the OS May 23 12:09:09 it says it needs --no-abrev May 23 12:09:34 it also says "please install 1.7.6 or newer" May 23 12:09:44 if you can't stand linux, what's the use for the kernel source? May 23 12:09:58 i intend to use the linux kernel on the beaglebone May 23 12:10:33 it doesn't seem that there's any other OSes with good driver support for the am3359 May 23 12:10:58 why oh why do i get this "its ugly, i cant stand it, it doesnt do what i want it, i don't want to learn it, so §$%§$& why don't you just fix my problem"-feeling again. May 23 12:11:22 sneak: just don't use linux, stick to u-boot only. May 23 12:11:29 ++ May 23 12:11:34 or even better, starterware. May 23 12:11:40 hush shell is really powerful :) May 23 12:12:04 and saves you from all these kernel bugs May 23 12:12:24 it can boot, it can fire up your super cool non-os applications... it must be THE NEXT BEST THING!(TM) May 23 12:12:47 you could even link your app directly into u-boot. May 23 12:12:50 LetoThe2nd: i don't want anyone to fix my problems. i'd just like the source code to the software that shipped on this device i just bought. May 23 12:13:18 i'd rather not have to execute a bunch of convoluted scripts to get that source code. May 23 12:13:18 repos that fetch other repos and patch this and that May 23 12:13:24 sneak: and you have been told how to get it. that you lack the proper tools for the job is your problem, not ours. May 23 12:13:28 whatever happened to "here's a device, here's a link to a tarball." May 23 12:13:39 thats last century May 23 12:13:42 the proper tools in this case should be restricted to "wget" May 23 12:13:42 +1 May 23 12:14:01 sneak: you're free to change it. May 23 12:14:02 ogra_: apparently May 23 12:14:12 LetoThe2nd: no, i'm not. May 23 12:14:12 the new way is to use sourced May 23 12:14:21 sneak: make up the tarball, give it to the world. it open source. May 23 12:14:31 sneak: there is an easy way to get the source May 23 12:14:42 av500: i'm waiting May 23 12:14:49 sneak: write an email to the beaglebone people requesting the GPL code May 23 12:14:56 av500: shh, don't give away "zcat /proc/kernel/source" May 23 12:15:13 sneak: per GPL they are obliged to comply May 23 12:15:19 i've tried it on both angstrom running on the beaglebone as well as the latest osx (the most common development environment on earth) with all the gnu garbage installed (autotools, gawk, etc etc) May 23 12:15:43 I am serious May 23 12:15:45 av500: technically there is nothign in the gpl that specifies email May 23 12:15:49 and technically i am in #beagle May 23 12:15:54 can anybody please hand me a kleenex, i burst in tears of laughing. May 23 12:15:57 funny, i work since 20 years with computers and never worked on the most common dev env May 23 12:16:07 sneak: #beagle is just an irc channel May 23 12:16:08 i'm not sure that i'm _not_ doing that. May 23 12:16:09 av500, oh yeah, make them stop *everything* to comply with your source request…way to bite the hand that feeds you. May 23 12:16:18 \o/ May 23 12:17:01 I am still serious, I dont see why one could not just download the source code that goes with a bone SW release May 23 12:17:46 which bone sw release? May 23 12:17:53 yeah, it doesn't make sense to me either. jumping through all of these git-repos-that-fetch-other-git-repos stuff is stupid. May 23 12:17:56 e.g. the one that comes with the bone May 23 12:17:56 mdp: the 14 feb one that ships on the microSDs May 23 12:17:58 when you buy one May 23 12:18:07 ahh, the circuitco production image May 23 12:18:14 mdp: any image on the web May 23 12:18:25 it's a 4gb card, no reason why the full (however patched) kernel sources shouldn't be on there. May 23 12:18:32 is it not an experimenting/educational device? May 23 12:18:46 no, that was the r-pi :) May 23 12:18:49 that's a good question, you ought to ask a cco employee May 23 12:18:53 (SCNR) May 23 12:18:53 ahh May 23 12:18:53 but those are vaporware May 23 12:19:05 mdp: indeed May 23 12:19:14 I would ask the cco s/w guy that does those releases, in fact May 23 12:19:23 mdp: i would have assumed that one or more of those would be in the official irc channel. May 23 12:19:24 he hangs out here sometimes May 23 12:19:29 maybe it's my fault for making that assumption May 23 12:20:05 I'm being facetious, of course. May 23 12:20:18 I suspect koen will comment RSN May 23 12:21:24 * mdp wonders if r-pi ships with the codez May 23 12:21:44 rpi is run by a broadcom employee May 23 12:21:57 i would expect intentional obfuscation/hoop-jumping from them May 23 12:21:58 shhh May 23 12:22:03 that's a sekrit May 23 12:22:35 once they actually ramp up production, it's a v. clever model for selling lots of their SoC May 23 12:22:38 until that time, it's vaporware, the market has priced their limited supply at the same price point as the bbone basically May 23 12:22:45 (if you check ebay) May 23 12:22:55 imagine if beagleboard.org had TIers involved…that would blow the top off of everything May 23 12:23:06 why would it? May 23 12:25:24 :) May 23 12:25:41 we do need clients to display sartalics May 23 12:26:02 Hi, May 23 12:26:14 smileys got outdated? May 23 12:26:14 or fewer sperglords on irc May 23 12:26:14 xor May 23 12:26:30 indeed May 23 12:26:40 I have compiled a library for the C64 arch. DSP, and was wondering how to use it along with a natively running ARM application. May 23 12:27:08 aholler, yes, doomed by the web people and their markup May 23 12:28:00 mbixx988: dsplink? May 23 12:29:50 av500: Sorry? I am referring to the "HD video capable TMS320C64x+ DSP." as listed at http://beagleboard.org/hardware May 23 12:29:55 "Intended to work standalone or in conjunction with existing BeagleBoard and BeagleBoard-xM, the BeagleBone Community Development Board is targeted at application developers enabling easy access to industry standard I/O's." May 23 12:30:02 av500: the answer to the gpl source request would still be a link to http://www.angstrom-distribution.org/building-angstrom May 23 12:30:04 easy access == ship with the code. May 23 12:30:38 i mailed cco re: gpl. hopefully they'll live up to the "easy access" part. May 23 12:30:47 sneak: http://www.angstrom-distribution.org/building-angstrom May 23 12:30:55 sneak: that's how you get the source and buildscripts May 23 12:31:08 koen: tried it, on two machines, beaglebone included. didn't work. May 23 12:31:13 too bad May 23 12:31:18 the gpl doesn't cover that May 23 12:31:40 i should not be required to execute code (scripts) to get the source. May 23 12:31:53 wget isnt code ? May 23 12:32:00 its not May 23 12:32:02 if you don't believe me, mail rms. May 23 12:32:02 git is. May 23 12:32:03 in that regard May 23 12:32:03 your email program isn't code? May 23 12:32:04 oebb.sh is May 23 12:32:10 _your_ code. May 23 12:32:11 koen, seems like a reasonable effort to comply..you even went above and beyond the "on request" and just make it available to everybody regardless of distribution…nice. :) May 23 12:32:17 or aangstroem's May 23 12:32:35 it also of course possible to pick all patches by hand, they're all freely available. May 23 12:32:45 fact is, it's basically inexcusable that the kernel source doesn't _ship with the device_ May 23 12:32:51 LetoThe2nd: where? May 23 12:32:59 (this is what i've been asking for an hour.) May 23 12:33:07 koen: I would not accept that May 23 12:33:09 why would that be inexcusable May 23 12:33:17 granted, its not pretty - but i don't think that gpl needs you provide everything in one chunk. May 23 12:33:30 LetoThe2nd: i'd be happy with a url to a directory full of patches. May 23 12:33:30 sneak: no, you are bi***ching about not getting one big tarball. May 23 12:33:37 sneak: tried googling for "beaglebone kernel source" ? May 23 12:33:56 every link on that page of google results is purple (visited) May 23 12:33:56 so yes May 23 12:34:57 don't even get me started on the hoops one has to jump through to even get mkimage to work on the bbone May 23 12:35:47 please start May 23 12:35:59 let me tell you a secret: the software isn't built on a bone May 23 12:36:13 koen: then why does it ship with a compiler? May 23 12:36:23 (don't answer that, it's rhetorical.) May 23 12:36:28 koen: so per your reasoning a windows user is not entitled to get the source code? of a GPL product he bought? May 23 12:36:32 sneak: see http://git.yoctoproject.org for example, and the meta-ti tree. its full of patches, go ahead and pick whatever you like. May 23 12:36:47 LetoThe2nd: no, not what he likes May 23 12:36:51 what *ships* May 23 12:37:04 exactly May 23 12:37:12 av500: i thought it was weird, too. May 23 12:37:42 being able to actually BUILD what ships is another point May 23 12:37:44 sneak: that complain is as old as beagleboard.org May 23 12:37:46 av500: like i said - its not pretty, and i agree on that. but i don't agree on "i want that now, and will go on ranting just because i belive its my god given right to get the tarball i deserve" May 23 12:38:00 LetoThe2nd: but he deserves a tarball May 23 12:38:04 or a git and a hash May 23 12:38:11 or a svn and a version number/tag May 23 12:38:29 like in ye olde days May 23 12:38:32 git to fetch oebb.sh, oebb.sh to fetch bitbake, bitbake to fetch patches May 23 12:38:45 bash and awk and weird git version to run oebb.sh May 23 12:38:53 sure, i can jump through all of these hoops. May 23 12:39:00 or you could be normal and do as av500 suggests May 23 12:39:11 and live up to the "easy access" phrase in the product description. May 23 12:39:16 and the source code is already there during a build May 23 12:39:26 all that needs to be done is to run tar May 23 12:39:30 solved May 23 12:39:37 or, better yet, throw the patched sources on the SD card May 23 12:39:59 av500: that would be a 30gb tarball if you do it like that May 23 12:40:10 av500: well, if you do it on the buildserver, a 260gb tarball May 23 12:40:15 aholler: if that complain is so old, why has not anybody just adressed and fixed it? thats my point. because complaining is easier then fixing. May 23 12:40:20 koen: the sources angstrom uses for 99% of the software don't have special patches. May 23 12:40:23 koen: 260gb tarred sources? May 23 12:40:34 there are already wgettable links to the sources for, say, python and perl and node. May 23 12:40:40 LetoThe2nd: those who could fix it, never did May 23 12:40:41 the kernel is the only unique part May 23 12:40:48 man git-archive ;) May 23 12:41:01 and the kernel is not 260gb of tarred sources. May 23 12:41:04 aholler: and those who could just don't care. the usual. May 23 12:41:13 koen: what, you have to source code to all windows version since 1.0 too? May 23 12:41:21 can haz? May 23 12:41:26 I'll send you a 260GB hdd May 23 12:41:36 dr-dos as well May 23 12:41:40 \o/ May 23 12:42:02 speaking of May 23 12:42:02 ah, finally we are approaching "oh it doesn't fit onto the sd... well then just the patches. and there only those that i am personlly interested in." May 23 12:42:15 i wonder if qemu would run on the bbone to get dosbox games running on a tv-out May 23 12:42:32 there is qemu for arm May 23 12:42:50 i admit i'm in with koen on that. all or nothing, and since all is impossible, its nothing then. May 23 12:43:16 a git repo from the kernel is about 500mb, that should fit on the sd-card. a one-liner which includes a git-url and a commit-number is even smaller May 23 12:43:31 aholler: it does not have to be a git May 23 12:43:37 the GPL does not talk about git May 23 12:43:42 it talks about source code May 23 12:43:57 sure. thats the a*-way May 23 12:44:14 the what? May 23 12:44:24 the way archos uses May 23 12:44:29 I would not accept something in tar format May 23 12:44:40 git-patches are much better, they include comments May 23 12:44:46 sure May 23 12:45:01 maybe GPLv4 will cover that May 23 12:45:21 so one could just add the output from git-format-patch on top of a well known kernel-release May 23 12:45:29 ontothe sd-cards May 23 12:46:06 along with a description, git clone foo; git am dir/* May 23 12:46:13 * ogra_ doesnt want to see that with the TI panda kernels May 23 12:46:22 that would be wonderful May 23 12:46:37 i don't care about tracking kernel dev, i would just like to be able to modify the software that is running on it presently May 23 12:46:39 ogra_: arnd and andy don't agree on rebasing May 23 12:46:42 sneak, I've got all the source code from a build and it comes out to a 90MB .tar.bz2 May 23 12:46:55 the GPL does not cover rebasing either May 23 12:46:56 koen, heh, yeah May 23 12:47:08 exosyst: linux-3.4.tar.bz2 is 76MB May 23 12:47:29 where's that other 14mB of bz2 come from May 23 12:47:29 platform support? May 23 12:47:29 did you distclean? May 23 12:47:38 * av500 finds himself in the strange position having the explain the GPL to open source heroes.... May 23 12:47:50 av500: strange, innit? May 23 12:48:56 sneak, Has built binaries and patches applied May 23 12:49:26 i already have the binaries... May 23 12:49:26 in fact, i already have _most_ of the source May 23 12:49:41 mind doing a "make clean" and diffing against a kernel.org tree? i bet that bz2 is substantially smaller. May 23 12:49:54 i'll even host it. May 23 12:51:45 git format-patch v3.2; tar ...; done May 23 12:55:27 or just don't ship a sd-card with the board. May 23 12:55:46 aholler: and no images on the net May 23 12:56:02 only build.sh May 23 12:56:54 sneak, I can send you the archive I have. I don't want to mess with cleans and diffs May 23 12:57:58 exosyst: i'd be grateful. May 23 12:58:15 sneak, You have an FTP/scp account I can throw it? May 23 12:58:57 "python -m http.server" serves the current directory at :8000/ May 23 12:59:07 i'd rather not swap credentials May 23 12:59:22 sneak, Yeah that's not gonna work on this network May 23 13:00:11 if i had spare boxes around for you to log in to, i would just be jumping through all these openembedded hoops. :( May 23 13:00:26 i'm a kicksend.com user if that helps. May 23 13:04:33 sneak, dropcanvas May 23 13:04:51 sneak, http://dropcanvas.com/e0g0g even* May 23 13:05:36 (uploading still? i get a blank page) May 23 13:06:24 Yeah - net here is dreadful over lunch May 23 13:08:27 thanks again for helpin' out May 23 13:11:54 sneak, no probs. There was a better sharing site where you share it, click the link and it does a direct transfer rather than me > server upload > you but I can't remember it May 23 13:16:43 Half way uploaded... wow this connection sucks :D May 23 13:24:15 somebody has running fedora 17 on beaglebone? May 23 13:31:39 I can only imagine how painfully slow building linux on the beaglebone will be May 23 13:32:21 about 4h May 23 13:33:13 aholler, Oh dear. I read somewhere that with an SSD, an i7 running at stupidHertz and with stupidGB of RAM it can be built in a minute. May 23 13:33:23 yes May 23 13:33:27 takes 4h on the panda on a fast USB disk May 23 13:33:34 Nice article about the beaglebone being used as a server http://www.ewanleith.com/blog/956/my-60-arm-server May 23 13:33:37 so should be a bit more on the bone May 23 13:33:45 anyway, howoften do people build a kernel? May 23 13:33:55 depends what they're messing with :D May 23 13:33:56 about once per hour May 23 13:34:07 oh, thats an ubuntu package build i'm looking at ... disregard the numbers, that does multiple build runs May 23 13:34:09 sometimes every minute May 23 13:34:22 you should not call clean with every make ;) May 23 13:34:33 I dont May 23 13:34:58 on my ws a full kernel needs about a minute May 23 13:35:11 on the beagleboard around 3h or such May 23 13:35:22 (usb-disk) May 23 13:36:43 what I'm still missing is vm-setup to easily debug the (x86) kernel inside a vm. May 23 13:37:06 might be the easiest way to use kgdb May 23 13:41:45 I wonder if I am in trouble for trolling the panda list? May 23 13:44:32 you are fine May 23 13:47:22 heh May 23 13:47:29 I finallly got around to reading the response May 23 13:47:40 looks like someone lowered expectations quickly :) May 23 13:49:49 yes, a spoilsport.. May 23 13:50:30 pfft - a watched upload never completes May 23 13:51:01 hi, i am trying to get network connection working from my beaglebone to my pc running ubuntu 12.04 May 23 13:51:26 i had setup a dhcp server on the system, but bone fails to acquire an IP from that May 23 13:52:03 k4rtik, I'd recommend setting a static - it's what I did on my bone May 23 13:52:19 I am manually trying to configure the IP, but still "route add default gw 192.168.91.10" gives me invalid argument May 23 13:53:06 one more thing, I am on TI Android ICS image currently May 23 13:54:12 exosyst: could you tell me how to do that? May 23 13:54:30 Can you ping 192.168.91.10? If not then it won't work as a gateway May 23 13:55:01 k4rtik, Not sure on that one, so does ifconfig eth0 192.168.1.5 netmask 255.255.255.0 up not work? May 23 13:55:07 fiola: nope, 91.10 is my PC's IP. But I can't ping it May 23 13:55:22 exosyst: that command worked, but not the route one May 23 13:55:43 k4rtik, So you have an IP address then? May 23 13:56:09 android has ifconfig? May 23 13:56:19 Well you have to take routing one step at a time, sorting out the local IPs and masks etc before it'll reach the next hop etc May 23 13:56:44 k4rtik, udhcpc should grab you a DHCP address AFAIK. Not sure if it's on the image May 23 13:57:25 fiola: I just have a single crosswire connection between my PC and bone, isn't that a single hop May 23 13:57:32 Don't use ifconfig peeps, it's old, and doesn't work well with IPv6. Use iproute2 package. ip route add default via 192.168.91.10 dev eth0 --- equivalent to your ifconfig May 23 13:57:55 ip route ls -- shows what you have currently May 23 13:58:21 fiola: android has iproute2? May 23 13:58:30 exosyst: i tried discovering with DHCP too, but that failed May 23 13:58:35 Dunno, but BeagleBone does May 23 13:58:36 av500: yes it has all these :P May 23 13:58:59 jkridner: can you confirm or deny that c69c933acf48b22411dd5cca3d0b2e2a9d148ae1 of bonescript should be in the a6 image? May 23 13:59:23 fiola, I didn't know that - I doubt many people will bother with it until ifconfig disappears May 23 13:59:32 fiola: ip route ls gives default via 192.168.91.10 dev eth0 May 23 13:59:34 192.168.91.0/24 dev eth0 proto kernel scope link src 192.168.91.15 May 23 14:00:09 91.15 is what I assigned using ifconfig eth0 earlier for bone May 23 14:01:42 Well if they're both on 192.168.91.0/24 then you shouldn't even need a route on either machine, no vias. May 23 14:02:15 fiola: but it still doesn't ping May 23 14:02:27 fiola: anything else I can try? May 23 14:02:37 koen: on a call... will look in a bit. May 23 14:02:51 Run a tcpdump on the far end to see if anything is arriving May 23 14:03:31 it says no suitable device found May 23 14:03:40 jkridner: audio on dvi cape working, emberassingly enough the fix was hidden in an e2e post May 23 14:03:51 great! May 23 14:04:11 k4rtik: prolly needs to be root to run tcpdump in promiscuous mode May 23 14:04:46 fiola: right, my bad May 23 14:06:09 No filewall on either device hopefully, or that would interfere. May 23 14:06:42 fire* May 23 14:06:42 fiola: it gives lot of lines of data, shall i stop my wifi temporarily May 23 14:07:08 No don't stop wifi, use a more restricted tcpdump. eg May 23 14:08:34 tcpdump host May 23 14:09:48 Or ... tcpdump net 192.168.91.0/24 May 23 14:13:37 busybox knows ip since some years May 23 14:15:43 sneak, My end says it's done, I loaded up the site in chrome but couldn't see it - can you? May 23 14:17:01 When the hell did libusbx appear? May 23 14:23:10 exosyst: it is not showing up, no :( May 23 14:23:14 sneak, Well, if you're still there - the link is now http://files1.jetbytes.com/c1c27dc1d0774220b3de7bf3eca3b03b it's p2p May 23 14:23:27 downloading now, thanks :) May 23 14:23:57 dusend was the service I was thinking about but it got shuttered not so long ago. It was p2p and was very cool May 23 14:24:06 I should write one of these for my own site! May 23 14:24:06 that's pretty interesting May 23 14:24:07 i wonder how it works May 23 14:24:28 i'm surprised there isn't some simple wordpress plugin for doing this or something May 23 14:24:56 sneak, No idea. It's fantastically simple. I had hoped that the dropcanvas one would work. But p2p is always better for this sort of thing. May 23 14:25:32 sneak, says you cancelled? May 23 14:26:17 sneak, http://files1.jetbytes.com/cb78e372ce8dad1c76973f476dd2ebba - it does however require a stable net connection. May 23 14:26:43 it's still downloading May 23 14:27:13 killed it, using new link May 23 14:27:24 Is *that* one working? May 23 14:27:42 no, it sent the http request but no response yet... May 23 14:27:51 fuck it, i'll fire up a linux vm for beagle development May 23 14:28:03 fiola: sorry, i went offline trying to get tcpdump May 23 14:28:09 i'll just install debian and the whole toolchain and jump through all these stupid hoops May 23 14:28:18 maybe i'll even post patches for people who don't want to deal with the bullshit May 23 14:28:41 sneak, I'd recommend it. It's always good to post stuff May 23 14:29:48 at least ubunturd plays well with virtualbox May 23 14:29:54 going through irc logs, and sending a proper tcpdump soon May 23 14:30:21 k4rtik: Ah welcome back. May 23 14:31:04 :) May 23 14:33:05 k4rtik: Seems a but odd to me that they're not communicating when they're both on the same IP segment though (the gateway must not be specified for that to work, since it's on same /24). May 23 14:33:53 fiola: yes to me too May 23 14:33:59 k4rtik: Double check that they're both /24 and have same netmask May 23 14:35:33 k4rtik: Since we're only checking ping currently, narrow your tcpdump to showing ICMP only --- tcpdump icmp May 23 14:36:58 okay May 23 14:36:58 Then try your ping from the beagle again. You should get two lines for every ping, a to and a from May 23 14:37:34 till now, i am getting destination host unreachable from bone May 23 14:38:27 fiola: double checked, same netmask. OTOH does gateway matter on the other side? May 23 14:39:16 fiola: no icmp data from tcpdump, on bone same host unreachable May 23 14:39:41 The gateway shouldn't have any effect because default route is less specific than the implicit route of sending packets on one's own IP segment May 23 14:40:42 In other words, a host always knows that packets with a dest address within the /N subject of its own interface has to go out on that interface May 23 14:41:28 Just to be sure, remove all routing statements on beagle May 23 14:42:04 (I'm assuming you're talking to it on serial console) May 23 14:42:43 fiola: yes, using screen for that, i had restarted the board after getting disconnected from here May 23 14:42:53 so no routing info currently May 23 14:43:12 just ifconfig eth0 to 90.15 and proper netmask May 23 14:44:32 I guess you'd better check that the broadcast address has all-1's in the host part, ie. ends in .255 May 23 14:44:53 (On a /24) May 23 14:45:29 ifconfig on my PC reads inet addr:192.168.91.10 Bcast:192.168.91.255 Mask:255.255.255.0 May 23 14:45:50 you might have a look at iptables -L -v May 23 14:45:53 Looks OK May 23 14:45:56 and ifconfig eth0 on bone reads: May 23 14:45:57 eth0: ip 192.168.91.15 mask 255.255.255.0 flags [up broadcast running multicast] May 23 14:46:46 aholler: what should that report? May 23 14:46:50 oebb.sh complains that /bin/sh is a shortcut to something that isn't bash May 23 14:47:11 why doesn't it just use /bin/bash instead of complaining May 23 14:47:12 what the fuck May 23 14:47:19 my WTFs/minute using all the software surrounding the beaglebone has rarely been higher May 23 14:47:27 k4rtik: don't know, depends if you have a firewall enabled which might block the beagle May 23 14:47:44 sneak: oebb.sh is not meant to be run on the bone May 23 14:47:47 but you know that already May 23 14:47:55 i'm running it on a fresh ubuntu 11 install May 23 14:48:13 thats dash May 23 14:48:24 I assume May 23 14:48:33 yes, i realize that May 23 14:48:33 then why state May 23 14:48:35 aholler: i got: May 23 14:48:40 Chain INPUT (policy ACCEPT 0 packets, 0 bytes) pkts bytes target prot opt in out source destination May 23 14:48:43 Chain FORWARD (policy ACCEPT 0 packets, 0 bytes) pkts bytes target prot opt in out source destination May 23 14:48:46 Chain OUTPUT (policy ACCEPT 0 packets, 0 bytes) pkts bytes target prot opt in out source destination May 23 14:48:59 so, no fw. good May 23 14:49:03 (or bad) ;) May 23 14:49:16 oebb.sh actually outputs "/bin/sh is a symlink to dash, please point it to bash instead" May 23 14:49:16 instead of just using /bin/bash to exec shit May 23 14:49:55 k4rtik: It's a long shot, but I'm wondering if your cable isn't bad :P May 23 14:50:12 also oebb.sh fails silently if gawk isn't installed May 23 14:50:19 such a clusterufck May 23 14:50:21 I assume that both pieces of equipment are modern enough for crossover lead detection to be automatic May 23 14:50:56 fiola: or the ethernet port on bone, 'cause the wire works well when i connect to other systems from PC, just transferred over 1 GB really fast over that cable May 23 14:50:57 no, only gbit-if support crossover by default May 23 14:51:34 aholler: bone doesn't have gigabit, is that the source of trouble? May 23 14:51:40 on plain old 100mbit-ifs you need luck. May 23 14:51:48 aholler: auto-mdix is mandated but the gbit spec, yes, but you can add it to your 100mbit design if you want May 23 14:52:16 koen: therefor "by default" May 23 14:52:20 Erm ... I haven't seen any equipment that doesn't do auto crossover for a decade. And I have tons of Ethernet gear here :P May 23 14:52:35 * fiola goes looking for cables to check May 23 14:52:56 sneak: a rather complex build system having prerequisites is nothing to rant about here May 23 14:53:01 hi there! May 23 14:53:13 i was complaining about silent failures and bad code. May 23 14:53:21 I'm trying to get the mt9p031 sensor to work May 23 14:53:22 not prerequisites. (bash was installed.) May 23 14:53:27 * LetoThe2nd would really vote for reducing the amount of foul language a bit. it helps nobody. May 23 14:53:37 has anyone already achieved that? May 23 14:53:37 +1 May 23 14:53:49 * k4rtik wants to go for dinner, catch you guys later May 23 14:53:52 LetoThe2nd: your idea is bad and you should feel bad May 23 14:54:11 ah yes. indeed. May 23 14:54:21 *plönk* (finally) May 23 14:55:01 go read children of the anus of the sandworm of the emperor of dune and relax May 23 14:55:01 or at least try May 23 14:55:37 Oh crap. Hint: don't try to push Ethernet lead tino beaglebone by resting the other end on the table. Because the other end has an SD card stick out ..... May 23 14:55:45 into* May 23 14:55:47 haha May 23 14:55:57 fsck time May 23 14:56:34 that's basically my whole day May 23 14:56:40 fsck this, fsck that, fsck everything May 23 14:56:44 fsck you, fsck him May 23 14:57:00 get a valium May 23 14:57:03 koen: I see you have got the mt9p031 working... does it work with the oe mainline kernel (3.2.16)? May 23 14:57:53 my beagle just grown up, it got trained ;) May 23 14:57:55 jkridner: ping May 23 14:58:19 pong May 23 14:58:58 * jkridner puts on op jacket May 23 14:59:20 fiola, I kept bumping the sd card on my bbone until I got a case for it May 23 14:59:54 Mojito: very sensible. I'll do that asap May 23 15:00:41 sp00k: haven't tried with 3.2 yet May 23 15:01:14 The one at adafuit.com is good May 23 15:01:26 adafruit.com May 23 15:02:32 do I also need to configure the VPFE Video Capture Driver (for beagleboard)? May 23 15:03:54 because after booting the cloud9-gnome-image I don't see a /dev/media or a /dev/video0 (OMAP 3 Camera support is enabled) May 23 15:07:20 sp00k: there's no vfpe, only isp May 23 15:07:43 sp00k: the camera stuff is on my todo list, but far, far down May 23 15:08:39 ok, but the ISP should work? May 23 15:09:16 yes, the ISP driver has improved a lot May 23 15:09:35 what's missing the the glue for the sensor driver and isp into the beagleboard board-omap3beagle.c file May 23 15:09:51 k4rtik: I've just confirmed that beagle running default Angstrom needs no routing statements in order to reply to incoming pings on same subnet. (As expected) May 23 15:10:00 the (old) patches from javier or the 3.0 ones would be a good starting point May 23 15:10:31 I got an implementation (board-omap3beagle-camera.c) of max galemin May 23 15:10:58 http://blog.galemin.com/2012/03/buildroot-2012-02-for-beagleboard-xm-with-li-5m03-mt9p031-camera-support/ May 23 15:11:02 the javier ones are better IMO May 23 15:11:08 max cuts too many corners May 23 15:11:16 (but if it works for you, go for it) May 23 15:11:27 thx for the hint May 23 15:11:43 what should I enable to get the ISP working? May 23 15:12:07 right now I can't see the /dev/media to use media-ctl May 23 15:12:27 that's what those patches are for :) May 23 15:13:17 the patches which are already included in the linux-mainline of oe? May 23 15:13:33 or the patches from javier? May 23 15:17:59 sorry, but I really don't get it... Are you saying that the ISP should work out of the box with the OE mainline kernel? May 23 15:21:13 teh ISP works, but you need to add stuff to the boardfile to activate it May 23 15:22:07 I'm just asking because if I modprobe omap3-isp, it says "not found" May 23 15:22:56 (and I have configured the "OMAP 3 Camera Support" as a module) May 23 15:23:15 Is there a document somewhere about how the new systemd-based network config works? May 23 15:23:32 fiola: I used google May 23 15:24:10 google indexed info mostly refers to pre-systemd May 23 15:24:51 Because of this change, most BeagleBone docs about startup seem to have got obsoleted May 23 15:25:44 Is there a Beagle wiki we can use to make doc pages? May 23 15:26:02 elinux is used often May 23 15:26:03 elinux May 23 15:26:09 kk May 23 15:27:12 if you need a tutorial for systemd: http://www.h-online.com/open/features/Booting-up-Tools-and-tips-for-systemd-1570630.html May 23 15:27:25 cheers May 23 15:27:50 but for connman you have to search yourself May 23 15:27:54 fiola: the network config is connmann now May 23 15:28:23 Oh, I thought connman was using systemd. Hmmm May 23 15:28:37 systemd just starts stuff May 23 15:29:49 Oh neat, elinux seems fairly comprehensive on beagle. Worth adding to that May 23 15:51:53 * k4rtik is back May 23 16:04:53 COPAIN! May 23 16:16:45 Oh lord, I'm leaving a huge trail of *~ files behind me every time I edit something on beaglebone. Is that a vim config setting? May 23 16:17:09 I guess so May 23 16:17:17 it annoys the crap out of me May 23 16:17:35 use joe May 23 16:17:44 Ditto. I'll try to chase it up, it's annoying the hell out of me May 23 16:17:49 littering May 23 16:18:04 set nobackup May 23 16:18:15 scromp++ :-) May 23 16:18:38 can just put that in your .vimrc May 23 16:18:47 Yep May 23 16:19:57 I already have stuff like :syntax off, and :set noautoindent in there, but I guess :set nobackup must be the default on Gentoo May 23 16:20:51 Oh maybe there's a global default settings file for vim, would make sense May 23 16:21:26 Crofton is the one to blame for the default vimrc in angstrom May 23 16:21:42 Yes there is, /etc/vim/vimrc May 23 16:22:09 heh May 23 16:22:18 to be fair, before Croftons change it was running old-school mode May 23 16:22:20 the previous situation was worse May 23 16:22:28 I can't begin to describe how annoying that was May 23 16:22:31 I think I just installed hte default May 23 16:22:48 the busybox vim was less annoying that the real one in old school mode May 23 16:23:00 that's how bad the vim defaults are without a vimrc May 23 16:23:12 BeagleBone A5 default doesn't have an /etc/vim ... I'd better strings the binary first May 23 16:25:20 just copy the one from gentoo May 23 16:29:29 The vim binary had no "etc" visible in strings, although it might have been doing it craftily so I checked anyway. Result: it doesn't use /etc/vim/vimrc. However, the usual per-user .vimrc's work as expected. May 23 16:30:12 No strace on my beagle yet, or I'd have used that to check May 23 16:30:44 opkg install strace :) May 23 16:31:10 Whoop, hehe. Now a find / to remove all the littering of *~ I've done without realizing it May 23 16:31:44 koen: one step at a time. So far I only checked networking up to my gateway. :P May 23 16:31:54 * fiola is slow but methodical May 23 16:32:07 * mranostay yawns May 23 16:32:14 * mranostay yawns violently May 23 16:32:37 * fiola sticks a chili in mranostay's open gob May 23 16:34:50 I think my first opkg is going to be man :P May 23 16:47:47 jkridner: you can opkg upgrade to .23 or dd http://dominion.thruhere.net/koen/angstrom/beaglebone/Angstrom-Cloud9-IDE-GNOME-eglibc-ipk-v2012.05-beaglebone-2012.05.23.img.xz May 23 16:48:01 jkridner: give it a try, it's close to final May 23 16:54:39 fiola: you had mentioned IPv6 somewhere, do you think configuring a DHCP server with IPv6 support will help my bone acquire an IP address? May 23 16:58:24 after trying everything out, it seems to me that something is probably wrong physically, either the port or something else on the board May 23 16:58:46 k4rtik: IPv4 and IPv6 are entirely separate, so I don;t think so. In any case, BeagleBone does IPv6 out of the box, which is a lot better than most distros May 23 17:00:59 All the networking "just worked" on my beaglebone, after I chose an IP address, both IPv4 and IPv6. I've figured out connman config now I think, although haven't rebooted to test it yet May 23 17:01:48 Currently trying to figure out why opkg doesn't see any opkg feeds. It can reach the hosts, I checked their port 80 May 23 17:01:49 fiola: but if it is the case that Bone supports ipv6 out of the box and doesn't support ipv4 so nicely, plus I received "eth0: No IPv6 routers present" in dmesg in both bone and PC May 23 17:01:50 what does PSP mean May 23 17:02:36 fiola: connman, is this related to my query? May 23 17:03:15 k4rtik: IPv6 has autoconfiguration, so it's actually easier to configure than IPv4. People just get stuck on the long addresses May 23 17:03:55 that's why you also use e.g. mdns with ipv6 May 23 17:03:58 fiola: oh, probably that's why I am unable to find any documentation on configuring DHCP for IPv6 on ubuntu May 23 17:04:11 k4rtik: No, your problem is more basic, it seems you have no MAC-level comms at all. You need the basic ping to work before you bother to configure connman I think May 23 17:04:32 right, at least ping should work May 23 17:04:35 dhcpv6 is not really popular. preferred method is RA. May 23 17:05:28 k4rtik: Yeah, DHCP is pretty much not needed for IPv6, although there is a DHCPv6 for more extended config. But nodes discover each other automatically when you link them with copper :P May 23 17:05:46 Yep, as dm8 says May 23 17:06:04 dhcp is for wimps, at least that is what the ipv6-designer though ;) May 23 17:06:05 RA? May 23 17:06:16 just start radvd somewhere, done May 23 17:06:49 router advertisement May 23 17:07:51 k4rtik: I was going to test what 'bone does when given a crossed lead, but I failed to find an RJ45 adapter to let me splice in the cross so not checked yet May 23 17:08:18 Always have half a dozen adapters around when I don't need one May 23 17:11:54 :D May 23 17:14:43 Ah, opkg isn't even attempting a network access, no wonder install fails May 23 17:14:50 * fiola scratches head May 23 17:15:37 Does opkg need to be enabled or something first? May 23 17:18:15 Oh god, connman sets resolv.conf to localhost May 23 17:18:17 lol May 23 17:28:22 My cable got upgraded to 100 meg for free, yay! May 23 17:29:26 tanstaafl May 23 17:29:35 indeed May 23 17:36:57 is that like when you forget to pay your cell phone bill and they don't disconnect you out of "courtesy"? May 23 17:40:44 Gah. Without strace I can't tell why opkg isn't going out on the net. And I can't install strace because opkg isn't working. May 23 17:40:51 I hate chicken and eggs May 23 17:48:33 fiola: do you have done opkg update? May 23 17:48:51 NOT upgrade May 23 17:49:35 koen: the bonescript on the 5/21 image does seem to match the git and includes analog.js, but I do have a couple of updates on an SD card that I'd like to push upstream, including a fix for analog.js. May 23 17:49:48 for some reason, my USB SD card adapter isn't mounting. May 23 17:49:55 aholler: nope. Does it come with a local-only version, which needs updating first? May 23 17:49:59 I see a SCSI device, but no /dev/sdX. May 23 17:50:27 http://pastebin.com/fS85AHh0 May 23 17:50:28 fiola: I know almost nothing about angstrom, just that you need opkg update before anything works May 23 17:50:45 and that upgrade breaks things May 23 17:50:49 Interesting. Thanks aholler, I'll do the magic incantation May 23 17:50:56 seems the problem was seen on the list a week ago. May 23 17:51:42 aholler: upgrade breaks things? May 23 17:52:09 acoording to this channel opkg upgrade does something bad May 23 17:52:40 I've been opkg upgrading all month, no troubles on the bone May 23 17:53:06 of course, every new kernel breaks usb in different ways, but such is TI May 23 17:53:41 koen: I've just answered because no one else does May 23 17:55:11 koen: so much fun ahead, huh? :P May 23 17:55:18 i don't even know the difference because update and upgrade May 23 17:55:26 s/because/between/ May 23 17:55:41 (on angstrom) May 23 17:56:21 same as with apt May 23 17:57:05 aholler: update == sync in gentoo speak May 23 17:58:36 koen: oh I see! :-) May 23 17:58:36 I use debian, ubuntu and fedora too, just not angstrom ;) May 23 17:58:49 jkridner: I'm ready to disable MUSB host in the beaglebone kernel May 23 17:59:31 So "update" == "tell me about all the new packages out there" May 23 17:59:59 just like apt-get May 23 18:00:18 koen: that doesn't seem useful. May 23 18:00:40 jkridner: how usefull is it currently? May 23 18:00:41 To be fair, that's exactly what opkg --help says it does :-) May 23 18:02:12 more useful than if it was disabled. May 23 18:02:24 So, same as Gentoo, opkg is just working off a local list of packages. And in BeagleBone's case, the local list == the installed list. May 23 18:02:58 and that is zero when delivered May 23 18:04:04 Whereas now, after an opkg update, my local cache contains a whole lot more than just the installed packages, so presumably it will now try a network access when asked to install. May 23 18:04:05 it's populated when deliveres, but updating it is adviced May 23 18:09:37 Woohoo, I now have strace, #1 install :P May 23 18:10:26 I guess tcpdump has to be #2. Odd that debug tools are missing from the base distro. May 23 18:12:09 tcpdump installed as #2, but doesn't work. And strace tells me why: socket(PF_NETLINK, SOCK_RAW, 12) = -1 EPROTONOSUPPORT (Protocol not supported) May 23 18:12:25 * fiola googles May 23 18:13:00 I assume I just need a module, rather than a new kernel May 23 18:24:34 does the beaglebone really have a SGX530 GPU? May 23 18:24:44 http://www.youtube.com/watch?v=ps47MOHF9x8 May 23 18:24:48 this video claims that May 23 18:24:57 how would video output even work May 23 18:26:06 heeen, http://www.ti.com/litv/pdf/spruh73e <- read start on page 184 May 23 18:26:58 If I understand it right, the TI SoC has the PowerVR GPU built in, but since the BeagleBone doesn't provide the graphics interface hardware, you won't get the graphics unless you attach a DVI-D cape May 23 18:27:34 so you could use the gpu for gpgpu stuff? May 23 18:27:50 It's just like with Ethernet. The SoC has the Ethernet MAC built in, but if the BeagleBone didn't have an Ethernet PHY on the board, you'd have no networking either. May 23 18:27:51 heeen, read the tech ref manual, it's long for a reason May 23 18:28:02 but to generate an actual signal from the framebuffer you need external hw May 23 18:28:14 ok May 23 18:29:14 heeen's right that in principle one could use it for GPGPU ... but PowerVR is closed AFAIK, same as most other GPUs. May 23 18:31:24 closed? May 23 18:31:37 more closed May 23 18:31:44 I meant as in use OpenGL to do for example signal processing May 23 18:31:44 no source, no api May 23 18:31:47 most closed May 23 18:31:48 should be possible? May 23 18:32:38 bradfa: that section only describes the gpu features May 23 18:33:07 heeen, there's also a table of contents at the beginning May 23 18:33:09 Same as Broadcom's GPU, very lame. Open source ARM world really needs an open GPU. Even MALI-400 is closed. May 23 18:33:25 fiola, if you can make that happen, millions would love you May 23 18:33:31 koen: apparently the 5/20 kernel release had the USB bug, but an update overnight fixed it. May 23 18:34:49 heeen, sorry, I'm grumpy, I thought the TPM has more info on the sgx May 23 18:34:59 "The SGX530 module does not include any external interface pins." May 23 18:35:03 bradfa: Attaching an FPGA is probably the easiest was forward for GPGPU community. BeagleBone has no shortage of low-level I/O's at least May 23 18:35:37 fiola, external fpga will kill your $ budget for consumer electronics, never mind power and form factor May 23 18:35:59 if you want hard care GPGPU, buy x86 and NVIDIA / AMD May 23 18:36:05 hmm. so what pins does the dvi shield hook up to if the gpu does not have any pin outs May 23 18:36:33 heeen, page 1268 May 23 18:36:43 bradfa except for the low power mobileFPGAs :) except those have too little gates for anything exciting. May 23 18:36:45 Worth noting that MALI-400 driver is being reverse-engineered already. And if you look at Nouveau vs the very closed nVidia precedent, there might be some light at the end of the ARM tunnel. May 23 18:37:29 so there is some low level video output May 23 18:37:31 mdp, for various definitions of "anything exciting" :) May 23 18:37:38 but to generate an actual signal from the framebuffer you need external hwd the shields convert that to DVI or hdmi May 23 18:37:49 oops May 23 18:37:56 koen: http://arago-project.org/git/projects/?p=linux-am33x.git;a=commit;h=2cdf485be20a9536093bf99306c7223954f9cd98 seems to have a fix for USB MSC host May 23 18:38:00 and the shields convert that to DVI or hdmi May 23 18:38:29 heeen, I believe the sgx is a graphics processor, it takes vectors and triangles and renders them into a matrix that can be displayed on a screen, it doesn't actually do the screen interfacing May 23 18:39:12 bradfa, within the current context ;) May 23 18:39:17 musb, never ending fountain of joy... May 23 18:39:58 I think it's a mistake to reinvent the same wheel on every device. If the Pi people insist on giving us a commodity 1080p graphics display for $25, why reinvent it again for *more* than $25? My engineering background says "reuse, don't reinvent". May 23 18:41:38 jkridner, that looks like framebuffer, not musb, am I missing something? May 23 18:41:51 if there wouldn't be that archaic armv6 glued to it May 23 18:42:30 sorry, not that particular commit, but that commit has the fix in it somewhere. May 23 18:43:53 jkridner, http://arago-project.org/git/projects/?p=linux-am33x.git;a=commit;h=a9e9e758646ee41289c0030645bbdd43c711e9c1 May 23 18:43:55 that one? May 23 18:44:54 fiola: regarding the "reuse", I only have two monitors which are used by my desktop PC, two playstation3 and two ARM boards May 23 18:45:28 fiola: and if I ever get RPi, there is still a spare HDMI connector available for it :) May 23 18:46:22 ssvb: That sounds good. Are you using Synergy from a single keyb/mouse? May 23 18:47:14 fiola: a single keyboard/mouse and x2x May 23 18:48:16 I've never used x2x, but use Synergy regularly. Installing x2x on Gentoo to see how it works. May 23 18:48:37 They sound similar. May 23 18:49:07 yes, probably they are mostly the same May 23 18:49:45 Once Pi boards are available ex-stock (ie. around 2027 :P) I'm intending to dedicate one to being nothing than a LAN Synergy node. May 23 18:50:29 * ssvb wonders if this setup survives the imminent transition to wayland, sounds like a lot of troubles are coming May 23 18:50:57 bradfa: might be it. May 23 18:51:04 wyland is the one which doesn't care about network May 23 18:51:24 jkridner, that's the only recent thing I see touching musb May 23 18:51:47 no more ssh -X or -Y :/ May 23 18:51:47 I admit, the whole am335x USB kernel config is bassackwards May 23 18:52:03 aholler: yes, but I do care, and it is quite important for me May 23 18:52:13 having to enable OTG to get Intrav(somethingsomething) in order to have USB host support... May 23 18:52:26 ssvb: I too, I use ssh -Y a lot May 23 18:52:47 and that OTG is 2 pages below the Intrav(somethingsomething) selection which doesn't come active for selecting till OTG is sorted May 23 18:52:56 in menuconfig May 23 18:53:43 Intraventra, that's my intrav(somethingsomething) May 23 18:53:51 Inventra May 23 18:54:19 which apparently is musb... May 23 18:54:28 * bradfa learns something! May 23 18:56:01 Is this an example of what Linus meant when he said that the ARM world is nuts, with each vendor doing their own thing? May 23 18:56:23 fiola, if this was Linus' kernel, then possibly, but it's not, it's TI's fun thing May 23 18:56:48 whole other ball game May 23 18:56:59 bradfa, that's the mentor brand of a line of IP cores…Inventra..fwiw May 23 18:57:05 Oh, sorry, thought we were talking about BeagleBone's Linux, hehe May 23 18:57:13 fiola, we are May 23 18:57:27 actually the musb IP core is quite common in the ARM world... :/ May 23 18:57:33 * fiola scratches head May 23 18:57:39 * bradfa too May 23 18:58:02 musb has other vendor specific IP around it too May 23 18:58:55 if you haven't yet noticed that even within TI parts, the DMA is sometimes handled by the mentor dma core, sometimes by SDMA (known to some as "OMAP DMA") and then in am335x's case..it's the CPPI 4.1 DMA IP core May 23 18:59:37 mdp, where's edma fit here? May 23 18:59:48 it doesn't May 23 18:59:59 so that's in hardware too, just not in kernel yet May 23 19:00:00 ? May 23 19:00:07 not integrated with musb May 23 19:00:15 oh, I meant just in general May 23 19:00:22 cppi 4.1 dma is a usb specific dma core May 23 19:00:30 ok May 23 19:00:47 edma is a general purpose dma engine…like sdma…only filled with more mostly useless features May 23 19:01:31 that provides the dma controller for most of the other ip cores…spi, uarts, etc. that you use May 23 19:01:45 is e for enhanced? May 23 19:01:53 yes May 23 19:01:55 or are there any good app notes on this stuff? May 23 19:02:22 the TRM is quite good alone... May 23 19:02:27 k, will read May 23 19:02:37 however, theere's some pages on processors... May 23 19:02:38 * bradfa looks innocent after telling heeen to read trm... May 23 19:03:11 Windows seems so alien. May 23 19:04:20 bradfa, it's much more flexible than a few other dma engines I've worked with..you can pretty much define an arbitrary number of transfer types..from byte swapping to whatever scheme you want May 23 19:04:25 Hi, looking for an efficient fft for my bone. FFTW hasnt shown impressive and OpenMax Angstrom doesnt have fft but only codec to my knowledge. Anyone has som hints for me? May 23 19:04:47 bradfa, good bedtime reading May 23 19:05:10 mdp, sounds like it, will load trm on ipad and try not to drop it on the floor when passing out May 23 19:05:38 I do a lot of this pdf reading on my tablet as well these days May 23 19:06:21 Beaglo, what would fftw have to do to impress you? May 23 19:06:48 mdp, so the current work going for edma, none of that is in PSP kernel, right? May 23 19:07:17 no, this is for mainline, specifically on mach-davinci/ platforms atm May 23 19:08:08 still in the get it all going and converting drivers from using the old raw edma driver mode..then will post an rfc for the dmaengine conversion May 23 19:08:19 Right now it takes some 80uS to run an 256 pt FFT ...accordong to OpenMax DL (not available to angstrom) it can run it at 3.8uS using Neon May 23 19:08:41 bradfa, after that, the plan is refactor the old edma stuff up into the dmaengine edma wrapper…to enable on am335x May 23 19:09:48 Beaglo, can't you build openmax dl yourself? May 23 19:11:30 Yes but to my knowledge i then have to use the realview arm compiler May 23 19:12:14 bradfa, to clarify, none of this work is applicable to the PSP kernel as the drivers are already hacked to use the proprietary edma api or a mix of that and the sdma2edma wrapper hack May 23 19:12:59 mdp, ok, thanks May 23 19:13:02 just rewrite the musb-driver and everyone is happy May 23 19:15:15 aholler, my solution is to just not use usb on beaglebone. problem solved. May 23 19:16:13 don't forget the other beagles May 23 19:16:33 I have USB working... May 23 19:16:40 Given IPv6, there's barely a point to USB anymore. It's plug'n'play. May 23 19:17:01 ? May 23 19:17:16 fiola, mice on ipv6? May 23 19:17:52 on the bluetooth cape? May 23 19:18:19 bradfa: the manufacturers aren't ready for it, but the protocol is. And since there are mice on IPv4 (I have one, HP wifi mouse), no reason why not on IPv6. May 23 19:18:42 fiola, didn't know it could do that... May 23 19:18:44 just need a ps/2 cape May 23 19:18:45 neat May 23 19:19:43 It would be wonderful if all devices were available to anything on your LAN. No need for all this Synergy or x2x stuff we're using. May 23 19:19:59 And gigabit doesn't cost much anymore May 23 19:20:09 a wifi-mouse? how often you have to recharge it? May 23 19:20:20 twice a day? May 23 19:20:47 PoWiFi ;) May 23 19:20:50 aholler: I bought it speculatively. Hasn't been used since I discovered Linux has no drivers for it, sadly. But it will. HP is a major brand. May 23 19:21:26 it's probably sitting on a witricity pod May 23 19:21:34 seems like a feature that should be built into your handset…no need for a separate mouse device May 23 19:22:16 bradfa, usb works anecdotally May 23 19:22:42 aholler: It says "9 month battery life" on the packet :P May 23 19:23:21 with wifi? I don't believe that if it has normal dimensions May 23 19:24:40 aholler: If it can turn on power only for the packet transfers, no reason why it shouldn't be frugal. After all, duty cycle can be extremely low at high Mbps May 23 19:26:23 Maybe not so low for a Quake/CS fanatic, but for normal computing use :P May 23 19:28:29 jkridner: I have all the psp musb patches + more for 3.2.x stable May 23 19:33:04 does lxr not work anymore? May 23 19:35:43 I tried pulling down the 2.6.18, but got the same issue. May 23 19:38:15 MVL5 kernels? May 23 19:38:47 bradfa: I know what a GPU is ;) May 23 19:39:05 general purpose unix May 23 19:39:33 so the ps3 has special purpose unix? May 23 19:40:07 you're on the right track as always May 23 19:40:11 I just don't know at which point they stopped to embed features into the SoC, as in we're going to integrate the gpu and a framebuffer but no hdmi generator May 23 19:40:35 I've been building my own kernels on x86 forever. Is this going to be any different here for BeagleBone when I start with that? (Not for a while yet, still exploring) May 23 19:41:19 yes, it's more adventures May 23 19:41:24 haha May 23 19:42:30 many people have this one on their bucket list May 23 19:42:42 I think it is advised to use TI ppa builds? May 23 19:42:46 sky diviing…everest…beaglebone kernels May 23 19:42:48 if you want omap stuff? May 23 19:42:59 someone correct me May 23 19:49:38 MVL5 kernels?!!?? May 23 19:58:27 bradfa: I think the external interface is lvds? May 23 19:58:41 so it could interface some lcds directly May 23 19:59:00 lcds as in panels, not devices to plug in May 23 20:00:12 ds2, it's the future May 23 20:00:28 mdp: indeed May 23 20:02:31 Wow man. The jury managed to see through the ridiculous patent claims of Oracle, despite total lack of technical experience. Verdict: no patent infringement in Android. May 23 20:04:54 just to tell you, because it was long fight (for me ;-): all my TVs and Apple products now can see DLNA/uPNP server (mediatomb), which is Ubuntu on beaglebone. May 23 20:04:59 hmm, someone has renamed the names for exported gpios from gpioNR to gpiochipNR May 23 20:06:18 now you have the problem how to connect storage ;) May 23 20:06:27 aholler: connected May 23 20:06:38 via USB May 23 20:07:03 and that works reliable with musb? May 23 20:07:27 just testing May 23 20:07:39 I think it depens on recoding as well May 23 20:08:28 but there is 10pm in my house = I can not test as all people are watching something already ;-) May 23 20:10:54 aholler: reliable and musb never should be in the same line May 23 20:11:39 Therefor I'm asking May 23 20:12:12 hm, I am very new in ARM... What can happen? May 23 20:12:19 musb can be reliable May 23 20:23:33 If I make mrproper and make clean, when I try and rebuild it blows away the am335x-pm-firmware.bin - Is there a patch I can apply to fix this? May 23 20:24:32 mrproper blows away your config too May 23 20:24:52 it is designed to do that May 23 20:24:52 yeah but I've created a beaglebone_defconfig May 23 20:25:31 so that's not an issue. I'm making a patch I can apply to a vanilla 3.2.14 kernel but that firmware is the last bug May 23 22:05:41 1 May 23 22:20:28 2 May 23 22:20:53 3 May 23 22:25:07 4 May 23 22:25:09 o_O May 23 22:34:48 one eye shouldn't be bigger than the other. you may want to get that checked May 23 23:06:51 mranostay lol May 23 23:07:21 The logging portion of my beaglebone powered home alarm system has been running perfectly for almost a weke. May 23 23:07:23 week* May 23 23:09:54 * mranostay robs KeatonT house for beta testing May 23 23:10:00 lol May 23 23:10:54 It's an ongoing process. May 23 23:24:50 is anybody at all familiar with the battery function of the beagle bone? May 23 23:27:18 i'm refering to the battery charge stuff on pmic expansion May 23 23:28:15 no but I'm interested in knowing. May 23 23:28:19 So if you find out. May 23 23:28:20 Let me know. May 23 23:30:42 ok yeah i haven't see much doc on it May 23 23:32:02 it is a stock battery charger May 23 23:32:06 nothing unusual about it May 23 23:32:17 you need to add analog glue May 23 23:42:43 ds2: what do you mean analog glue? Are you just talking about wiring it up? May 23 23:55:29 I mean exactly that May 23 23:55:50 do the right components to make it match your battery May 24 00:31:53 welcome prpplague May 24 00:33:10 greetings May 24 00:34:05 indeed. May 24 01:28:17 Well, rev 2 of my beaglebone stepper shield is out for fabrication May 24 01:28:23 Fingers crossed it meets expectations :) May 24 01:34:02 Hello everyone. I am currently working with a BeagleBone, and am trying to load matplotlib onto it. The command I am using is "python setup.py install", and it is returning an error "ImportError: No module named distutils.core". Has anyone every successfully added matplotlib? May 24 01:34:20 jtr001-UCSD: I can fix that May 24 01:34:37 You're missing a dependency... One of the python-* packages. I forget what they named it. Give me one sec May 24 01:35:07 I did some looking around and some sites told me I am missing setuptools May 24 01:35:08 Oh duh: python-distutils May 24 01:35:19 opkg install python-distutils and you should be good May 24 01:35:27 Ok thanks, I will try that May 24 01:36:05 BTW: If you're going to use matplotlib with Python you should try it out inside Gate One... You can print your graphs to stdout and they'll display in the terminal :) May 24 01:36:26 I just sent Koen a new gateone_git.bb that fixes all the speed problems May 24 01:37:35 Thanks...that worked for me (the matplotlib) May 24 01:37:43 I am having a hell of a time doing "bitbake cloud9-image"... Keep getting checksum errors May 24 01:37:51 I am not sure what you mean by "inside Gate One" May 24 01:38:08 I was able to fix the netbase package myself but now I'm getting checksum errors on debianutils and the version difference is too great for a simple fix May 24 01:38:08 Sorry I am very new to this, just trying to work through it =X May 24 01:38:29 Gate One is the web-based terminal emulator that ships with the Beaglebone May 24 01:38:35 Though the version that ships is DOG slow May 24 01:38:46 I just got myself a Beaglebone and fixed it May 24 01:38:59 I've logged onto Cloud9, how do i access Gate One? May 24 01:38:59 I even made a pretty little logo that will be displayed by default now :) May 24 01:39:05 Gate One should be listening on https May 24 01:39:10 So same URL, just https May 24 01:39:27 Though on my 'bone it didn't start up by default for me May 24 01:39:42 I had to login to the device and start it up by hand... /var/lib/gateone/gateone.py May 24 01:39:50 Not sure why the avahi service didn't start it up May 24 01:40:00 I don't fully understand how Avahi works May 24 01:40:27 I understand zeroconf/rendezvous but does it start the service on-demand or something? If that's how it is supposed to work, well, it isn't working May 24 01:41:01 Sorry, I wish I could help, but that is way over my head May 24 01:41:36 Yeah, I'm usually pretty good at this stuff but I think my head has been in too much Python code lately :) May 24 01:41:53 Haha, I think I understand that feeling May 24 01:42:04 This is my first encounter with a microcontroller May 24 01:42:05 BTW: You can learn more about Gate One here (and try the demo: Just press the ESC key!): http://liftoffsoftware.com/ May 24 01:42:15 This ain't no microcontroller May 24 01:42:24 Beaglebone is WAAAY better than a microcontroller :) May 24 01:42:30 Much more powerful May 24 01:42:32 yeah May 24 01:42:36 Like, 1000x more power May 24 01:43:06 I was thinking about adding USB and ethernet to my Arduino a while back and it would've cost me like 2x the cost of a Beaglebone. It was silly May 24 01:43:17 Not to mention how slow it would've been May 24 01:44:07 ATMega chips were just not meant for TCP/IP. Though they're great as USB perhipherals (oddly not with the default chip that comes with an Arduino though, go figure!) May 24 01:44:34 riskable: There's V-USB, if you don't want to do high-speed USB May 24 01:44:55 And yes, the price of ethernet shields etc is remarkable. Even more so is the cost of wifi shields May 24 01:45:01 I ordered a little USB wifi adapter for my Beaglebone... Why? I was purchasing something entirely different and Amazon suggested it to me. For $7 shipped how could I say no? hehe May 24 01:45:21 * nickjohnson paid way too much for one in a local shop, then saw Adafruit's $20 ones :/ May 24 01:45:25 That alone just shows how much better the 'bone is May 24 01:45:38 * riskable got the same exact chip as the Adafruit one for $7 May 24 01:45:48 Well, it's more that the various integrated components that the Arduino has to use are overpriced May 24 01:45:56 ok I have two A5 beaglebones next to each other, one is stable in ubuntu-arm and the other is flaky May 24 01:46:00 thoughts? May 24 01:46:03 nickjohnson: Really? I wouldn't know May 24 01:46:10 flaky bone is stable in angstrom May 24 01:46:13 They integrate things like the TCP stack on the module, and there's a smaller market than for consumer wifi, which has a huge competitive market May 24 01:46:18 trelane: How are they powered? May 24 01:46:29 nickjohnson: I suppose that makes sense May 24 01:46:30 riskable, 5v PSU May 24 01:46:42 tried switching PSU no change May 24 01:46:43 trelane: Is one flaky on the same PSU? May 24 01:46:46 yeah May 24 01:46:46 Ahh May 24 01:46:53 Same OS image both of them? May 24 01:46:55 flaky one is also flaky on another PSU May 24 01:46:56 yes May 24 01:47:03 Define "flaky" May 24 01:47:16 riskable, random hard freezes including the watchdog (heartbeat led stops) May 24 01:47:32 Bad board May 24 01:47:37 I mean, what else could it be? May 24 01:47:42 Sunspots? May 24 01:47:50 Maybe a bad solder somewhere May 24 01:47:51 I think it's a bad board May 24 01:47:52 Could be a flaky SD card May 24 01:47:54 Check the caps May 24 01:48:00 Though actually, if the watchdog freezes, no May 24 01:48:01 nickjohnson: Oooh, good suggestion May 24 01:48:08 riskable, thing runs GCC in angstrom May 24 01:48:12 barely boots in Ubuntu May 24 01:48:16 (Well, probably no. Weird things happen with complex systems. :)) May 24 01:48:23 nickjohnson: Actually, if the SD card has problems it can cause issues with the watchdog because the SD card driver lives in kernel space May 24 01:48:37 riskable: Oh, is the watchdog a kernel driver, rather than a hardware watchdog? May 24 01:48:48 I've seen MANY MANY times where a bad drive can cause even the strangest of nonsensical misbehavior May 24 01:48:52 trelane: Tried swapping their SD cards around? May 24 01:48:56 nickjohnson, yeah May 24 01:49:13 I had a Linux host with a bad controller once that would constantly report 999 degrees on the CPU thermistor. Replaced the controller and suddenly the temperature was showing up properly again May 24 01:49:20 I take it it didn't solve your problem? May 24 01:49:34 riskable: I think that falls under my "weird things happen with complex systems" heading ;) May 24 01:49:49 nickjohnson, no May 24 01:49:58 and this card functions correctly in the other bone as well May 24 01:50:05 Bad board May 24 01:50:12 the problem is it's stable in angstrom :/ May 24 01:50:36 Oh, then it is likely the chip itself May 24 01:50:45 Not all CPUs are made equal May 24 01:50:47 right May 24 01:50:49 that's my thought May 24 01:50:56 Especially $5 CPUs that are stamped out by the millions May 24 01:51:16 Is there a difference if you power via USB instead of the 5V? May 24 01:51:24 ...because on USB power it will be underclocked slightly May 24 01:51:24 You could try the angstrom kernel in Debian May 24 01:51:33 Though I can't imagine userspace could be doing much to cause the flakiness May 24 01:51:49 It could be a flaky i2c thing May 24 01:51:54 I really odn't want to spend that much time on it :/ May 24 01:52:09 I think it's hardware, problem is proving it when it doesn't watn to fail in angstrom May 24 01:53:08 This may sound silly but "try rebooting". Get it booted up under your ubuntu image then issue a 'reboot' and see if that fixes it May 24 01:53:34 It might be something that happens inside of init on that initial boot but subsequent ones will be fine May 24 01:53:43 Certain init parameters can do funky things like that May 24 01:53:53 Not sure what ubuntu uses by default on the beaglebone though May 24 01:54:31 Arg, how the heck do I get cloud9-image to compile? It feels like it is way out of date May 24 01:54:38 ./oebb.sh update doesn't do squat May 24 01:56:33 eliminated the console chip as the problem May 24 01:57:12 riskable, shutdown -r now hard locks the board May 24 01:57:23 Aha! May 24 01:57:35 Have you got anything plugged into the USB? May 24 01:57:47 the console usb yes, the other usb no May 24 01:57:52 OK May 24 01:57:55 this is a bare board sitting on a plastic bench May 24 01:58:23 It might be the DDR2 RAM has some bad spot that Angstrom doesn't use but Ubuntu does May 24 01:58:40 ...but hard hang on shutdown says "bad board" May 24 01:59:02 hang is instant May 24 01:59:08 shutdown -r now May 24 01:59:11 *LOCK* May 24 01:59:26 ubuntu@omap:~$ sudo shutdown -r now May 24 01:59:26 [sudo] password for ubuntu: May 24 01:59:26 Broadcast message from ubuntu@omap May 24 01:59:26 (/dev/ttyO0) at 18:25 ... May 24 01:59:26 The system is going down for reboot NOW! May 24 01:59:27 ubuntu@omap:~$ May 24 01:59:31 that's from the console May 24 02:00:07 is there a way to do a memtest? May 24 02:00:18 I am trying to find a way to get this to generate an error May 24 02:00:49 Good question about the memtest May 24 02:00:56 I'm surprised there isn't a memtest image May 24 02:01:01 Maybe I should make one May 24 02:01:03 * trelane needs to prove to $VENDOR that this is broken so I can get an RMA May 24 02:01:09 Though memtest might be an x86-only program May 24 02:01:14 it is May 24 02:01:44 * trelane tries to think of a way to force high memory usage on the angstrom unit with lots of reads and writes so it'll fail May 24 02:07:27 trelane: How about this? python -c 'a = []; for i in xrange(1000000000): a.append(i)' May 24 02:07:45 Just increase the size of the range. Make it grow and grow May 24 02:07:52 You could even have it print where it's at May 24 02:07:53 right works if the memory isn't already mapped May 24 02:08:14 Well, it won't tell you WHERE in memory there's a problem but you can see how much you're using before it breaks May 24 02:08:15 * trelane is assuming that angstrom is writing to that ram and never reading it May 24 02:08:30 Well, Python needs to read it every time you append() May 24 02:08:45 ...since it is a list which is mutable May 24 02:09:04 If you want to *just* write use a dict May 24 02:09:51 If you want to write then read use a dict... python -c 'a = {}; for i in xrange(100000000): a[i] = i+1' May 24 02:10:04 Then each time you write it will definitely also read :) May 24 02:10:59 Anyway, it is time for me to hit the sack May 24 02:11:04 enjoy May 24 02:11:09 thanks for the help May 24 02:11:12 I'm going to blow away my setup-scripts dir and try again from scratch May 24 02:11:19 (overnight since it takes forever =) May 24 02:17:43 hey thanks for all the help riskable May 24 02:17:55 No prob... Did it work out OK? May 24 02:18:07 yeah we can use gate one now May 24 02:18:22 its super laggy its kinda funny May 24 02:18:33 Yeah, it is slow May 24 02:18:41 An update will be out soon that fixes it May 24 02:19:08 you must be a wizard for fixing it yourself May 24 02:19:08 You can grab my pull request by hand here if you want: https://github.com/koenkooi/meta-oe/pull/1 May 24 02:19:19 Yeah, I'm sort of an expert May 24 02:19:30 * riskable is the author of Gate One May 24 02:19:42 http://farm5.staticflickr.com/4119/4876696414_d3f62d18be_z.jpg May 24 02:19:44 hehehe May 24 02:20:23 Gate One 1.1 should be out soon regardless... Just finishing up a few minor features and writing documentation May 24 02:20:48 Keep getting distracted with big "potential" clients that want weird stuff May 24 02:20:55 Well, not that weird, actually May 24 02:21:11 All good features to add... But suddenly having to prioritize things tends to put off new releases :) May 24 02:21:20 Well, with that I'm off to bed May 24 02:21:21 Have fun May 24 02:21:35 laters May 24 02:55:40 Is anyone familiar with cloud9? May 24 02:57:06 Running my html/js code I get a standard error from within one of the node.js modules May 24 02:57:41 ^ node.js:134 throw e; // process.nextTick error, or 'error' event on first tick May 24 02:58:25 Where are the node.js libraries stored on cloud 9? **** ENDING LOGGING AT Thu May 24 02:59:58 2012