**** BEGIN LOGGING AT Sun Mar 26 10:59:56 2006 Mar 26 11:11:57 good day Mar 26 11:12:42 hi all Mar 26 11:12:48 hi cockroach Mar 26 11:13:18 CVS kernel gives me this: Mar 26 11:13:21 42949554.780000] ixp400_eth: ethernet 0 using NPEB and the PHY at address 0 Mar 26 11:13:29 [42949554.800000] Division by zero in kernel. Mar 26 11:13:52 * likewise inspects the source Mar 26 12:23:29 heh Mar 26 12:23:33 RMK is funny Mar 26 12:23:35 Just say no to Python (tm) Mar 26 12:28:49 morning Mar 26 13:12:56 Ifaistos, I'm also working on ppc kernel. maybe we can exchange notes? Mar 26 13:14:23 HopsNBarley : good :) Mar 26 13:15:04 Ifaistos, I've had no problems building 2.6.11, i'm just starting to experiment with later revs. Mar 26 13:15:52 HopsNBarley : i got 2.6.12 and 2.6.15.4 to build but there are issues when compiling modules Mar 26 13:16:13 Ifaistos, I'll take a look. I saw the patch you mentioned. Mar 26 13:16:13 HopsNBarley : external modules like hostap and madwifi Mar 26 13:16:41 Ifaistos, I hit this in 2.6.16: error: `SERIAL_PORT_DFNS' undeclared here Mar 26 13:16:53 Ifaistos, this in arch/ppc/boot/common/ns16550.c:20 Mar 26 13:17:21 Ifaistos, I'm pretty sure this is from not having the include files right. i.e. the specific arch i'm using isn't set up correctly. Mar 26 13:17:34 HopsNBarley : for which ppc you are compiling ? i am doing 405gp Mar 26 13:31:41 I'll create a "mplayer-common" package which contains working configuration for mplayer and mplayer-atty Mar 26 13:31:54 if that's ok.. Mar 26 13:32:51 CoreDump|home: with -ac mad -framedrop ?p Mar 26 13:33:13 and correct rotation for SL-Cxxxx Mar 26 13:35:48 correct rotation? Mar 26 13:35:58 what kind of funky Xserver config do you have?s Mar 26 13:36:16 X? i was running it from the VT Mar 26 13:41:45 mplayer in X will mess up the whole desktop Mar 26 13:45:01 ? Mar 26 13:45:06 mplayer -vo x11 -fs -zoom ? Mar 26 13:49:40 * CoreDump|home whistles innocently while removing mplayer-atty Mar 26 13:50:40 mplayer doesn't compile with -vo x11 Mar 26 13:50:46 err Mar 26 13:50:58 it is not compiled with -vo x11 Mar 26 13:58:21 koen: ok, rotation in X with -vo sdl is correct Mar 26 14:11:14 you might want the w100 driver with mplayer, unless X has gained accel Mar 26 14:16:07 hmmm Mar 26 14:16:13 w100 Mar 26 14:17:01 mithro: is there any documention about w100? Mar 26 14:21:55 some Mar 26 14:22:07 but things would have change quite a bit since I last looked at it Mar 26 14:23:28 There is very little documentation about the w100 Mar 26 14:31:38 hows the opensource aticore stuff going? Mar 26 14:36:32 mithro: I haven't had time but I know some other people working on it and its nearly complete Mar 26 14:56:41 ok, it looks like mplayer isn't able to set options to a specific -vo (and not use them on other -vo's) Mar 26 14:57:20 the only way to make mplayer work correctly in X _and_ VT would be to use a wrapper for mplayer Mar 26 14:57:58 the most important stuff you want to put in mplayer.conf is ac = mad and framedrop Mar 26 14:58:22 that accound for 95% of the 'bugs' user complain about Mar 26 14:58:25 accounts* Mar 26 14:58:28 koen: i want mplayer to work correctly in X and VT Mar 26 14:58:42 that includes correct rotation Mar 26 15:24:38 03rpurdie 07org.oe.dev * r2e6f4fc5... 10/packages/ (5 files in 3 dirs): Update several DEPENDS to match recent xlibs renaming Mar 26 15:26:02 Good evening all Mar 26 15:29:47 Ping mickeyl Mar 26 15:31:48 hi uv1 Mar 26 15:31:57 hi RP Mar 26 15:32:49 Have problems to build linux-uml. Kernel build complains "PTRACE_SETOPTIONS' undeclared". Any ideas ? Mar 26 16:09:33 03coredump 07org.oe.oz354x * r39783c0e... 10/packages/gpe-bootsplash/ (11 files in 3 dirs): gpe-bootsplash: Update to version from .dev to allow for bootsplash themes Mar 26 16:09:37 03coredump 07org.oe.oz354x * r1d2119c9... 10/packages/mplayer/ (4 files in 2 dirs): Mar 26 16:09:37 mplayer: Add mplayer-common. Mar 26 16:09:37 mplayer-common includes a default configuration for mplayer which should work on all devices. Also added mplayer-common to the mplayer and mplayer-atty .bb's Mar 26 16:09:41 03coredump 07org.oe.dev * r4aa63cca... 10/packages/mplayer/ (4 files in 2 dirs): mplayer: Add mplayer-common, ported from 3.5.4.1 Mar 26 16:14:28 CoreDump|home: ao=oss is slower as ao=also according to Xoroa Mar 26 16:17:53 koen: sadly, ao=alsa doesn't work on SL-Cxxxx right now Mar 26 16:18:05 ah, I see Mar 26 16:20:28 where i can find info apout tslib ? there is a specific irc channel ??? Mar 26 16:23:59 CoreDump|home: not even with the latest alpha? Mar 26 16:24:15 gremlin[it]: not that I know off Mar 26 16:24:23 gremlin[it]: kergoth_ is the tslib maintainer Mar 26 16:24:27 alsa sound itself works just fine, but mplayer refuses to work with -ao alsa Mar 26 16:25:05 mplayer is a piece of crap when you look how its being developed Mar 26 16:25:34 thank koen ... Mar 26 16:27:54 after i finish the framebuffer the customer notice teh ts don't work ... grrrr Mar 26 16:28:13 gremlin[it]: ipkg install tslib-tests Mar 26 16:28:50 gremlin[it]: that will give you some utils to debug tslib Mar 26 16:31:58 koen no way i wan't abel to build a functional image with oe ... Mar 26 16:33:03 i know the error ... when i do a read (c library) to the file descriptor of ts device i got 'invalid argument' ... Mar 26 16:33:40 same error if i try to do a `cat /dev/misc/ep93xx_ts` ... Mar 26 16:35:57 but tha strange thing is that ts_print utility work ... at least it print out some numbers ... mistere for me ... Mar 26 16:45:51 CoreDump|home, RP: could one of you merge the multiple heads? Mar 26 16:46:36 which branch? Mar 26 16:47:09 the one with multiple heads :) Mar 26 16:49:06 hello Mar 26 16:49:06 monotone: successful exchange with monotone.vanille.de Mar 26 16:49:09 merged Mar 26 16:51:45 when bitbake goes into do_package stage, what is ${WORKDIR} the current directory? it seems it's not even using my Makefile Mar 26 16:52:19 CoreDump|home: #PREFERRED_PROVIDERS += " virtual/libsdl:libsdl-qpe" Mar 26 16:52:29 comment that out in local.conf and in its place do... Mar 26 16:52:35 PREFERRED_PROVIDERS += " virtual/libsdl:libsdl-x11" Mar 26 16:52:54 i wish that were documented somewhere Mar 26 16:52:55 jnc: oh yeah, I did that already. thanks tho ;) Mar 26 16:53:28 of course, having that in local.conf prevents you from compiling opie SDL stuff Mar 26 16:53:35 oh yeah Mar 26 16:53:39 good thing you mention that Mar 26 16:53:44 CoreDump|home: not really Mar 26 16:53:49 since i just started an opie-image build. heh Mar 26 16:54:00 the shlib stuff should pickup libsdl Mar 26 16:54:08 koen: it does Mar 26 16:54:11 the wrong one.... Mar 26 16:54:19 that's a bug Mar 26 16:54:22 =) Mar 26 16:55:04 you know, can it be so hard to make libsdl selectable through a debian-like alternatives system? Mar 26 16:55:21 CoreDump|home: reopen http://bugs.openembedded.org/show_bug.cgi?id=115 Mar 26 16:55:31 03koen 07org.oe.dev * rfac7719b... 10/packages/upnp/libupnp_1.3.1.bb: libupnp: add 1.3.1 Mar 26 16:57:45 kernel/qmemorymanager_qws.h:129: error: cast from ‘char*’ to ‘int’ loses precision Mar 26 16:57:49 oh fun Mar 26 16:58:13 heh Mar 26 16:58:37 holy crap. I have no idea what this bug is all about lol Mar 26 16:58:56 * CoreDump|home reopens anyways as it is a very annoying bug Mar 26 17:05:38 evening Mar 26 17:07:08 CoreDump|home: it's not too difficult to understand Mar 26 17:07:34 CoreDump|home: libsdl-qte and libsdl-x11 are not binary compatible Mar 26 17:07:43 right Mar 26 17:07:47 that's the end of it Mar 26 17:07:53 heh Mar 26 17:08:03 thats not a bug, it's fact =D Mar 26 17:08:37 exactly. that bug report is mostly a discussion on what is already known Mar 26 17:10:50 would someone on ia32 build host please test this patch? http://bugs.openembedded.org/attachment.cgi?id=496 Mar 26 17:12:02 jnc: libsdl-qte and libsdl-x11 are binary interchangable for apps that link against one of them Mar 26 17:12:35 which is what #115 is about, that OE doesn't know how to handle virtual/ Mar 26 17:12:43 oh? Mar 26 17:12:57 there was mention of a spurious main() function Mar 26 17:12:58 the -DSDLMAIN was fixed eons ago Mar 26 17:13:05 ah =) Mar 26 17:13:10 okay so they are binary compatible huh Mar 26 17:13:41 would this make sense to handle at the package manager level? Mar 26 17:13:49 it already is Mar 26 17:14:02 when #115 was closed it worked as expected Mar 26 17:14:18 oh then why suggest to reopen it Mar 26 17:14:34 the packages depended on libsdl_1.2.x and your package manager will select the appropriate one from the feeds Mar 26 17:14:52 that sounds okay Mar 26 17:15:04 I suggested to reopen it since CoreDump|home was seeing Depends: libsdl-qte Mar 26 17:15:19 actually I wasn't Mar 26 17:15:21 ohh, not depends libsdl_1.2.x Mar 26 17:15:35 i was seeing gpe-apps compiled with QPE sdl Mar 26 17:16:01 ah, I misunderstoodn Mar 26 17:16:10 that's pretty harmless Mar 26 17:16:20 but annyoing =\ Mar 26 17:16:26 provided you only have libsdl-x11 visible to ipkg Mar 26 17:16:52 it's one of the things you have to pay attention to as a distro developer :) Mar 26 17:17:28 heh I hate it when mplayer/gpe doesn't run because it can't find libqpe or something like that Mar 26 17:19:22 koen: how about we dump some of this into the sanity check bb Mar 26 17:19:22 that should only happen if you screw up your feeds Mar 26 17:19:31 and spit out warnings that we as developers can grok ;) Mar 26 17:20:29 anyways... i'm just talking for apparently no reason. back to reviewing amd64 build host patches Mar 26 17:28:27 03mickeyl 07org.oe.dev * r4559dcad... 10/packages/libetpan/ (7 files in 4 dirs): libetpan: add 0.43, remove 0.36 Mar 26 17:31:31 * france is back (gone 11:24:52) Mar 26 17:57:54 * france is away: Away Mar 26 18:07:57 ok Mar 26 18:08:22 i have this newly set up image of familiar running Mar 26 18:08:28 and i am very happy Mar 26 18:08:53 now, i am trying my utmost to make the boot process a bit more verbose, so i can actually see whats happening Mar 26 18:09:42 i have tried setting the VERBOSE property in /etc/defaults/rcS to "yes" and "very", but i am not noticing any effect whatsoever Mar 26 18:09:59 what am i be doing wrong? Mar 26 18:10:35 dunno about fam, but check /etc/init.d/rc* Mar 26 18:13:54 yay! Working hotkeys =) Mar 26 18:15:20 ok CoreDump|home, i agree with you... i did check the init.d/rcS script, which is the very first script that runs... it has a few lines that say that if $VERBOSE = very, echo "Init".... etc etc. set VERBOSE to very, but no effect :( Mar 26 18:15:45 CoreDump|home: have you tried the softkeys on the lcd yet? Mar 26 18:16:27 koen: nope, they do not work Mar 26 18:17:29 [ "$VERBOSE" = very ] && echo "INIT: Running $i..." Mar 26 18:18:43 now, if VERBOSE is set to very in the config file, what could possible reasons be for not seeing it... or is it perhaps echoing this back over the serial line instead of on the screen Mar 26 18:19:42 CoreDump|home: gpe or opie? Mar 26 18:19:48 gpe Mar 26 18:20:02 opies are working fine for ages ;) Mar 26 18:24:32 hmm it looks like matchbox grabs F11 Mar 26 18:25:29 indeed it does Mar 26 18:31:24 mickey|bbl: could you please take a look at bugs #345, #535 and test the patches on ia32 build host? they look like the correct thing to do for amd64 build host here Mar 26 18:32:46 mickey|bbl: i'm quite sure they are okay for amd64 users, please test for ia32 build host :) Mar 26 18:33:30 CoreDump|home: f11 is used for fullscreen Mar 26 18:33:52 right Mar 26 18:33:59 CoreDump|home: the "fullscreen" hardware buttons are usually mapped to F11 key Mar 26 18:34:19 you already knew that probably Mar 26 18:34:21 :/ Mar 26 18:34:30 =) Mar 26 18:35:11 * france is back (gone 00:37:16) Mar 26 18:35:46 hmm i can't use the same hotkey mutiple times, i have to wait 5-10s for it to become usable again Mar 26 18:39:04 * jnc looks confused Mar 26 18:39:11 where the heck are my commits for sqlite? Mar 26 18:39:19 i don't see it in .dev anywhere Mar 26 18:40:39 Ping. Mar 26 18:40:42 pong Mar 26 18:40:55 Cool. Mar 26 18:41:17 oh wait there it is... 2.8.17 sqlite Mar 26 18:41:21 I'm trying to set the date through the command line on Familiar. Does anyone know the currect format to do such a thing? Mar 26 18:41:44 Jason_Foss: busybox docs i think Mar 26 18:43:28 ohhhhh.... snap. i hardcoded this as a config option Mar 26 18:43:30 damn i'm good Mar 26 18:43:49 OPTS="-fPIC -D'INTPTR_TYPE=int'" Mar 26 18:56:43 hi all Mar 26 18:56:57 hey rob_w Mar 26 18:57:02 maybe a no brainer for some of you : branch org.openembedded.oz354fam083 has multiple heads Mar 26 18:57:02 Mar 26 18:57:17 but for me it is like a himalaya currently Mar 26 18:57:19 ;-) Mar 26 18:58:03 monotone merge and ignore it Mar 26 18:58:11 ~lart usb Mar 26 18:58:12 * ibot beats usb into protomatter with the andromeda galaxy Mar 26 18:58:28 rob_w: it means that the many monotone servers are a little out of sync Mar 26 18:58:38 not a big deal in the world of monotone development Mar 26 18:58:41 I got the wif dongle to work (with a powered hub) and no mather how I remove it, it always kills my keyboard Mar 26 18:58:42 jnc: no, that's wrong Mar 26 18:58:57 koen: well sometimes people push changes to different servers Mar 26 18:58:58 jnc: it means that we have multiple heads and someone forgot to merge Mar 26 18:59:10 i know in a perfect world everyone pushes to monotone.vanille.de Mar 26 18:59:18 i seee Mar 26 18:59:21 not much choice Mar 26 18:59:22 03jnc 07org.oe.dev * r028a611c... 10/packages/uicmoc/ (2 files in 2 dirs): Mar 26 18:59:22 Add a couple of 64-bit buildhost cleanup patches to the tree, but do not use Mar 26 18:59:22 them yet. These need to be further tested by devs with ia32 buildhost, and Mar 26 18:59:22 activated by uicmoc maintainer. Thanks to Darren Edmundson for patches (OE bugs Mar 26 18:59:22 #345, #535). Mar 26 18:59:30 thank goodness Mar 26 18:59:38 i was testing to see if it would only commit those two files Mar 26 18:59:46 and it did :) Mar 26 18:59:56 jnc: that's a nonsense commit, please revert it Mar 26 19:00:14 it's not, mickey is going to test this later and enable the patches Mar 26 19:00:14 jnc: either activate the patches, or wait for the maintainer to activate them Mar 26 19:00:26 jnc: do *not* add useless files to OE Mar 26 19:01:12 jnc: this is not how we do it Mar 26 19:01:19 jnc: so please revert that commit Mar 26 19:01:23 well.... i don't want to put the maintainer through all the trouble of going to a bugzilla thing, downloading a patch, coming up with a name. i mean, that's not a waste of time? Mar 26 19:01:31 no Mar 26 19:02:11 if it was a waste of time we wouldn't have bugzilla, would we? Mar 26 19:02:28 so stop littering the metadata with unused files Mar 26 19:02:34 it's not being used very effectively. i'll revert the change, no problem Mar 26 19:02:47 i don't understand your objection Mar 26 19:04:46 ah , now my tree feels updated ,,, Mar 26 19:04:53 koen: it's not like mickey can magically access my home directory on my computer Mar 26 19:05:04 jnc: why would he need to? Mar 26 19:05:09 for the patches Mar 26 19:05:17 03jnc 07org.oe.dev * rb2304690... 10/packages/uicmoc/ (2 files in 2 dirs): disapproval of revision '028a611cef6b234bc056ae3e11b0b983d6762f4e' Mar 26 19:05:18 the patches are in bugzilla Mar 26 19:05:38 okay so how are patches in bugzilla tracked Mar 26 19:05:50 everything and its mother is assigned to oe@handhelds.org Mar 26 19:05:59 when bugs are assigned, people bitch about it Mar 26 19:06:04 that's how we use bugzilla Mar 26 19:06:29 jnc: people bitch because you force stuff onto them Mar 26 19:06:40 force? Mar 26 19:06:50 i'm not holding a gun to your head demanding you fix a bug or you die Mar 26 19:06:55 that's hardly force. Mar 26 19:07:12 you're assigning a bug to people without asking them for consent Mar 26 19:07:20 which is forcing stuff onto people Mar 26 19:07:26 if a bug is interesting to me, and i've commented on it, i would not complain if someone assigned it to me and told me to keep on top of fixing it Mar 26 19:07:47 if someone wants to fix a bug, he assigns it to him/herself Mar 26 19:08:00 exactly Mar 26 19:08:00 okay Mar 26 19:08:24 i'm doing wifi/amd64/gcc stuff.... with this mentality you speak of, someone has to go out of their way to ask me if i want the bug Mar 26 19:08:44 that's called communicating Mar 26 19:08:46 instead of just assigning it to me, in which case it shows up in my list and i can see if i want to work on it or not Mar 26 19:08:56 try it sometimes, its pretty usefull Mar 26 19:09:22 i'm taking that to be a hint of sarcasm :/ Mar 26 19:09:50 it sounds like a waste of time to me, and inefficient Mar 26 19:10:27 bugs default to oe@, a mail gets sent to the list so every dev sees and and a dev can *choose* to pick up a bug Mar 26 19:10:35 fairly efficient Mar 26 19:10:38 or not pick it up Mar 26 19:10:52 yes Mar 26 19:11:03 there's several hundred bugs that are not ever picked up Mar 26 19:11:06 that is the whole point Mar 26 19:12:38 bugs that are picked up, often are not re-assigned Mar 26 19:12:49 within your boundaries, you work on what you want Mar 26 19:12:59 this makes it very difficult to figure out which bugs are not being worked on Mar 26 19:13:46 RP: ping Mar 26 19:13:51 we have bugdays from time to tome Mar 26 19:13:52 ahh ,, all this cpu time feels so greatly used by oe baking ,,, there you know why you have all those GHz `s Mar 26 19:14:10 s/tome/time/ Mar 26 19:14:10 that's all well in good, but if i want to work on bugs that haven't been touched in a while, there's a basic flaw in trying to search for a bug that is neglected Mar 26 19:14:10 rob_w: heh indeed Mar 26 19:14:38 in that a bug which is neglected, looks exactly like a bug that is being actively worked on Mar 26 19:14:54 (the flaw is that people tend to not assign bugs to themselves) Mar 26 19:14:57 a bug that's actively worked on has fresh comments Mar 26 19:15:29 the bug tickets are not discussion forums Mar 26 19:15:38 that's what the ML is for, yes? Mar 26 19:15:44 not quite Mar 26 19:16:16 when you work on a bug you can put your progress in the comments Mar 26 19:17:04 yeah so if you're going to the trouble of putting your progress in the comments, that's a big if... only you and a couple other devs seem to do this, okay given that. why not assign the bug to yourself if you're working on it? Mar 26 19:17:34 my apologies are due if i offended you by assigning bugs you were actively working on to, you. Mar 26 19:17:38 because I might have worked on it during coffee break and leave it alone after that Mar 26 19:19:04 i agree with koen, assignments should only be used durring an active action state Mar 26 19:19:42 assigning something to someone who may not look at it for a year is useless when someone else who may like to fix the proplem now is willing to do it Mar 26 19:19:51 problem* Mar 26 19:19:59 emte: exactly Mar 26 19:20:02 well, then they reassign it to themselves Mar 26 19:20:14 if they're so motivated about it Mar 26 19:20:18 that shouldn't be a problem Mar 26 19:20:23 people by nature will not reassign themselves a task Mar 26 19:20:35 exactly my point Mar 26 19:20:37 when will the oz354fam083 be merged back to dev ? Mar 26 19:20:41 it's always assigned to oe@handhelds.org Mar 26 19:20:46 would you willingly assign yourself all the glibc tasks open? Mar 26 19:20:58 rob_w: never Mar 26 19:21:02 emte: yes, if i had an understanding of who liked to work on what Mar 26 19:21:08 ah Mar 26 19:21:26 and there inlines the problem Mar 26 19:21:35 so all apps need to move to oz354fam083 ? Mar 26 19:21:48 right now though, if no one assigns themself bugs, then it's completely unclear who has worked on what Mar 26 19:21:52 if there were dedicates devs to a specific issue auto assigning may make sense Mar 26 19:21:58 dedicated* Mar 26 19:21:59 rob_w: hmm? all apps are aleady in 3.5.4 Mar 26 19:22:01 except through monotone log, and that only shows devs with access to commit Mar 26 19:22:12 regardless of the inefficiencies of the bug tracking system, untested unused patches are considered cruft and there is no sense in adding them to the tree. once they're tested and modified to atleast build, if not work, they can be added with the other files required to make the whole thing function. that way you can revert that one change instead removing multiple commits at different points by different peo Mar 26 19:22:12 ple Mar 26 19:22:14 new apps should not be added Mar 26 19:22:32 TheMasterMind1: thanks :) i see that now Mar 26 19:22:34 CoreDump|home, ? hmm Mar 26 19:22:44 jnc, unassigned means that it is open to anyone who wishes to take it, or if it is fixed while fixing something else all the better Mar 26 19:22:46 CoreDump|home, where do those go then ? Mar 26 19:23:09 rob_w: you got something wrong here ;) Mar 26 19:23:29 3.5.4 is the stable branch, only bugfixes allowed (more or less) Mar 26 19:23:41 is this declared in OE ? -> ${CONFIG_SITE} Mar 26 19:23:46 .dev is the HEAD branch where development is continued Mar 26 19:24:46 it's a lot of duplication. all bugzilla function is verbosely duplicated on the mailing list, the mailing list kind of spills off into bugzilla tickets, and the monotone log only shows devs who have commit access and (invariably) are committing to solve bugs still assigned to oe@handhelds.org, so it's all confusing to me how exactly people are supposed to do their workflow and not get yelled at Mar 26 19:25:16 .dev is supposed to be a superset of .branch, so there's no need to merge stuff back. Except perhaps some distribution specific things that came in to polish the release and turn out to be of general use Mar 26 19:25:21 CoreDump|home, .dev is therefore the main branch and the other is just a release "prepare" branch Mar 26 19:25:23 thanks for being patient btw Mar 26 19:25:36 rob_w: exactl< Mar 26 19:27:31 koen: Can the EABI features be turned on in Linux Kernel 2.6.16 and still function (in non-eabi mode) on non-eabi system? you've done EABI work, maybe you'll know... Mar 26 19:27:35 it felt to me that people keep working on the oz354 one and i found myself running into issues whith .dev ,, so they then told me that these fixes are in .oz354 Mar 26 19:27:56 trying to sort out the defconfig situation for linux-openzaurus-2.6.16 kernel and angstrom-2006.9 Mar 26 19:28:18 the defconfig is not distro specific at the moment Mar 26 19:28:38 angstrom requires EABI function, other distros do not Mar 26 19:29:16 jnc: you can run an EABI kernel on a non-EABI rootfs when you have the oldabi compats bits set Mar 26 19:29:27 jnc: angstrom only requires that for ARM cpus Mar 26 19:29:31 ah Mar 26 19:29:50 however, you do need an EABI toolchain to compile an EABI kernel Mar 26 19:30:01 hmm. that sounds okay Mar 26 19:30:25 so you need a seperate defconfig for eabi Mar 26 19:30:26 would you suggest it might work to have defconfig-cxx00 default to AEABI features being built? Mar 26 19:30:30 ohhh Mar 26 19:30:55 eg packages/linux/linux-openzaurus-2.6.16/linux-eabi/defconfig-cxx00 Mar 26 19:30:59 it would work in the case of building angstrom, but somehow fail if you were building with a non eabi toolcahing hmm Mar 26 19:31:12 s/cahing/chain/ Mar 26 19:31:15 I'd rather have it decide at buildtime by the kernels that support EABI Mar 26 19:31:27 i.e. if TARGET_ABI = "EABI" Mar 26 19:31:34 mickey|writing: would have to do that with some kind of sed-fu :( Mar 26 19:31:35 echo >> CONFIG_AEABI=y Mar 26 19:32:02 i don't like that. makes it harder to look at what your defconfig is going to be when you compile Mar 26 19:32:18 mickey|writing: yeah, some python foo for that would be nice Mar 26 19:32:30 jnc, for some things it doesn't make sense to provide fine granular defconfigs that need to be synced Mar 26 19:32:30 mickey|writing: although I'm not clear on linux-gnueabi vs linuxeabi Mar 26 19:32:39 koen: me neither. lets wait for bero explaining that Mar 26 19:32:43 Is there some angstrom channel opened? :) Mar 26 19:32:51 not yet Mar 26 19:33:18 might be a good idea to register it... Mar 26 19:33:22 i'm abandoning openzaurus for my Z, in favor of angstrom and the new toolchain Mar 26 19:33:25 i guess koen did that already Mar 26 19:33:25 jnc: you can still look at the .config or defconfig in the compile dir Mar 26 19:33:31 CoreDump|home: way ahead of you ;) Mar 26 19:33:35 good Mar 26 19:33:48 JustinP: it still seems like a hack =) Mar 26 19:34:04 would rather this be implemented with some kind of class just for kernel building Mar 26 19:34:54 thanks for volunteering! ;-) Mar 26 19:34:58 hah. yeah Mar 26 19:35:08 well the idea of a conditional patch always bothers me Mar 26 19:35:18 how do you bitbake -w that? Mar 26 19:35:18 bbl Mar 26 19:36:50 mickey|writing: you are the uicmoc maintainer ;) Mar 26 19:37:01 new fun patches for you Mar 26 19:37:05 i'm afraid that's true Mar 26 19:37:06 heh Mar 26 19:37:14 please attach them to some bug and I'll have a go at testing Mar 26 19:37:33 okay, OE bugs #345 and #535 Mar 26 19:37:38 *nod* Mar 26 19:37:43 will check asap Mar 26 19:37:51 (asap being somewhere next week) Mar 26 19:37:54 <--- busy as hell Mar 26 19:37:57 thanks Mar 26 19:38:03 np Mar 26 19:39:08 think i could commit them to .dev? i'll be around to revert if people complain. it's just substituting type 'int' for type 'int_t' making it work on 64-bit and 32-bit compile hosts equally Mar 26 19:39:38 the other way around actually, it does s/int/int_t/ Mar 26 19:39:39 jnc, why dont you test and append? Mar 26 19:39:57 emte: no ia32 hardware that can build OE Mar 26 19:40:26 no i mean for your replacement Mar 26 19:40:36 my repla... Mar 26 19:40:48 substituting type 'int' for type 'int_t' Mar 26 19:40:48 * jnc scrolls up "what on hell was i talking about a few min ago?" Mar 26 19:41:05 oh no i got the english wrong Mar 26 19:41:35 you can still test and replace Mar 26 19:41:37 i meant everywhere there is an 'int' being used wrongly, it should instead be fixed to use int_t Mar 26 19:41:46 not all uses of 'int' are wrong Mar 26 19:41:53 the same way my flag works Mar 26 19:42:10 just use it where appropreate Mar 26 19:42:18 oh Mar 26 19:42:26 my goal is not to add more crap to OE Mar 26 19:42:37 not saying yours is crap, but mine would certainly be Mar 26 19:42:38 =) Mar 26 19:42:39 keeps compatability until all platforms can be verified Mar 26 19:43:17 it's not an OE problem, it's a bug in the uicmoc sources which can be fixed and made portable Mar 26 19:43:17 not all hosts are intel or amd Mar 26 19:44:04 the bugs list uicmoc-native and uicmoc3-native Mar 26 19:44:04 if it comes from upstream then it is an OE problem ... Mar 26 19:44:29 so yeah, i'm quite certain that build hosts will be ppc, amd, and intel Mar 26 19:44:59 there are also a few arm build hosts that i know of around too ... Mar 26 19:45:03 emte: no, upstream can always have problems too Mar 26 19:45:19 emte: in fact, all of these 64-bit patches should be pushed upstream Mar 26 19:45:24 JustinP, correct, but until upstream fixes them it is an OE issue Mar 26 19:45:25 jnc: have you been on that? Mar 26 19:45:48 JustinP: i've tried with other packages Mar 26 19:46:21 JustinP: i'm sort of afraid to approach TT about anything at all. Mar 26 19:46:23 it would be best....unless we get them pushed upstream we're going to have maintenance fun when we update versions... Mar 26 19:46:26 heh Mar 26 19:46:30 yeah, don't know about them Mar 26 19:47:12 get visions of men in black suits jumping out of black helicopters down ziplines and auditing my mother's computer for Qt-win32 Mar 26 20:11:12 Any front-end chess boards already in OE? Mar 26 20:11:43 xboard/eboard equivs? Mar 26 20:26:38 man we're getting raped Mar 26 20:27:22 we are? Mar 26 20:28:16 must be watching the playoffs Mar 26 20:29:04 yea, gmu v uconn Mar 26 21:07:53 03coredump 07org.oe.dev * r2e715e00... 10/packages/mplayer/mplayer_1.0pre7.bb: mplayer: fix a typo in PR Mar 26 21:09:56 base-files installs an fstab with usbdevfs mounting at /proc/bus/usb, however newer kernels require usbfs Mar 26 21:10:48 so i create a DISTO/fstab with what i want correct? Mar 26 21:11:00 or is it MACHINE Mar 26 21:11:20 good question Mar 26 21:11:22 both should work Mar 26 21:11:24 I'd got for machine Mar 26 21:11:47 s/got/go/ Mar 26 21:12:01 that would be probably best. Mar 26 21:12:27 alright, thanks Mar 26 21:12:32 mickeyl and I discussed the DISTRO override craziness at FOSDEM Mar 26 21:13:27 udev installs an init.d script which uses the --oknodo option with start-stop-daemon, but its not supported by busybox's start-stop-daemon Mar 26 21:13:40 that's a bug Mar 26 21:14:02 * CoreDump|home ran into that one, too Mar 26 21:14:10 a bug with what? should the fix be a patch to the udev script or busybox Mar 26 21:14:19 both would be fine Mar 26 21:14:27 I'd prefer a patch to busybox, though Mar 26 21:14:50 if you fix busybox, then it potentially fixes other packages. it also may bloat busybox (minor concern) Mar 26 21:15:17 koen|tv: there has to be _some_ reason we patched a bazillion initscripts instead of busybox Mar 26 21:15:47 CoreDump|home: you would think that Mar 26 21:15:53 CoreDump|home: patch script 30s, patch busybox 30m Mar 26 21:16:27 "busybox is scary, lets not touch that" Mar 26 21:16:34 right Mar 26 21:16:48 ~lart busybox Mar 26 21:16:49 * ibot frags busybox with his BFG9000 Mar 26 21:17:06 "Busybox: The nail in a scripters coffin" Mar 26 21:17:07 heh. i've done patches to busybox before and gotten yelled at by the uclibc folk for adding too much code Mar 26 21:17:20 i had a check to see if a file existed before deleting it Mar 26 21:17:59 which (never even occurred to me) could be done without Mar 26 21:22:45 ~lart usb Mar 26 21:22:46 * ibot declares usb a moron Mar 26 21:23:06 XorA|gone: what happened? Mar 26 21:23:39 Bernardo: on my Shuttle XPC if I have anything plugged into usb on boot except my keyboard and mouse, it doesnt detect the CF reader Mar 26 21:23:51 XorA|gone: which linux kernel? Mar 26 21:23:57 or is that windows Mar 26 21:24:06 jnc: its a bug in the BIOS Mar 26 21:24:14 maybe so Mar 26 21:24:17 which linux kernel? Mar 26 21:24:34 jnc: 2.6.15 Mar 26 21:24:43 try .16, some usb stuff was fixed Mar 26 21:25:01 jnc: its a bug in the BIOS Mar 26 21:25:02 are you using an usb hub? Mar 26 21:25:26 jnc: no Mar 26 21:25:45 i've got an Shuttle, and the card reader never seems to work. amd64 Mar 26 21:26:01 one of the sn85g3 i think Mar 26 21:26:33 jnc: Ive got V1 of that board, probably fixed in later revs, it also doesnt occur if you have ps2 keyboard Mar 26 21:27:12 jnc: if you have a PS2 keyboard disable USB keyboard and mouse in BIOS and that should fix it Mar 26 21:27:18 XorA|gone: did people report the problem you're experiencing for something other than Linux? Mar 26 21:27:33 jnc: yes, it happens in windows as well Mar 26 21:27:38 ah okay Mar 26 21:27:46 i wonder if that's fixable Mar 26 21:28:00 jnc: and was all over the forums 2 years ago when I bought the shuttle Mar 26 21:28:38 jnc: BTW is your shuttle unstable in linux, or is it just the overhearing problem on V1 board I see Mar 26 21:29:19 i have no troubles with noise or overheating Mar 26 21:29:27 it's the most reliable PC i have ever owned Mar 26 21:29:53 using the cpufreqd to scale down the CPU when not used helps Mar 26 21:30:00 on bootup i see many modprobe pnp:XXX failed errors. how can i find out what modules it wants? Mar 26 21:30:13 jnc: I think its just a V1 board fault, again it was all over the forums Mar 26 21:31:23 anyway I need to reboot to get cardreader working Mar 26 21:31:27 TheMasterMind1: try 'logread' Mar 26 21:32:12 mine is an SN85G4 v3 i think Mar 26 21:33:09 nope nothing Mar 26 21:33:21 its hotplug trying to modprob Mar 26 21:33:45 ooh and new drivers Mar 26 21:33:53 * jnc wanders off Mar 26 21:39:55 wow. 2 points up, 8 seconds left Mar 26 21:44:47 jnc: ahah, a hard power down and finally it works again :-) Mar 26 21:44:55 anyway time for bed for me, cyas Mar 27 06:12:31 morning Mar 27 06:57:15 hi all Mar 27 06:57:26 how can I add new fs type support for IMAGE_FSTYPES? Mar 27 07:29:39 hi Mar 27 07:34:35 good morning all Mar 27 07:36:10 morning all Mar 27 07:38:13 hey do13_ Mar 27 07:38:40 hi koen Mar 27 07:40:35 what's a working gcc / glibc combination for an i686 target? Mar 27 07:44:59 morning Mar 27 07:45:55 hi XorA Mar 27 07:47:04 hey XorA Mar 27 07:48:34 koen: is it not policy that all .bb files in OE be MIT/X11 licensed? Mar 27 07:48:45 http://lists.debian.org/debian-devel/2006/03/msg01084.html - gcc 4.1 on mips Mar 27 07:49:24 XorA: I think that's a nice goal, but afaik, it isn't a strict demand Mar 27 07:49:31 hey hrw|work Mar 27 07:59:58 koen: so how is EABI on this rainy morning? Mar 27 08:01:47 XorA: still not 100% stable in OE Mar 27 08:04:17 http://womble.decadentplace.org.uk/c++/syntax-errors.html Mar 27 08:10:43 meeting... Mar 27 08:20:44 koen: glibc-initial_2.4.bb sets FILESPATH to 'glibc-cvs' ? Mar 27 08:20:56 koen: Does this work for you? Mar 27 08:21:14 do13_: I can build 2.4, don't know much about it's internal magic Mar 27 08:23:00 koen: do you build glibc-initial_2.4.bb? I can't build glibc-initial_2.4.bb, because the patches won't apply Mar 27 08:23:33 good morning Mar 27 08:23:56 hey Bernardo Mar 27 08:24:05 do13_: glibc-intermediate and glibc Mar 27 08:24:16 those are in my work/ dir Mar 27 08:25:30 No glibc-initial? Mar 27 08:25:39 no sign of that Mar 27 08:42:07 Hi, I tried to compile apr but it failes with : checking whether TCP_NODELAY and TCP_CORK can both be enabled... configure: error: cannot run test program while cross compiling, may I suppres that test somehow? Mar 27 08:42:38 sws: I suspect you have to update the site files in org.oe.dev/site/ Mar 27 08:42:48 sws: or disable the test in the configure script Mar 27 08:44:21 may i savely disable that test? can TCP_CORK be enabled for PXA255 Mar 27 09:18:57 <_law_> morning al Mar 27 09:18:59 <_law_> l Mar 27 09:19:34 <_law_> someone build gconf-dus today? i get the following error http://pastebin.com/624659 Mar 27 09:19:48 <_law_> also updated pango today Mar 27 09:20:16 hi guys Mar 27 09:20:30 hey _law_ & obergix[work] Mar 27 09:26:30 RP: ping Mar 27 09:27:01 hrw|work: pong Mar 27 09:27:24 RP: /dev/rtc0 -> /dev/rtc - can you add symlink in udev rules? Mar 27 09:27:26 RP: ping! any mail? Mar 27 09:27:41 mithro: nothing so far... Mar 27 09:27:56 damn slow snail mail :P Mar 27 09:28:22 mithro: :) Mar 27 09:28:31 hrw|work: ok, I'll look into that Mar 27 09:29:17 RP: we have OE bug for this - 758 iirc Mar 27 09:29:31 I'm a touch depressed atm as a flaw has been discovered in the LED trigger code :-( Mar 27 09:29:41 Still trying to see a way around it... Mar 27 09:29:41 RP: ;( Mar 27 09:30:36 hi schurig Mar 27 09:30:43 hi schurig Mar 27 09:31:45 hi Cwiiis Mar 27 09:31:54 hrw|work: hi! Mar 27 09:31:58 schurig: did you got my mail about wlags 722? Mar 27 09:32:00 Hey hrw|work Mar 27 09:32:06 hrw|work: are we still with monotone? Mar 27 09:32:10 schurig: yes Mar 27 09:32:31 hrw|work: hmm, I got some mail about wpa-supplicant and GNU TLS, but nothing about wlags Mar 27 09:32:34 schurig: looks like monotone 0.26 will win SCM war (is many times faster then 0.25) Mar 27 09:32:57 schurig: ok. so I ask directly: you added wlags 718 into OE - do you have 722 working? Mar 27 09:33:28 hrw|work: no, I haven't. With a huge patch orgy I was able to compile it, but the driver then didn't work with my card Mar 27 09:33:51 schurig: ok Mar 27 09:34:03 schurig: wcf54 needs 722 iirc? Mar 27 09:34:21 hrw|work: what's wcf54 ? Mar 27 09:34:26 RP: should get something this week then Mar 27 09:34:33 schurig: linksys wcf54 - 802.11g cf card Mar 27 09:34:33 3-5 days they said Mar 27 09:35:49 mithro: ok, sounds intriguing :) Mar 27 09:37:57 hrw|work: I have some artem cards, they work with 718 Mar 27 09:38:29 hrw|work: I also once took the firmware out of 722 and put that into 718, that made the roaming behavior better, but produced from time to time strange lockups, so I reverted tha Mar 27 09:38:32 that Mar 27 09:39:00 ah Mar 27 09:39:23 one user asked about wcf54 support - I tried to get 722 buildable but failed Mar 27 09:39:46 hrw|work: you need some patches to get it compile Mar 27 09:40:34 hrw|work: do you have a screenshot of a recent opie on a clamshell? Mar 27 09:40:42 schurig: I know - used your ones from 718 Mar 27 09:40:53 koen: no Mar 27 09:41:09 I have "bugs-722.patch", "kernel-pcmcia-722.patch" and "kernel-26.patch" Mar 27 09:41:26 and "include-722.patch" Mar 27 09:42:07 I can send them to you as a basis, if you want. If you get the card working, I'd be interested into getting a fixed bugs-722.patch back ;-) Mar 27 09:42:18 http://ars.userfriendly.org/cartoons/?id=20060326&mode=classic Mar 27 09:42:56 hrw|work: however, the pcmcia-722.patch is for 2.6.13-somewhat, pcmcia has changed (again) to 2.6.15 (which I'm currently using) Mar 27 09:43:09 hrw|work: but that is an easy fix Mar 27 09:43:21 schurig: no such card here so it will have to wait - have other things in todolist **** ENDING LOGGING AT Mon Mar 27 10:59:56 2006