**** BEGIN LOGGING AT Wed Apr 11 02:59:57 2007 Apr 11 03:41:20 hi * Apr 11 03:42:16 I've just update gcc/glibc version in board's conf/distro files, how do i tell bitbake rebuild all rootfs files ? Apr 11 03:42:57 If i using bitbake my-board-image -c rebuild, it only re-assembly packages to file.image in tmp/deploy ? Apr 11 06:30:43 good morning Apr 11 06:31:10 good morning Apr 11 06:50:04 hi all Apr 11 07:02:00 wow. a build actually completed! Apr 11 07:09:10 psokolovsky_: is opie-image currently buildable in .dev? Apr 11 07:38:18 morning Apr 11 07:39:28 koen|away: If it breaks something, we won't do it ;) Apr 11 07:59:11 morning all Apr 11 07:59:53 morning Apr 11 08:00:43 morning Apr 11 08:01:28 morning Apr 11 08:18:09 morning Apr 11 08:20:31 hi XorA Apr 11 08:20:52 * XorA shoulda brought the collie to work Apr 11 08:29:56 hi greentux Apr 11 08:31:05 hi hrw... Apr 11 08:31:16 hrw did you visit the moko workshop in frankfurt? Apr 11 08:33:56 greentux: no, as I am not part of openmoko coreteam Apr 11 08:44:54 * hrw wants good working OS for nokia770... Apr 11 08:50:52 hrw: aaaahh hrw, You "are" the hidden corememeber... Apr 11 08:51:58 greentux: ? Apr 11 08:52:17 zecke|away: hello Apr 11 08:53:06 greentux: or you know something which I do not know Apr 11 08:53:12 hi schurig zecke Apr 11 08:53:24 hrw: hi Apr 11 08:55:42 hi Liam Apr 11 08:59:34 hrw: no right, but you are a core OE dev, so you are not so far from core moko devs... Apr 11 08:59:41 hi liam Apr 11 09:00:52 RP: thanks for answering my email on the non-deterministic issue. Apr 11 09:01:29 likewise: np. Things aren't perfect but should be better in later bitbake versions Apr 11 09:01:55 RP: later as in later than 1.8 rev 812? Apr 11 09:02:14 likewise: no, bitbake 1.8 is "later" in this case Apr 11 09:03:07 greentux: beiing OE core dev does not give me extra powers into this direction. Apr 11 09:03:55 greentux: working with OH gave me larger power when it comes to openmoko informations Apr 11 09:04:13 RP: OK, but in case of multiple providers without explicitly a preferred provider (and possibly only when combined with collections) it works undeterministic. Maybe by design, but this may fool ppl. I would consider bailing out instead of having bitbake selecting a provider "at random". Apr 11 09:05:24 RP: What is the design decision for resolving multiple provider ambiguity? (sp?) Apr 11 09:05:33 likewise: It does it deterministically in that it should always do the same thing. That behaviour was part of bitbake before I did anything with it and I've just preserved it. I agree an option to disable it might help people see whats going on more... Apr 11 09:05:55 likewise: The design is that it should try providers in turn until one works Apr 11 09:06:17 RP: Hmm, then I have a counter-case to debug. Apr 11 09:06:17 I'm not claiming thats right, just that its how it was written Apr 11 09:17:48 morning Richard, Marcin, all Apr 11 09:17:58 hi Liam Apr 11 09:18:07 ~prod XorA Apr 11 09:18:18 * ibot zaps XorA with a high voltage cattle prod Apr 11 09:18:20 likewise: You hadn't mentioned it was a collections problem. That probably gets messy :-( Apr 11 09:18:39 RP: I think I did mention that. lemme check Apr 11 09:18:46 I thought that was more fifers stylee with sheep :-) Apr 11 09:19:03 RP: In my email, I did (first line) Apr 11 09:19:44 likewise: Yes, you did although I didn't realise it was a problem with the collections specifically Apr 11 09:20:21 likewise: This is the same local.conf in both cases? Apr 11 09:20:36 RP: Yup, everything the same. See this to demonstrate what I mean: http://www.pastebin.ca/434241 Apr 11 09:20:37 hi, all! is there any target to produce scratchbox toolchain? Apr 11 09:21:32 RP: grabbing some coffee. bbiab Apr 11 10:09:27 likewise: Have you a PREFERRED_PROVIDER set? If not, how should bitbake choose one? Apr 11 10:09:53 * RP suspects it comes down to the order of the files on the disk :/ Apr 11 10:29:02 anybody knows about sbox toolchain package? Apr 11 10:30:15 so how to create sbox toolchain from OE toolchain to play with task-sdk-sbox generated SDK? Apr 11 10:48:11 * XorA is so bored of crappy half assed embedded arm stuff Apr 11 11:04:36 I wonder if http://www.arm.linux.org.uk/developer/patches/viewpatch.php?id=4302/1 fixes Irda crash on Z's Apr 11 11:06:18 XorA: check Apr 11 11:07:37 hrw: I am Apr 11 11:08:04 someone used Dell D400 laptop? Apr 11 11:16:09 RP: No, I did not have PREFERRED_PROVIDER set. However, bitbake should in all cases respect the priority of each bitbake collection and should not arbitrarely switch (as I showed, it switched between collections during subsequent bitbakes on the same package) Apr 11 11:17:20 RP: With PREFERRED_PROVIDER set, everything works. Therefore my proposal to either (1) respect bitbake collections priorities, or (2) bail out if bitbake is going to choose an arbitrary resolution for a package. Apr 11 11:17:52 XorA: Ah well, there are worse things to be bored off Apr 11 11:17:59 XorA: s/off/of Apr 11 11:40:06 likewise: It should respect collections priorities, I'm not sure why it doesn't Apr 11 11:40:54 hrw: D800 & D820 Apr 11 11:40:59 RP: Any idea how I could debug this? -vvv -D gives no hint on why bitbake chooses between them Apr 11 11:41:11 cbrake: wb Apr 11 11:41:38 likewise: Start adding print statements to providers.py I guess Apr 11 11:41:48 likewise: hello -- still working through the bbcollections issue ... Apr 11 11:43:05 cbrake: Yes, but pinpointed the problem some more. Seems a combo of not setting a PREFERRED_PROVIDER and using bbcollections, in which case I can get one machine to switch collections during subseq. bitbake invocations on the same package. Apr 11 11:43:20 RP: tnx. Apr 11 11:46:07 cbrake: d400 is 12" - I am thinkin about buying it or ibm x31/x32 Apr 11 11:48:53 03xora 07org.oe.dev * rfc80304b... 10/ (1 packages/aspell/aspell_0.60.5.bb): aspell_0.60.5.bb : fix packaging problem with .debug thanks insanity Apr 11 11:49:45 hrw: ibm is probably a little more rugged machine. My D800 has served me well, but it does not feel as rigid (it flexes more) than IBM machines, or the newer Dell machines (820). I think with the 820, they used more magnesium vs plastic. Other than that, I don't have any complaints about the 800. Apr 11 11:50:35 * XorA has a D420, nice and light Apr 11 11:51:01 * likewise has an Inspiron 8100, heavy and dark Apr 11 11:51:10 XorA: 12"? Apr 11 11:51:43 RP: can you change some occurences of "bb.msg.debug(1, " to "bb.msg.error(" in bitbake-1.8? E.g. all of the "No providers of build target" messages should be errors, not debug statements Apr 11 11:51:55 hrw: yeah Apr 11 11:52:04 RP: otherwise you do a bitbake, it suddenly stops and you don't really see why (except when you run with -d...) Apr 11 11:52:30 schurig, RP: +1 from me Apr 11 11:52:53 XorA: worth using? Apr 11 11:53:06 RP: also, some other messages might be changed from .note to .debug, because they aren't that interesting for users, e.g. "Resolving missing task queue dependencies" Apr 11 11:53:21 hrw: I like it as I have bad back, very easy to carry, keyboard is quite good for typing, runs debian Etch perfectly Apr 11 11:53:30 Sadly its not as simple as that as the multithreading considers all possiblities and sometimes a possibility doesn't have a provider but another possibility does. Changing it will make a ton of errors appear Apr 11 11:54:14 schurig: I agree it needs reworking but at the moment its on hold until we get the UI and can add something higher level to decide whether to show a set of errors or not Apr 11 11:54:46 schurig: I think the resolving message was there as it can take an age on a world build and users wanted to know what was going on :/ Apr 11 11:55:01 * RP can't win Apr 11 11:55:11 RP: my first suggestions is a bugfix, the second one an enhancement, so if you see this as a bugfix, then please apply :-) (or add me as a developer, so that I can do it, my berlios account must be holgerschurig) Apr 11 11:55:18 XorA: I need to create sheet with list of all available auctions with it and probably will choose one as they are easier to get then x31/x32 (and cheaper) Apr 11 11:55:32 RP: okay, than scrap my second suggestion Apr 11 11:55:53 hrw: has a SIM socket :-) but I dont know how to power up GSM in linux Apr 11 11:55:56 schurig: The first suggestion is not a bugfix as it will spew errors to users when things currently work :-( Apr 11 11:56:07 RP: I have usually less than <100 .bb files, because I build them in bunches. Apr 11 11:56:15 RP: then make it .note, not .error :-) Apr 11 11:56:29 schurig: it used to be, I got compaints :/ Apr 11 11:56:51 schurig: I agree it needs looking at but its not as simple as it looks Apr 11 11:56:53 XorA: ipaq h5000 have sim socket too Apr 11 11:57:07 XorA: but nothing use it Apr 11 11:57:08 RP: but I think it should be .error, because later you do "raise bb.providers.NoProvider(item)" Apr 11 11:57:17 if `whoami` | grep schurig then :-) Apr 11 11:57:33 hrw: does it have GSM hardware? Apr 11 11:57:44 schurig: That is trapped elsewhere? or perhaps not raised if in -k mode? Apr 11 11:58:22 schurig: I don't have the code handy but I'm 100% sure its not as simple as you think it is, I've had a lot of problems with that code Apr 11 11:58:24 XorA: no Apr 11 11:58:35 hrw: is that the one with figureprint scanner? Apr 11 11:58:42 yes Apr 11 11:58:53 hrw: then I know what it is for Apr 11 11:59:07 XorA: what for? Apr 11 11:59:18 hrw: your fingerprint is stored on smartcard Apr 11 11:59:27 hrw: for the decent figureprint software Apr 11 11:59:48 hrw: I worked on that Apr 11 12:00:02 schurig: If you mail me with a list of the messages you think should be notes/errors, I will look at them Apr 11 12:00:03 XorA: ipaq do not use that socket at all Apr 11 12:00:16 XorA: nevermind does it is pocketpc or linux Apr 11 12:00:40 hrw: look for a company called something line Precise Biometrics, Im pretty sure they use it in their software Apr 11 12:00:51 hrw: it is used within NIST in the US Apr 11 12:01:23 RP: only the two "No providers of build target" should be changed, because I think their current setting is plainly a bug. Bitbake (without -d) stops, but doesn't say why it stopped and why it couldn't build what it was asked to build. Apr 11 12:01:41 RP: in lib/bb/taskdata.py Apr 11 12:04:53 schurig: Look at the call sites of add_provider_internal - I suspect a better solution is to add an error before the raise in add_unresolved Apr 11 12:05:18 schurig: The other call site already has that Apr 11 12:07:14 schurig: Does http://www.rpsys.net/openzaurus/temp/bb-noprovider.patch solve the problem? Apr 11 12:13:29 RP: no Apr 11 12:13:40 RP: see output of those two changes: http://pastebin.ca/434424 Apr 11 12:16:16 schurig: ah, its for a failure somewhere down the task chain :/ Apr 11 12:16:29 RP: I'm still not getting why you have a problem. A bb.msg.note vs. a bb.msg.debug doesn't change any execution logic. Therefore it's an easy change, that cannot make any harm. Apr 11 12:16:51 but I agree that bb.msg.error might have some undesired side-effects :-) Apr 11 12:17:24 schurig: I worry about the case where you have a large respository and not all provider paths always work Apr 11 12:17:49 RP: then you have some NOTEs that you can ignore Apr 11 12:18:02 schurig: Users don't like NOTEs like that :/ Apr 11 12:18:03 RP: but now it stops building you are clueless why it stopped Apr 11 12:18:13 schurig: I agree that is a problem Apr 11 12:18:17 RP: users don't like to stay clueless ;-) Apr 11 12:18:30 schurig: But they don't like warnings that don't apply either Apr 11 12:18:52 schurig: Anyhow , I agree it needs fixing, its just a question of how to fix it Apr 11 12:20:01 * RP back shortly Apr 11 12:20:46 RP: hmm, if you build 5000 packages and in the middle of this it says "Can't build foo because barf is missing", then what problem makes this text? Apr 11 12:34:58 Hi. Has anyone hints on solving the circular dependency error when trying to compile opie-image for OpenZaurus-unstable ? Apr 11 12:41:50 hrw: seems to make irda happier anyway Apr 11 12:42:03 hrw: <7>pxa_ir: sir receiving error instead of an oops :-) Apr 11 13:08:22 schurig: Its not that simple. Imagine you have 3 providers of "foo". One of those providers depends on package "bar1" which doesn't exist. Another depends on "bar2" which doesn't exist. THe final one depends on "bar3" which does. If you then see two "No provider ..." messages, users tend to worry. We have that problem a lot in OE.dev :-( Apr 11 13:18:46 hello Apr 11 13:19:04 would someone help me resolving "VTYPE" problem? Apr 11 13:43:02 morning Apr 11 13:49:48 how do I change preferred version for specific package? Apr 11 13:53:33 PREFERRED_VERSION_specific_package="2.7.18" Apr 11 13:56:01 is that a shell command, parameter? Apr 11 14:01:26 param in local.conf Apr 11 14:11:39 is there some tool to automate bug reporting to OE bugzilla? I mean package build failures. Apr 11 14:33:59 hm, there are lots of those :-) Apr 11 14:34:13 why do I need to fix things manually...? Apr 11 14:37:21 djay-il: ? Apr 11 14:39:11 yea ? Apr 11 14:39:24 why, can't I complain a little? :-) Apr 11 14:40:13 and why the ... I can't compile libgvtype? Apr 11 14:40:23 libgpevtype Apr 11 14:41:06 djay-il: and why gcc does not build on my amd64? Apr 11 14:41:15 hehe :-) Apr 11 14:41:26 gcc sucks monkey bum Apr 11 14:41:36 I hear there lots of problems with amd64 Apr 11 14:41:51 actually there are very few amd64 problems Apr 11 14:42:35 well, I have a problem building binary locales on it, how's that? Apr 11 14:42:58 and why don't I have mimedir file? should I?... Apr 11 14:43:03 * djay-il is puzzled Apr 11 14:43:04 ENABLE_BINARY_LOCALE_GENERATION = "0" Apr 11 14:43:27 XorA: done that Apr 11 14:44:04 any answer for my above question please? Apr 11 14:44:29 djay-il: not a problem then :-) Apr 11 14:44:38 XorA: would you help me with something? Apr 11 14:45:10 djay-il: only if its simple enough for my stupid head Apr 11 14:45:35 well, lets see Apr 11 14:46:09 is there a way to have a package included in an image without using DISTRO_EXTRA_RDEPENDS? Apr 11 14:46:26 I tried DISTRO_EXTRA_RRECOMMENDS, but the package didn't end up in the image Apr 11 14:46:28 xuumbi: yes. you can create own image Apr 11 14:46:54 hrw: sorry, what do you mean "create own image"? Apr 11 14:47:20 XorA: take a look: http://pastebin.ca/434655 Apr 11 14:47:43 DISTRO_EXTRA_RDEPENDS works, but if I try to remove a package listed in there later on I have to use "ipkg -force-depends remove" Apr 11 14:48:06 djay-il: looks like a missing dependency Apr 11 14:48:11 it seems like I should be able to include a package in an image without making ipkg see it as a dependency of the distro Apr 11 14:48:16 maybe try and build mime-support Apr 11 14:48:18 xuumbi: image recipe Apr 11 14:48:55 mime-support... Apr 11 14:49:00 * djay-il is trying now Apr 11 14:49:15 hrw: ah, so right now I use "bitbake bootstrap-image", are you saying I could create my own "bootstrap-myimage"? Apr 11 14:49:21 djay-il: rather libmimedir Apr 11 14:49:27 xuumbi: yes Apr 11 14:49:28 ~fish koen|gprs Apr 11 14:49:34 hrw: done that Apr 11 14:49:34 * ibot slaps koen|gprs around with a large trout Apr 11 14:50:07 oh, how can I reinstall package? Apr 11 14:54:05 hrw: ok thanks Apr 11 14:56:17 wait, how can I tell what provides mimedir? Apr 11 15:30:22 can anyone tell me how to generate the Packages file? Apr 11 15:31:56 ipkg-make-index Apr 11 15:34:16 ah yes, thanks Apr 11 15:36:34 XorA: I had fish yesterday, today is a beef day :) Apr 11 15:37:16 ~prod koen Apr 11 15:37:39 * ibot zaps koen with a high voltage cattle prod Apr 11 15:38:08 if it's about libgcc and/or linking problems, I'm not home.... Apr 11 15:38:26 koen: nah, I just -force-overwrite on those Apr 11 15:38:47 koen: just nearest thing to beed I could think of Apr 11 15:39:29 XorA: on armv4t you need LD_PRELOAD=/usr/lib/libgcc_s.so.1 with the current feeds :( Apr 11 15:39:51 * koen has a clue why, but is afraid to dig deeper Apr 11 15:41:01 koen: Did you understand that bitbake bug I reassigned? Apr 11 15:41:37 RP: do_patch also fails sometimes and has the same symptons Apr 11 15:43:37 koen: Probably a different bug :/ Apr 11 15:44:03 koen: because everying depends on /usr/lib/libgcc.so.1? Apr 11 15:56:29 03florian 07org.oe.dev * rf058afdd... 10/ (1 packages/meta/meta-sdk-sbox-gpe.bb): meta-sdk-sbox-gpe: Use a nicer name. Apr 11 15:57:33 XorA|gone: I suspect because of ambiguous renaming Apr 11 16:18:12 http://www.theage.com.au/news/phones-pdas/palm-unveils-a-new-foundation/2007/04/11/1175971136619.html Apr 11 16:21:15 Crofton: hmm... anything new in there?! Apr 11 16:21:39 hum well... i got it Apr 11 16:21:47 Palm != Palm source Apr 11 16:33:03 ~curse person which broke X Apr 11 16:33:05 May the fleas of a thousand camels infest your most sensitive regions, person which broke X ! Apr 11 16:42:45 03koen 07org.oe.dev * ra97e0415... 10/ (3 files in 2 dirs): Apr 11 16:42:45 gcc-cross 4.1.2: bump PR till I properly fix libgcc (again) Apr 11 16:42:45 * Yes, I suck, sorry for the breakage) Apr 11 16:43:38 funny how a PR can hide a big problem in our toolchain Apr 11 16:56:55 koen: you added bootchart into OE? how to use it? Apr 11 16:57:20 hrw: read the .bb, that should have all the hints Apr 11 16:57:30 hrw: psokolovsky_ added a bootchart-image, btw Apr 11 16:59:17 looks like on progear it does not created log at all ;( Apr 11 16:59:27 root=/dev/hda1 vga=0x317 quiet init=/sbin/bootchartd Apr 11 16:59:29 ;( Apr 11 17:00:32 that sucks Apr 11 17:06:52 03florian 07org.oe.dev * r974375c5... 10/ (1 packages/gpephone/iksemel_1.2.bb): iksemel: Fix dependencies Apr 11 17:06:57 03florian 07org.oe.dev * r8d044c26... 10/ (4 files in 3 dirs): iksemel: Move to its own directory. Apr 11 17:07:02 03florian 07org.oe.dev * r78613968... 10/ (1 packages/gpephone/gloox_0.8.1.bb packages/gloox): gloox: Move to its own directory. Apr 11 17:09:13 ok, bootchart is not working here Apr 11 17:15:18 hrw: you used bash, etc? Apr 11 17:17:51 mickey|zzZZzz: wake up! Apr 11 17:19:21 psokolovsky_: bash is on device Apr 11 17:20:40 hrw: /bin/sh -> /bin/bash? Apr 11 17:21:04 koen: bootchart has #!/bin/bash Apr 11 17:21:27 and /bin/sh -> /bin/bash negate any checks Apr 11 17:22:21 ah Apr 11 17:26:14 koen: looks like the red tape finally caught up with ewi546 :( Apr 11 17:27:43 hvontres|poodle: really? Apr 11 17:28:00 11 19:34 [FreeNode] -!- hrw [n=hrw@ewi546.ewi.utwente.nl] Apr 11 17:28:40 hrw: nevermind...:) Apr 11 17:32:47 hvontres|poodle: I've been away for the last two days, so I might not have noticed Apr 11 17:35:27 03florian 07org.oe.dev * r9f53c3a6... 10/ (4 files in 3 dirs): disapproval of revision '8d044c26b37ff724e3f1c34dde3455615ef675c3' Apr 11 17:35:32 03florian 07org.oe.dev * r89bba7eb... 10/ (1 packages/gloox packages/gpephone/gloox_0.8.1.bb): disapproval of revision '7861396822409ca6c3e24762209c49eec2b2d690' Apr 11 17:35:36 03florian 07org.oe.dev * r403e3759... 10/ (4 files in 4 dirs): iksemel: Move to its own directory. Apr 11 17:35:40 03florian 07org.oe.dev * r7b148974... 10/ (4 files in 4 dirs): gloox: Move to its own directory Apr 11 17:47:45 cu Apr 11 17:48:32 koen: any luck with the server yet? Apr 11 17:48:58 likewise: no, I was stuck in den helder that past 2 days Apr 11 17:49:10 koen: oh yes, not oosterhout Apr 11 17:49:57 I'll try to poke gerwinin when I see him online Apr 11 17:51:02 koen: do you know what server model this is about? Apr 11 17:51:09 koen: except that it is a supermicro? Apr 11 18:00:15 anyone has any experience with the hfc-s isdn chips ? Apr 11 18:00:28 * wicked is away: afk Apr 11 18:14:31 likewise: 1U supermicro is all I know Apr 11 18:14:41 (1U is the maximum allowed here) Apr 11 18:19:53 vivijim: got succesfull in tracking the machine-dependent issue yet? Apr 11 18:21:27 anyone know what happened to opencores.org? Apr 11 18:21:59 never mind, it seems just slow. Apr 11 18:29:14 Anybody here familiar with Angel/Angelboot? I have a dd dump of the boot-partition from my zaurus, but it does not to be a cramfs image Apr 11 18:29:17 likewise: unfortunately no :( ... But the owner of this machine will be out of office for a week, So I'll check it next week... the notebook of this same person got another different problem... I don't remember exactly but I believe that the build was breaking on gcc-cross compilation... some related to unrecognized "-Qy" option passed to arm-mamona-linux-gnueabi-as ... Then I said him: "Man, please, format your machines!" hehehe Apr 11 18:32:17 likewise: koen: I'm very excited because I could build some .deb packages with OE and I've already setup a deb repository and coded a debootstrap script... Then it is already possible easily to do a chroot to mamona environment running all over qemu... :) Apr 11 18:33:00 timtimred: hello Apr 11 18:47:15 03philippe 07org.oe.dev * rbcc945f4... 10/ (4 files in 3 dirs): (log message trimmed) Apr 11 18:47:15 gtk-webcore: remove obsolete patches Apr 11 18:47:15 packages/gtk-webcore/osb-nrcit_svn.bb Apr 11 18:47:15 * update bb recipe Apr 11 18:47:15 packages/gtk-webcore/files/NRCit_HTTP_auth.diff Apr 11 18:47:16 * remove patch applied to svn Apr 11 18:47:18 packages/gtk-webcore/files/NRCit_unmap.diff Apr 11 18:47:19 vivijim: iirc the -Qy problem is because /bin/sh is /bin/dash (instead of bash) Apr 11 19:00:45 likewise: Afaik, armeb Ångrström builds were still broken not long ago Apr 11 19:00:57 ~1 week or so, I think Apr 11 19:01:54 NAiL: Ah, I didn't follow lately, but I know months ago BE seemed fine for EP93xx and IXP4xx. Apr 11 19:02:36 NAiL, koen: Is there a bug issued somewhere for this? Does it affect a EABI BE combo? Apr 11 19:02:38 Yeah. I know koen built angstrom for ixp4xx be (slug, actually) some months ago. Apr 11 19:02:53 yes, eabi be. Apr 11 19:04:29 re Apr 11 19:04:50 FYI: on MTD someone posted a patch to speed up mounting JFFS2 on Intel Flash by factors Apr 11 19:05:08 (for anyone who cares) Apr 11 19:05:19 that's interesting Apr 11 19:05:34 koen actually in the machines that the build is working we are using /bin/sh -> /bin/dash and in machine that it is NOT working is /bin/sh -> /bin/bash ;) sounds strange? Apr 11 19:05:55 vivijim: that's just plain.... wrong ;) Apr 11 19:06:17 vivijim: anyway, using dash *will* land you in trouble sooner or later. Apr 11 19:07:09 NAiL: I believe this... I've just noticed it when I was trying to figure out the differences between these machines... Apr 11 19:08:37 vivijim: good point. The machine that gave me problems today (a Gentoo system) ran bash as well. Apr 11 19:09:24 are we still depending on bash-isms then? Apr 11 19:23:30 likewise: GNU software assumes bash Apr 11 19:23:55 koen: By design? Apr 11 19:24:33 koen: or by accident? :-) Apr 11 19:27:04 likewise: "Bash is a Unix shell written for the GNU Project" (from http://en.wikipedia.org/wiki/Bash)...so do the math :) Apr 11 19:30:28 hvontres|poodle: Yes, then GNU s/w will probably assume bash features, not mere posix. Apr 11 19:31:10 koen, hvontres|poodle; didn't know this actually. I kind of liked the switch to a posix shell in ubuntu, at least it is standardized. Apr 11 19:31:30 although gnu is quite a defacto standard :-) Apr 11 19:35:09 I like it for exposing bash-isms Apr 11 19:35:19 too bad I grew up with bashisms Apr 11 19:35:36 ah its bash-bashing day ;) Apr 11 19:35:43 stuff that shouldnt work in bash part 1: cd // Apr 11 19:36:06 or is that a cd issue? Apr 11 19:36:08 heh Apr 11 19:36:21 something is clearly wrong with that anyway Apr 11 19:37:29 what's wrong with that? Apr 11 19:37:37 it works? Apr 11 19:38:41 why shouldn't it? Apr 11 19:38:44 and doesnt work in for example csh Apr 11 19:39:29 why should it? is / called // also ? :) Apr 11 19:40:06 POSIX says that multiple consecutive slashes are equivalent to a single slash. Apr 11 19:40:14 also it doesnt happen in other shells Apr 11 19:40:24 sure, but bash sees it as two Apr 11 19:40:25 "ls /usr/////bin" will work just fine too Apr 11 19:40:57 tims-computer-2:~ tim$ cd // Apr 11 19:40:57 tims-computer-2:// tim$ Apr 11 19:41:00 they're equivalent: it's perfectly valid for it to see it as either. Apr 11 19:41:33 if you did "cd /" and bash set $PWD to "////", that would be valid behaviour (albeit curious). Apr 11 19:41:46 in terms of POSIX valid it may be, but surely this is not desirable behavior Apr 11 19:42:16 dunno. it's hard to imagine what real problem it would cause. Apr 11 19:43:29 indeed. :) Apr 11 19:43:48 Sorry, let's talk OE. I like bash. I like POSIX. No bashing please :-) Apr 11 19:47:29 so yeah. typing ls in packages/ sucks :) Apr 11 19:48:39 at some point is there going to be categorisation? :) Apr 11 19:52:37 likewise: can I pm you briefly? Apr 11 19:52:56 blaster8: no problem. Apr 11 19:55:33 * NAiL agrees with timtimred on ls in packages/ Apr 11 19:56:12 it appears to be especially bad on xfs filesystems Apr 11 19:59:01 NAiL: why? slow? Apr 11 19:59:29 yes Apr 11 19:59:35 it appears to hang for a while Apr 11 19:59:42 then pushes everything Apr 11 19:59:53 (at once) Apr 11 20:00:21 making eg. ctrl-c pointless Apr 11 20:06:48 there's over 1500 folders in packages/. Browsing through that is a pain. Apr 11 20:33:39 03pfalcon 07org.oe.dev * r581aa13a... 10/ (1 packages/gpe-ownerinfo/gpe-ownerinfo_svn.bb): gpe-ownerinfo svn: SCM versions are always DEFAULT_PREFERENCE = "-1". Apr 11 20:41:26 mtn update Apr 11 20:41:36 gah :) Apr 11 20:43:28 RP: is there some sort of bootloader prompt on the Zaurus? I haven't seen anything on my serial cable before the actual linux boot. Apr 11 20:44:19 hvontres|poodle: No, there isn't Apr 11 20:45:02 RP : bummer... Apr 11 20:46:30 likewise: koen: look if you can help me to figure it out: my gcc_4.1.1-r12 package doesn't have the /usr/bin/gcc just the /usr/bin/arm-mamona-linux-gnueabi-gcc .... I've tried to debug and saw that the do_install function create a gcc link at tmp/work/gcc-4.1.1-r12/image/usr/bin but it is not present in deb package... what can be wrong? Apr 11 20:47:34 vivijim: looking what I have in my tree Apr 11 20:49:57 vivijim: that's a gcc for the target? I don't build it. Inspecting the .bb Apr 11 20:53:08 vivijim: sorry I cannot help now. I was about to go to get some sleep Apr 11 21:12:31 mtn dioff Apr 11 21:12:39 * RP sighs :/ Apr 12 00:48:23 what is "do_stage" for? Apr 12 00:48:50 the online doc don't have anything, really, about what it is **** ENDING LOGGING AT Thu Apr 12 02:59:57 2007