=== apachelogger_ is now known as apachelogger === asac_ is now known as asac === Martyn is now known as Guest78883 [05:46] hi, does DEB_BUILD_ARCH = armel for the native builders? [08:02] morning [09:59] micahg: Yes; but it's usually a bad idea to check this [11:09] is there a link for geexbox image for arm platform? [11:53] looking at the u-boot packaging i now understand why lool didnt want to do the package split [11:53] sigh [11:56] no compat file, no build deps, debian/rules looks horrid ... [11:57] why cant people just use a proper build system instead of making up debian/rules by hand [12:04] ogra: I actually committed an u-boot-tools package in the Debian git repo this morning [12:04] oh, cool [12:05] this rules file is just painful [12:09] lool, any ETA when that will be in debian (i need it asap so i'm pondering to add it as patch for ubuntu) === loafh is now known as loafhead [13:40] GrueMaster, can you remember what failed in your testing of omap from natty master [13:53] lool: what's the proper way to check if something is an arm build vs not arm? [13:57] micahg: 'arm build' as 'build on arm' or as 'build for arm'? [13:57] hrw: build on arm [13:58] 'uname -m' may be but debian packagers will want to kill you for it probably [13:59] right, I was looking at DEB_BUILD_ARCH, but apparently, that's not the best way? [14:03] micahg: why you need check? [14:04] hrw: --enable-thumb2 fails the firefox build on non-armel builds [14:04] ah [14:05] rules2: ifneq (,$(findstring $(DEB_TARGET_ARCH),armel hppa ia64 sparc)) [14:05] micahg: so use DEB_TARGET_ARCH check [14:06] DEB_BUILD_ARCH = amd64 DEB_TARGET_ARCH = armel is 100% proper btw [14:06] * micahg is getting DEB_TARGET_ARCH not supported on maverick [14:07] let me check how gcc do that [14:07] micahg: It's uncommon to check the architecture of the buildd [14:07] micahg: Usually you test what you're building /for/ [14:08] DEB_HOST_ARCH is the Debian architecture of the machine the package will run on [14:08] rules.defs: TARGET_VARS := $(shell dpkg-architecture -f -a$(DEB_TARGET_ARCH) 2>/dev/null) [14:08] rules.defs:DEB_TARGET_ARCH ?= $(call vafilt,$(TARGET_VARS),DEB_HOST_ARCH) [14:08] ops [14:08] micahg: It would also be best to test the GNU CPU or something similar rather than the Debian architecture name; for instance the Debian armhf port is very similar to armel, but wont be covered by your test [14:09] sorry - it is done by different way in gcc [14:10] micahg: Perhaps you could explain your specific problem? I'm not sure I can comment on which var you should use, as you only asked about native builds but I might have missed when you explained what you're building [14:10] hrw: The packages like gcc which have the concept of a target architecture are fairly rare [14:10] lool: well, I need to add a build option to arm only for firefox [14:11] micahg: Right, that's a good example of a case where you'd test DEB_HOST_ARCH, and *not* DEB_BUILD_ARCH [14:11] micahg: The main difference is when cross-compiling; when cross-compiling, DEB_BUILD_ARCH might be amd64 while DEB_HOST_ARCH is still armel [14:12] ah, so HOST_ARCH is the right one, ok, that's good to know for the future [14:18] thanks lool [14:27] micahg, I think it's a simpler change to not check at all and let ff configure drop the option if it is not suitable for the current arch [14:45] janimo: I tried with it on for amd64 and the build failed [14:47] micahg, did it pass march arm to gcc? [14:47] micahg: Just saw your firefox changes; would you mind if we discussed these for a sec? [14:48] lool: actually, I'm about to catch a bug, can we chat in a few hours? [14:48] *bus [14:48] micahg: Sure; I will send you an email [14:48] janimo: I have to check the build log [14:48] lool, you mean the change in the latest upload or something else? [14:48] lool: thanks [14:48] janimo: Yes, the THUMB2 one [14:49] lool, I suggested that one to micahg to fix the FTBFS. Tested on a panda before that. [14:49] the details may need to be done differently but I think that flag needs to be passed [14:50] janimo: Yup; do you want to be Cc:ed? [14:50] lool, sure [14:50] Hmm apt-cache showsrc janimo wont give me your email address [14:50] jani at ubuntu.com [14:51] * lool <- tired [14:51] janimo: thanks [14:59] lool, is there any ETA for the debian u-boot upload or do we need to add the package split as ubuntu patch to get it in this week ? [14:59] (i think you disconnected above when i asked before) [15:08] ogra: I've contacted Clint this morning to ask him to review these changes; they will hit Debian NEW in any case, so that will probably take some time; however, it wouldn't be too bad to copy the NEW package in Ubuntu [15:09] lool, do you get any notification you could forward if that happens ? [15:12] (i mean if it hits NEW, I'll care for teh sync requests etc) [15:54] ogra: Unfortunately, one can't sync from NEW; if I'm doing the upload, I'll tell you once it's done [15:54] thanks [15:58] Morning (barely). [16:00] ogra: rsalveti: apw: The issues I saw with the previous omap kernel prior to the break was that it literally didn't boot. I couldn't get any feedback past u-boot loading the kernel. Not even output from earlyprintk. [16:02] I'm installing the new kernel now. [16:10] GrueMaster: yeah, I'll also try it here, and then try to fix it if it doesn't work [16:11] Hmm. Currently, it appears to be stuck on installing the linux-headers package. [16:13] installing things in general is slow i think if you have the older dpkg [16:13] doesn't seems to boot, will now debug to see why [16:13] Looks like I was getting a segfault loop of some sort. Rebooting. [16:14] rsalveti, thanks [16:14] rsalveti, are u gettting any console output at all ? [16:15] apw: nops [16:15] rsalveti, that sound fun to debug [16:16] apw: :-) [16:16] sounds like an issue with the console driver [16:17] do you have a monitor attached or do you try purely with serial ? [16:17] hm, ok, got something with earlyprintk [16:26] folks, the weather report for Dallas is cold next week Daytime 8.3C and nighttime -2.77 Colder then usual around here [16:26] For some of you that is warmer then you are at but for us it's cold [16:28] ouch, very cold for me hehe [16:28] it's around 30C everyday around here [16:28] good to know [16:29] I vote we relocate to Sao Paulo. [16:29] hehe :-) [16:29] at least i wont catch a cold [16:30] we have constantly around -3°C here (day and night) [16:34] ogra: apw: GrueMaster: ok, did boot, you now need to use ttyO2 instead of ttyS2, like panda [16:34] because of the newer console driver [16:35] awesome [16:35] but I'm now getting tons of mmc i/o errors [16:35] probably a different issue, more debugging [16:35] I had tried that when I tried the older kernel. [16:35] rsalveti, sounds good! [16:35] apw: will test it more around here, and compare the config with linaro's one [16:35] GrueMaster, the kernel he is testing is two -rc's newer than the one you tested before xmas [16:36] I know. [16:36] Just saying that I had tried both serial port settings. [16:38] GrueMaster, good to know, as that means things have improved by osmosis === njpatel is now known as njpatel|away [16:41] apw: ogra: GrueMaster: full maverick desktop up and running [16:41] rebooted and now I didn't get any i/o error [16:42] * rsalveti trying again [16:42] wohoo [16:42] apw, so can we get omap3 back ? [16:42] no calltrace during the bootlog [16:42] I believe this could be a good start [16:42] then we can just improve over the time, with more testing and etc [16:43] yeah [16:43] at least I believe it's good enough to get images [16:45] rsalveti, did you say that you have booted this one successfully ? [16:45] yes he did :) [16:45] apw: yup, 2 times [16:45] GrueMaster, does it boot for you too ? [16:45] who cares about GrueMaster :P [16:45] let me know try with a normal beagle [16:45] tried with a beagle-xM [16:45] *now [16:45] I'm having other non-kernel related issues atm. Will try again shortly. [16:45] he always manages to break even the super solid things ;) [16:46] It's my job, that's what I do. [16:46] I was hell with my toys. [16:46] lol [16:47] haha [16:48] ogra, ok tehcnically i can get this renabled yes, so you are going to use this kernel not the linaro in your images, thats the decision [16:49] and your team is going to sign up to keeping it tested and working [16:49] yes [16:49] and yes [16:49] :-) [16:50] bring it, big guy. === njpatel|away is now known as njpatel [16:52] rsalveti: Did your patch for the serial port error on XM ever get integrated into the maverick kernel? [16:52] i think i saw some noise about it this week [16:52] GrueMaster: it's commited already I believe [16:52] let me check again [16:54] yup, mail says "applied" [16:55] ok, haven't seen it in proposed yet. [16:55] not uploaded [16:55] yeah -proposed is somewhat behind Fix Committed over this xmas perios [16:55] not uploaded, it's commited at the master-next [16:55] period [16:56] ogra: apw: GrueMaster: ok, also working fine at my beagle c4 [16:56] so it seems to be good enough [16:56] yay [16:57] ogra, so we are saying beagle is the target for the omap kernel [16:57] apw, XM [16:57] apw: yep, xM will be our main use case [16:57] as it's now quite popular [16:57] and faster than a normal beagle [16:57] so the acceptance criteria for that omap kernel is 'works on Beagle XM' [16:58] ++ [17:00] yup [17:04] MADISON, Wis. - Madison police are looking for a burglar who gave back a bottle of tequila after the homeowner hit him in the head. [17:04] oops [17:04] wrong window ;) [17:27] apw: the config seems fine comparing with the older one from maverick and with linaro's one [17:28] rsalveti, thanks for the confirmation [20:52] My 8G Class 10 SD cards finally came in. Woohoo! [20:53] luck bastard [20:58] 8g class= nice, new class? :D [20:59] class 10 isnt that new [20:59] rsalveti, do you need any? i can buy them at the discounter and bring you some to the rally [21:00] ogra: would be good to have, I have none around [21:00] ogra: how much it cost usually around there? [21:00] i'll have to check, i'll just bring a few, if you dont want them (or they are to expensive) i'll take them back again [21:01] ogra: cool, would like at least 2 :-) [21:01] ok [21:01] np, just to know, here the problem is that I can't find them easily [21:01] or without spending too much money [21:01] I bought 2 8G for just under $30. Last cycle, they were $45 each. [21:02] nice, good price [21:02] amazon.com