**** BEGIN LOGGING AT Fri Mar 30 02:59:58 2012 Mar 30 06:50:25 is there any equivilent of packages.ubuntu.com for ubuntu armel and armhf? Mar 30 09:25:10 GrueMaster, feel free to remove https://code.launchpad.net/~gruemaster/flash-kernel/bug-961174/+merge/98874 if the corresponding bug is wontfix Mar 30 09:25:37 it showed up in the sponsoring queue and I almost considered uploading it before seeing in tiny green letters that the bug is wontfi Mar 30 09:25:46 LP UI is not the best Mar 30 09:25:56 tiny gray letters that is Mar 30 10:04:16 plugwash, you man a website to lookup packages, search for/in packages, etc? no Mar 30 10:04:20 *mean Mar 30 14:03:46 What is the minimum memory that will support ubuntu arm version? Mar 30 14:04:37 traisen: it depend if you want to run the full desktop Mar 30 14:05:17 traisen: a regular unity-2d session work ok using 512mb of ram Mar 30 14:05:28 the system then consume around 200mb on startup Mar 30 14:05:44 you can of course use 256mb of ram + some swap Mar 30 14:05:51 that will allow you to run most applications Mar 30 14:06:05 if you run headless like ubuntu-core Mar 30 14:06:14 then you should be able to make it with 64mb of ra Mar 30 14:06:16 m Mar 30 14:08:50 Thanks, Do you know if anyone has tried it on the Lego Mindstorms NXT ? 256kbyte FLASH- und 64kbyte RAM-Speicher Mar 30 14:10:13 do you know which arm cpu it are using? Mar 30 14:10:24 Atmel Mar 30 14:10:37 ok, do you know the variant? Mar 30 14:10:38 (or do you need more info) Mar 30 14:10:42 i need more info Mar 30 14:10:46 atmel make amny cpus Mar 30 14:10:49 many Mar 30 14:11:08 off hand i can tell you ubuntu will not boot on this device Mar 30 14:11:20 Just a minute Mar 30 14:11:20 but you can most likely write code for it using the ubuntu arm gcc Mar 30 14:11:54 if you have a way to upload code to this mindstorm device Mar 30 14:12:28 Currently using gnu tool chain, but ubuntu is better known then eCos Mar 30 14:13:21 But wouldn't the o/s have to be a favor of ubuntu, even if a minimal version? Mar 30 14:15:16 AT91SAM7S256 Mar 30 14:15:23 traisen, ubuntu hardly fits on a device with 1000 times as much memory Mar 30 14:15:49 traisen: FOR THIS KIND OF DEVICE I RECOMMEND OPENEMBEDDED OR SCRATCHBOX Mar 30 14:15:53 maybe the bootloader used on typical ubuntu arm hw would fit on the mindstorm :) Mar 30 14:16:03 OOPS Mar 30 14:16:07 damn capslock Mar 30 14:16:28 No desktop needed Mar 30 14:16:45 traisen: the problem are that a regular liux kernel require more than 256kb of flash Mar 30 14:16:45 traisen, ecos should be pretty good for it for the type of apps you'd probably run on it Mar 30 14:17:36 traisen: for a minimal embedded linux system you need at least 4mb of flash Mar 30 14:17:45 to store the kernel and a minimal root filesystem Mar 30 14:17:55 and more ram Mar 30 14:18:15 you can compare this lego device with your usb mousepointer Mar 30 14:18:37 the usb mouse probably have more flash Mar 30 14:18:46 if it are flash based Mar 30 14:23:31 MB not Kb :) eCos is working... thanks for mentioning other options, so I have something to compare against. Mar 30 15:35:47 hackish patch to get mongodb run its smoketest completed, now to clean up maybe and backport to Ubuntu's stable version hopefully before 12.04 Mar 30 15:36:45 * janimo` resists being sidetracked into debugging mongo for ppc on the porter machine. Mar 30 16:08:34 janimo`: Go for it, get sidetracked. Mar 30 16:08:48 janimo`: I thought mongo was heavily x86-centric. Just how large is your patch? Mar 30 16:10:07 (Or maybe I'm confusing it with another package that had a "don't bother trying to port this" note...) Mar 30 16:10:59 Ahh, there it is. Mar 30 16:11:01 The mongodb server depends on both little-endianness and unaligned memory Mar 30 16:11:01 access, which I believe means it can only work on i386 and amd64. We believe Mar 30 16:11:04 that the mongodb will be useful even it is not available for all Debian Mar 30 16:11:06 supported platforms. Mar 30 16:11:08 janimo`: Is the above no longer true? Mar 30 16:48:32 Howdy folks. Mar 30 16:48:43 We've got a test that's using callgrind, but it's failing to link on ARM. Mar 30 16:48:56 Could someone take a look? It seems like callgrind isn't built right on ARM. Mar 30 16:49:03 https://launchpadlibrarian.net/98928961/buildlog_ubuntu-precise-armhf.libdbusmenu_0.5.96-0ubuntu1_FAILEDTOBUILD.txt.gz Mar 30 16:50:51 tedg: You didn't get the memo that ARM is everyone's problem/responsibility now? ;) Mar 30 16:51:26 infinity, yes, that's why I'm helping -- by finding bugs :-) Mar 30 16:51:41 *smirk* Mar 30 16:51:48 janimo`: Know anything about valgrind on ARM? Mar 30 17:54:47 plugwash, you man a website to lookup packages, search for/in packages, etc? no <-- yes packages.ubuntu.com lets me search for/in packages on the official ubuntu architectures I was wondering if there was an equivilent for ubuntu armel and armhf Mar 30 17:55:34 plugwash, answer is no Mar 30 17:55:46 ok Mar 30 17:56:04 the easiest way is a chroot and apt-file/apt-cache Mar 30 18:29:01 plugwash: packages.u.c is maintained by frank@lichtenheld.de, feel free to bug him about adding ports. ;) Mar 30 18:31:49 plugwash: Of course, there's also Contents-$arch on the mirrors. Mar 30 18:52:52 infinity, mongodb is indeed x86 centric but not by design, they just did not care for portability much Mar 30 18:53:19 and there are not many places where it needs changing but my chjanges are not the most elegant Mar 30 18:53:55 so unaligned double and atomic int accesses cause traps and it is mostlt about finding and fixing those places Mar 30 18:54:26 infinity, I had a fire-and-forget ARM patch sent to valgrind once but not much knowledge more besides that Mar 30 19:01:02 infinity, as for the size of the mongo patch it is mongoarm.diff here, most recent comments Mar 30 19:01:25 that is against master which built and ran fine, applying the same to the 2.0 branch still crashes so I have some more digging to do Mar 30 19:01:58 infinity, regarding ppc, I probably need to debug it there too, after the single patch that is needed on ARM too (make char signed explicitely in one place) Mar 30 19:02:22 it still crashed and the build for some reason did not have debug symbols so gdb did not help me Mar 30 19:02:42 which is strange as the same default build on x86 and arm has debug symbols Mar 30 19:03:11 anyway I probably need to look at it some more, probably my last chance to play with ppc if I am reading my crystall ball correctly Mar 30 19:03:27 also my first time working on a ppc :) Mar 30 19:05:27 Heh. Mar 30 19:05:35 If you need access to a PPC machine to play, let me know. Mar 30 19:05:45 Although, I guess you have access to the porter in the DC too. Mar 30 19:06:01 But yeah, the claims that it's endian-specific don't bode well for PPC. Mar 30 19:06:09 Maybe that was overstated and/or incorrect. Mar 30 19:06:52 If you can make it both endian and alignment clean, I'm sure the Debian maintainer would love you forever. Mar 30 19:20:06 infinity, yes, it is the porter that I tried building mongo today and which crashed even after one fix which was also needed for arm Mar 30 19:20:33 infinity, ah indeed, it is probably endian specific too, but that luckily did not affect arm Mar 30 19:21:00 there was a mongo-nonx86 branch against 1.8 last year but it is very messy and the author no longer maintains it Mar 30 19:21:09 it had fixes for all sorts including arm Mar 30 19:22:08 infinity, I already had a few exchanges with antonin, pushin all but our upstart script to him to lessen delta Mar 30 19:22:30 and was surprised to find out that debian packages (or the system) can break if upstart scripts are installed Mar 30 19:22:42 because of some yet undeployed dpkg fix Mar 30 19:29:51 ah so that's wre my diskspace went Mar 30 19:30:00 * janimo` discovers a 21Gb log from Gnash Mar 30 19:31:07 wow Mar 30 19:39:24 highvoltage, possibly the effects of always open browser+gnash installed+infrequent reboots. Anyway, removed Mar 30 19:41:21 I use the #rmrf tag on identica/twitter for things like that. like http://identi.ca/notice/50474501 and http://identi.ca/notice/71778992 :) Mar 30 20:58:20 Hello people! I'm looking at a kernel config (versatile) for a recent kernel, I cannot seem to make em work (a part from the 2.6.22 which doesn't support ext4), any help is appreciated Mar 30 21:25:45 Mephisto__: Why versatile? It's long obsolete. Mar 30 21:25:51 Mephisto__: Surely, you want vexpress? Mar 30 21:27:28 well i'm a total newbie to this here's the system I'm trying to "emulate" Mar 30 21:27:42 http://www.sealevel.com/store/relio-r9-arm9-embedded-risc-computer.html Mar 30 21:29:42 Oh. Then you're almost certainly in the wrong channel, as Ubuntu won't run on that system (nor an emulated version of it). Mar 30 21:30:43 ARM9 is ARMv5, Ubuntu only supports ARMv7. Mar 30 21:31:06 Actually, that might even be v4t. Mar 30 21:31:22 Anyhow, you might want to try Debian, and they do have versatile kernels packaged as well. Mar 30 21:31:36 so it's versatile right? Mar 30 21:33:02 Erm. That system isn't the same thing as a versatile board from ARM, no. Mar 30 21:33:23 But they're perhaps vaguely similar product families. Mar 30 21:33:45 In either case, Ubuntu won't run on either one (old ARM versatile boards, the qemu versatile target, nor this machine you linked). Mar 30 21:34:09 I see Mar 30 21:34:17 thanks a lot for valuable informations and your time sir Mar 30 21:34:26 No problem. **** ENDING LOGGING AT Sat Mar 31 02:59:58 2012