**** BEGIN LOGGING AT Thu Jan 10 02:59:57 2008 Jan 10 06:25:49 Question -- do any of the tasks/images generate a fully working system in ramdisk format? Jan 10 06:26:15 Looking for an example, hopefully, of making a fully working Linux system in uImage format Jan 10 06:26:20 that can be loaded over a network, etc. Jan 10 08:08:46 hi all Jan 10 08:29:50 morning Jan 10 08:30:54 gm Jan 10 08:39:57 is there someone working 'bout oe on toshiba e740 (following form brighthand forum)? Jan 10 08:45:18 hi, people! Jan 10 08:46:35 hi lumag Jan 10 08:47:33 pictux: no idea, but adding a new device is easy Jan 10 08:48:27 flo_lap: some tips? Jan 10 08:51:19 morning Jan 10 08:51:57 pictux: conf/machine/eseries.conf already cover e740 Jan 10 08:53:12 morning hrw Jan 10 08:53:29 ah, even better :) Jan 10 08:53:46 I do not know how good supported it is Jan 10 08:54:35 well, I figured it ot. :) Jan 10 08:59:47 thx hrw Jan 10 09:03:08 morning all Jan 10 09:10:15 hi RP Jan 10 09:19:23 bb in 2-3h Jan 10 09:29:28 morning Jan 10 09:30:10 RP: ping Jan 10 09:35:25 hidiho Jan 10 09:53:22 ant|work: pong Jan 10 10:00:57 RP: hey..you're 24hours on line lately! Jan 10 10:01:46 RP: I'd like to ask your opinion about the "fortran" legacy... Jan 10 10:05:44 RP: ...and got interested by that old languages Jan 10 10:06:28 RP: ...and eventually found some interesting syntax / grammar extensions Jan 10 10:07:03 RP: ...you would eventually implement in order to ease the use of bitbake ;-) Jan 10 10:09:03 RP: ...and thanks to the "polite" syntax (PLEASE DO...) could calm down some spirits... Jan 10 10:09:48 ant|work: I think I understand what you're saying... Jan 10 10:09:52 RP: ...I lost the link years ago .... Jan 10 10:09:58 RP: ...http://www.muppetlabs.com/~breadbox/intercal/ Jan 10 10:10:06 RP: don't kill me ;-) Jan 10 10:10:49 ant|work: What are you suggesting this is useful for? Jan 10 10:11:02 RP: http://www.muppetlabs.com/~breadbox/intercal/intercal.txt Jan 10 10:11:16 RP: just read the first few lines... Jan 10 10:12:11 RP: ^_^ Jan 10 10:12:45 RP: "PLEASE GIVE UP" is my favourite Jan 10 10:13:13 ant|work: right ;-) Jan 10 10:15:41 RP: ... Jan 10 10:15:52 RP: INTERCAL wil be a future project... Jan 10 10:16:08 RP: but back to fortran, seriously Jan 10 10:16:40 RP: I don't get why libmpfr is needed... Jan 10 10:17:08 RP: and if I try to bitbake gfortran-native ... Jan 10 10:17:22 There is no gfortran-native Jan 10 10:17:27 RP: bitbake complains I need gfortran on host.... Jan 10 10:17:32 RP: errr.. Jan 10 10:17:44 RP: *fortran-native ? Jan 10 10:18:05 There is no fortran native package Jan 10 10:18:24 RP: ans as extensions to gcc? Jan 10 10:29:16 re Jan 10 10:37:43 florian: hi Jan 10 10:39:33 RP: gcc3-build.inc and gcc-4-build.inc basically differs only on the "enabled languages" Jan 10 10:41:17 ant|work: right Jan 10 10:41:31 RP: but for gcc3 should be F77 Jan 10 10:41:35 RP: http://wiki.linuxfromscratch.org/blfs/wiki/gcc3 Jan 10 10:43:35 RP: FORTRAN = ",f95" -> FORTRAN ?= ",f77" Jan 10 10:43:45 RP: ...seems INTERCAL...almost Jan 10 11:00:51 XorA: I hope you don't mind me bringing up bug 2430. Can you take a look? Jan 10 11:02:35 Laibsch: god, I forgot all about that one with the dropping of Opie Jan 10 11:04:35 ant|work: I really don't know enough about this :/ Jan 10 11:06:21 XorA: Would be nice if you spent some time considering it. Jan 10 11:06:38 Laibsch: Im not allowed to :-( Jan 10 11:07:06 why? Jan 10 11:07:13 its not OM stuff Jan 10 11:07:18 Your wife keeps you on a short leash ;-) ? Jan 10 11:08:57 XorA: Well, maybe you can tell me what needs to be done and I could eventually give it a try Jan 10 11:09:21 Laibsch: Ill put it in the bug Jan 10 11:09:27 cool, thanks Jan 10 11:10:29 Laibsch: in fact the solution is already there, a /etc/update-fonts.d that the font-utils and qtttfontdir stuff both put scripts into Jan 10 11:10:51 Laibsch: then in postinst for fonts you run all scripts in that directoty, not fc-cache Jan 10 11:15:01 RP: I found some more info Jan 10 11:15:07 RP: Starting from gcc 4.0 the new Fortran native compiler gfortran is introduced Jan 10 11:15:17 RP: The gfortran compiler is the GNU implementation of the new standard Fortran 95 ISO/IEC 1539-1997 that contains most features both for Fortran 90 and FORTRAN 77. Jan 10 11:16:08 RP: so, no gfortran for gcc3 Jan 10 11:18:40 ant|work: ok, so someone's made a mistake when trying to enable things :/ Jan 10 11:19:09 ant|work: It needs someone to spend some time and sort this mess out. I just don't know who's going to do it... Jan 10 11:24:05 If a project allows redistribution of their code in source and binary form provided that the copyright notice is reproduced in the documentation and/or other materials provided with the distribution, does that prevent putting said information into the -doc package? Jan 10 11:24:25 I am talking about http://cvs.sourceforge.jp/cgi-bin/viewcvs.cgi/efont/sazanami/README?rev=1.5&view=markup (scroll down to see the English version) Jan 10 11:26:24 Crofton: hey, any idea where the gumstrix OpenEmbedded is? Jan 10 11:26:25 XorA: I don't have /etc/update-fonts.d. Am I missing a package I should depend on? Jan 10 11:26:32 Crofton: did they put it in svn? Jan 10 11:27:00 Laibsch: you need to make a package to create that file, a /usr/bin/update-fonts which calls run-parts /etc/update-fonts.d/ Jan 10 11:27:18 Laibsch: and add appropriate scripts to font-utils and qtttttttfont-dir packages Jan 10 11:27:38 Laibsch: and change font postint scripts to call update-fonts Jan 10 11:28:46 XorA: The problem is that the necessary steps don't even yet work by hand. For example, calling update-qttffontdir recognizes ttf-sazanami fonts. But still they won't show up in Appearance of opie. Jan 10 11:29:00 Creating an image that has the font pre-installed seems to work Jan 10 11:29:12 Later installation onto an existing image does not Jan 10 11:29:18 For whatever reason Jan 10 11:29:33 Laibsch: well I have no idea how qttffontdir works Jan 10 11:29:41 Laibsch: I have run opie for 3 years or so Jan 10 11:29:45 havent Jan 10 11:29:45 me neither, at least not much Jan 10 11:36:34 zecke, you hey put it in svn Jan 10 11:36:38 ~lart sakoman Jan 10 11:36:38 * ibot takes out sakoman with the trash Jan 10 11:36:40 :) Jan 10 11:37:10 zecke, let me find url Jan 10 11:38:17 Crofton: found it :) Jan 10 11:41:33 ah Jan 10 11:41:34 me too Jan 10 11:41:55 sakoman, is doing a reasonable job Jan 10 11:41:58 Crofton: hehe, they should put my Qt4 stuff into it :) Jan 10 11:42:10 put it in angstrom stable Jan 10 11:42:21 he better move to that soon Jan 10 11:42:22 Crofton: be nice, I'm listening Jan 10 11:42:33 he tried it and his images came out 2m bigger Jan 10 11:42:36 :) Jan 10 11:42:39 I know Jan 10 11:43:05 your work makes a better example based on the stable branch Jan 10 11:43:41 hrw|gone: When you come back, can you help me with some font problems? Jan 10 11:44:50 Crofton: the only other issue that was immediately apparent was wifi not working on first boot Jan 10 11:45:32 I'm hoping to move early next week if I can beat down the 3 issues Jan 10 11:46:21 sakoman: hehe, I only hate svn but I understand what you need to do and want to provide to your customers, so good job Jan 10 11:47:06 I should time a pull/ipdate from svn versus mtn Jan 10 11:47:22 although there tend to be fewer changes in gumstix ... Jan 10 11:48:37 ok, I need to go look at slides Jan 10 11:48:44 get some sleep sakoman Jan 10 11:49:12 Crofton: just woke up! have a plane to catch Jan 10 11:49:38 6am flights are murder! Jan 10 11:50:00 urg Jan 10 11:50:08 a friend has one tomorrow Jan 10 11:50:38 Is today your presentation? Jan 10 11:52:38 yeah Jan 10 11:53:08 I need to get the past the bit about it sucking and get a better attitude, before I start :) Jan 10 11:53:19 anyway, have a good trip Jan 10 11:53:32 I need to hide for a bit Jan 10 11:54:00 Good luck! ttyl Jan 10 11:55:19 re Jan 10 11:55:22 Laibsch: pong Jan 10 11:57:25 Laibsch: in general what XorA told is best way to solve problem Jan 10 12:01:41 hrw: Thank you for the reply Jan 10 12:02:09 I agree that XorA's suggestion is the best solution, we already established that at the time I wrote 2430 Jan 10 12:02:38 But I need to get something working, also as a way to understand what is really going on before creating something like that. Jan 10 12:03:11 IOW, even calling the necessary steps by hand does not work now, so we are not there yet to create scripts Jan 10 12:04:08 Laibsch: qtttfontdir (or how I called that) does not work? Jan 10 12:04:20 hrw: Can you think of a reason why "bitbake test-image-with-sazanami" will show the ttf font in Appearance, but "bitbake opie-image" and then installing the font via ipkd and calling update-qtttffontdir will not? Jan 10 12:04:37 update-qttffontdir shows the font is there Jan 10 12:04:48 But it will not show in Appearance even after a reboot Jan 10 12:05:11 I think I can provide two qemu images for you if you want me to, to show the problem. Jan 10 12:05:24 would be nice Jan 10 12:05:33 OK, give me some time Jan 10 12:09:48 o Jan 10 12:09:50 k Jan 10 12:10:21 hello Jan 10 12:10:36 !oebug 3647 Jan 10 12:10:37 * * Bug 3647, Status: NEW, Created: 2008-01-08 07:16 Jan 10 12:10:38 * * valentin.longchamp(AT)epfl.ch: New mx31moboard machine support in OE Jan 10 12:10:39 * * http://bugs.openembedded.org/show_bug.cgi?id=3647 Jan 10 12:10:57 could someone quickly have a look and eventually merge this one ? Jan 10 12:11:01 i.mx31 is getting more and more popular Jan 10 12:11:30 yep, it rocks ... our boards now is getting pretty good Jan 10 12:13:04 do you have a link for the board? Jan 10 12:13:13 hrw: I noticed that too... and there are quite cheap boards available here. Jan 10 12:13:17 ~lart myself Jan 10 12:13:18 * ibot grabs a large, mis-shapened log, with squirrels, and beats myself until only the nuts remain ... which the squirrels run off with Jan 10 12:13:23 Longfield: please reuse linux_2.6.* and linux.inc for kernel as much as possible Jan 10 12:13:46 Longfield: default interfaces file (netbase package) should be enough for your board Jan 10 12:13:56 * Crofton wonders if we should move some of linux.inc's smarts into a bblcass ... Jan 10 12:14:14 Crofton: sure Jan 10 12:14:27 Crofton should be working rather than wondering :-) Jan 10 12:14:35 Longfield: and please add link to board website into machine config Jan 10 12:14:41 I am afraif linux.inc will collapse under it's own success ..... Jan 10 12:14:50 exactly Jan 10 12:14:58 and not on kernel bbclass ... Jan 10 12:15:08 :-) Jan 10 12:16:56 Crofton: the problem is that someone has to do it Jan 10 12:17:10 yeah Jan 10 12:17:17 always the problem Jan 10 12:19:30 hrw: OK ... do you have an example of a good reuse of these files for the kernel recipe ? Jan 10 12:20:25 Longfield: there is linux_2.6.22.6.bb recipe Jan 10 12:21:06 Longfield: you need to add SRC_URI_append_YOURBOARD only Jan 10 12:24:29 hrw: http://oz.leggewie.org/wip/test-image-qemuarm.ext2.gz is the "test-image-with-sazanami"-type Jan 10 12:24:51 sazanami is already installed by bitbake and visible in Appearance as it should be Jan 10 12:25:26 fetching Jan 10 12:32:53 hrw: You will see that sazanami is displayed in the list of fonts, but that it will show just gibberish in the sample area. But that is a separate issue Jan 10 12:36:29 ok, booting now Jan 10 12:37:42 re Jan 10 12:38:22 my daily question - http://bugs.openembedded.org/show_bug.cgi?id=3593 anyone? :> Jan 10 12:38:30 hrw: and I did post to OEML as you suggested ;> Jan 10 12:39:17 heh heh I think this method leads to one of two possible outcomes :-) Jan 10 12:40:11 Laibsch: show me /usr/lib/opie/lib/fonts/fontdir from bad image Jan 10 12:41:13 hrw: is there a "standard" field for board website in the machine config file ? Jan 10 12:42:17 Longfield: ## Website: maybe? Jan 10 12:46:25 OK fine ... and the last thing I could not remove according to what I read in the install of the defconfig ... here is how it looks now: http://pastebin.com/m2cb5822c. Is this defconfig stuff done by any other recipe ? Jan 10 12:48:22 Longfield: create dir packages/linux/linux-2.6.22.6/mx31moboard/ - put defconfig file there. use linux/linux_2.6.22.6.bb recipe Jan 10 12:49:02 ok Jan 10 12:49:31 Longfield: you only need to add one line there Jan 10 12:49:46 SRC_URI_append_mx31moboard = "http://mobots.epfl.ch/mx31moboard/linux-2.6.22-moboard.patch.bz2;patch=1" Jan 10 12:49:50 and thats all Jan 10 12:50:53 * hrw -> out Jan 10 12:58:26 hrw|afk: http://oz.leggewie.org/wip/plain_opie_image-qemuarm.ext2.gz is the plain image. For your convenience, I mounted the loop image and copied the sazanami fonts to / for easier installation on your end Jan 10 12:59:02 hrw|afk: The information about /usr/lib/opie/lib/fonts/fontdir (plus some more) is at http://rafb.net/p/8AwSCh17.html Jan 10 13:23:26 hello, i have a problem: | configure: error: C compiler cannot create executables http://pastebin.com/m4c4713dd Jan 10 13:24:56 GNUtoo: did u have a look at the config.log Jan 10 13:25:07 yes Jan 10 13:25:09 GNUtoo: did u have a look at the config.log Jan 10 13:25:52 is it because of that? cc1: Invalid option `-frename-registers' Jan 10 13:26:05 wrong gcc version Jan 10 13:27:00 wrong gcc version? i use an external toolchain Jan 10 13:27:29 <_diego_> external toolchain is often a problem Jan 10 13:28:02 <_diego_> becouse oe is tested with recent gcc, and you use an old gcc version Jan 10 13:28:22 <_diego_> (probably) Jan 10 13:28:49 yes but as the device has proprietary apps i am obligated to use their toolchain no? Jan 10 13:30:48 <_diego_> maybe... depends on proprietary apps, your gcc version and oe gcc version Jan 10 13:32:10 gcc version is 2.95.3-10 Jan 10 13:32:17 wow... that's OLD Jan 10 13:32:17 :) Jan 10 13:32:25 <_diego_> btw usually packet should not use particular gcc option if not strictly required Jan 10 13:32:34 <_diego_> s/packet/package Jan 10 13:32:57 <_diego_> really old ^^ Jan 10 13:36:28 <_diego_> GNUtoo: btw u can build a distribution with recent compiler/library and then run proprietary apps with old library Jan 10 13:36:57 _diego_, ok what about glibc? Jan 10 13:37:19 GNUtoo: what are you trying to do? add some packages to an existing firmware? Jan 10 13:37:27 Jin^eLD, yes Jan 10 13:37:35 I did that a couple of times Jan 10 13:37:40 <_diego_> look here for an example: http://wiki.openmoko.org/wiki/Gllin Jan 10 13:38:14 afaik you have to match the glibc version, rest should not matter to much Jan 10 13:38:32 ok Jan 10 13:39:42 s/to/too/ Jan 10 13:39:55 <_diego_> arm processor, probably he should pay attention with which abi he is using Jan 10 13:40:08 oh yes, indeed Jan 10 13:41:54 _diego_, for the abi is a machine.conf ok (http://pastebin.com/m7ec4631f) Jan 10 13:45:00 how do i use an external glibc as glibc 2.2.5-10 is not in oe (same for the kenrel 2.4.19-rmk7-omap1) Jan 10 13:45:02 ? Jan 10 13:45:15 that's tough of course Jan 10 13:45:51 <_diego_> it's possible to have different glibc version on the system Jan 10 13:46:16 <_diego_> what is difficult, is to say which executable should user which glibc Jan 10 13:47:52 no i mean inside oe(not to make it build glibc but use an external one like i did with gcc) Jan 10 13:47:58 <_diego_> an exemple is what openmoko does to launch gllin (a propetary program compiled with old tools) Jan 10 13:48:51 <_diego_> you should define a libtool-native Jan 10 13:49:06 conf/distro/sharprom-compatible.conf is a distro in OE that uses an external glibc and gcc Jan 10 13:49:18 <_diego_> i suppose Jan 10 13:49:39 thanks a lot Jan 10 14:04:47 Laibsch: according to http://rafb.net/p/8AwSCh17.html it looks like qtttfontdir was not called when ttf-sazanami was installed Jan 10 14:11:18 hrw: qttffontdif or update-qtttffontdir? Jan 10 14:12:01 As you can see, I called the latter by hand and still got no valid /usr/lib/opie/lib/fonts/fontdir Jan 10 14:12:08 s/valid/correct/ Jan 10 14:12:46 update-qtttffontdir Jan 10 14:13:06 Laibsch: you need to do "update-qtttffontdir /usr/share/fonts/truetype" >/usr/lib/opie/lib/fonts/fontdir Jan 10 14:13:28 Laibsch: as update-qtttffontdir do not update file - it parse fonts and output data Jan 10 14:13:45 All right, I see Jan 10 14:14:02 Thank you for the hint Jan 10 14:14:12 no problem Jan 10 14:14:26 Laibsch: iirc qpf.bbclass has that code Jan 10 14:17:53 In there it says update-qtfontdir and not update-qtttffontdir. There is also no redirection to the fontdir file, but an -f parameter Jan 10 14:18:41 hrw: Maybe the code in there is faulty? Jan 10 14:19:30 hi Jan 10 14:20:25 any reason for having notecase version 1.2.6 as the default version instead of 1.7.2? Jan 10 14:21:46 Bernardo: Look through mtn log, the reason might be stated in one of the commit messages Jan 10 14:22:14 why it looks for my toolchain at /usr/local/openpma/arm if i've not set this directory anwhere... Jan 10 14:22:40 s/it/does it look Jan 10 14:25:13 Laibsch: nothing there that I can find, only that koen added notecase 1.7.2 (as I had asked him then) in november Jan 10 14:25:22 the log only has "update to 1.7.2" Jan 10 14:27:09 Bernardo: mtn log --diffs conf/distro/include/angstrom-2007-preferred-versions.inc |less Jan 10 14:27:40 You can also always override the ?= locally Jan 10 14:28:34 Laibsch: you need update-qtttffontdir for ttf fonts and then update-qtfontdir for qpf ones Jan 10 14:30:09 Laibsch: thanks, I was looking at the logs only for the notecase packages Jan 10 14:30:10 hrw: OK, then the code is not in qpf-bbclass Jan 10 14:30:43 hrw: and there is no ttf-bbclass Jan 10 14:31:12 s/-/./g Jan 10 14:31:56 Laibsch: mtn rename classes/qpf.bbclass packages/qpf-fonts/qpf.inc was long time on my todolist Jan 10 14:32:03 hrw: Should I create ttf.bbclass modelled after qpf.bbclass? Jan 10 14:32:35 no Jan 10 14:32:49 rename qpf.bbclass to qpf.inc rather and change qpf-fonts recipes Jan 10 14:32:58 OK, I can do that Jan 10 14:33:13 But that by itself won't do any good to sazanami Jan 10 14:33:17 ??? Jan 10 14:33:18 which is ttf not qpf Jan 10 14:33:38 xchat bug Jan 10 14:34:04 grep -i /usr/local/openpma -r * returns nothing Jan 10 14:34:19 (i couldn't paste one line of console here) Jan 10 14:34:36 webkit-gtk is not compiling for me: http://rafb.net/p/XPW5EL40.html. Am I missing something? zecke, mickeyl? Jan 10 14:34:42 Laibsch: you need to implement that stuff which you discussed before: /etc/update-fonts.d/ Jan 10 14:34:51 Laibsch: the only entry of a preferred version for notecase was in october, when most versions were added. Jan 10 14:35:22 Laibsch: so under opie it will does not mater do you install qpf or ttf fonts as system will run-parts /etc/update-fonts.d/ and will generate fontdir for qpf and ttf fonts Jan 10 14:35:22 * mickeyl points to zecke Jan 10 14:35:30 * mickeyl calls "webkit mastr" Jan 10 14:35:42 Bernardo: october + november: I guess it can be updated, but you need to ask koen about it Jan 10 14:36:26 ok, thanks Jan 10 14:36:39 hrw: OK, I will try and cook something up and post it to the BTS for your and XorA's review Jan 10 14:36:56 anyone for my problem? Jan 10 14:37:21 GNUtoo: You copied from sharprom code? Jan 10 14:37:37 Laibsch: ok Jan 10 14:38:59 Laibsch, i didn't exactly copied because sharprom-toolchain-native doesn't seems to be my toolchain (PREFERRED_PROVIDER_virtual/libc = "sharprom-toolchain-native") Jan 10 14:39:46 instead i added ASSUME_PROVIDED += " virtual/libc " Jan 10 14:42:56 GNUtoo: You need to look into the sharprom stuff and understand where the setting comes from to look under /usr/local for the toolchain Jan 10 14:43:03 I cannot help you with that, though Jan 10 14:44:12 ok but i didn't copied from sharprom... Jan 10 15:01:34 hrw: Do you have an idea why the sazanami fonts show nothing but gibberish when chosen? Jan 10 15:02:39 GNUtoo: Maybe http://blog.leggewie.org/?p=11 has some information for you on how it can be done for SharpROM and how you can adapt that to your situation Jan 10 15:02:52 Laibsch: no idea Jan 10 15:02:55 thanks a lot Jan 10 15:03:05 Laibsch: it was long time ago when I used opie last time Jan 10 15:14:34 bye all! Jan 10 15:18:53 GNUtoo http://www.openembedded.org/user-manual&dpage=commonuse_prebuilt_toolchain Jan 10 15:19:00 I guess that should be where the info is Jan 10 15:19:32 ok thanks a lot Jan 10 15:50:48 Is there a fundamental difference between a plain old .inc files with basically nothing but function definitions and a bbclass? Jan 10 15:51:24 I was thinking about separating some common functionality for the Zaurus line either an inc file in conf/machine/ or a bbclass Jan 10 15:51:36 into either Jan 10 15:54:42 Laibsch: .inc files are traditionally like .conf files, .bbclass files are equal to .bb files Jan 10 15:54:58 Laibsch: but somewhere along the lines .inc files became as powerful as .bbclass Jan 10 15:55:18 It really depends more on the context of what the functions do Jan 10 15:56:08 RP: What would you advise in this case? Jan 10 15:56:21 I want to separate out the installkit function, for example Jan 10 15:56:28 Laibsch: why? Jan 10 15:56:43 hrw: One single point to fix things Jan 10 15:56:55 For example, the function is defective for collie Jan 10 15:57:16 Laibsch: If its in a .inc I'd leave it in a .inc for now Jan 10 15:57:21 Laibsch: include/zaurus.inc would be best Jan 10 15:57:31 RP: It is not separated out, yet Jan 10 15:57:46 Laibsch: Where is it currently? Jan 10 15:58:00 in the individual machine configs Jan 10 15:58:45 Laibsch: Its in zaurus-2.6.inc isn't it? Jan 10 15:59:13 Let me check again to be sure Jan 10 15:59:23 It certainly is in the collie config Jan 10 15:59:52 collie-2.6.inc ? Jan 10 16:00:03 That does look duplicated but only for collie... Jan 10 16:00:19 OK, in that case, I will fix the inc file and delete the entry in collie Jan 10 16:00:27 Laibsch: Sounds good Jan 10 16:00:30 I had not checked other machines, but just assumend Jan 10 16:00:33 assumed Jan 10 16:00:38 Alright, then Jan 10 16:04:45 i don't find sharprom-toolchain-native anywhere...how can i create one for my device? (openpma-toolchain-native) Jan 10 16:05:39 GNUtoo Have you read my blog? Jan 10 16:05:51 Laibsch, i've read the link Jan 10 16:06:50 OK, then maybe you should give a quick try to compile for Sharp ROM Jan 10 16:07:08 Once that is done, you should probably know much better how to achieve your initial goal Jan 10 16:07:09 compiling for sharp rom works Jan 10 16:07:15 and it even runs on my device Jan 10 16:07:29 Then I don't understand why you ask for sharprom-toolchain-native Jan 10 16:07:40 That toolchain is *external* Jan 10 16:07:48 (vim but i don't have the coulors even with the syntax files...) Jan 10 16:07:49 Downloaded from somewhere Jan 10 16:07:58 yes that's what i did Jan 10 16:08:11 but i'll like to do the same for my device Jan 10 16:08:37 GNUtoo: Have you read conf/distro/sharprom-compatible.conf ? Jan 10 16:08:53 That should be the starting point I was trying to get you to. Jan 10 16:09:07 I think everything you need to know should be in there Jan 10 16:09:14 Laibsch, yes that's there that i found that...PREFERRED_PROVIDER_virtual/arm-linux-gcc-2.95 = "sharprom-toolchain-native" Jan 10 16:10:26 GNUtoo Come on Jan 10 16:10:28 the sharprom-toolchain-native isn't magic it must have been defined somewhere...i tried with grep -r but i didn't find anything Jan 10 16:10:33 Found it in 2 seconds Jan 10 16:10:35 packages/sharprom-toolchain/sharprom-toolchain-native_1.0.bb Jan 10 16:10:42 ahh Jan 10 16:10:56 i thought it was a config file... Jan 10 16:12:13 thanks a lot... Jan 10 16:12:39 rgrep is a good idea Jan 10 16:12:43 so is locate ;-) Jan 10 16:13:20 I am glad we found it without bothering mickey_away who is the original author but notoriously busy ;-) Jan 10 16:13:35 I guess you finally should be ready to go Jan 10 16:14:08 is he the one who reverse engeenered the zaurus sd card driver? Jan 10 16:14:22 or is my memory so bad? Jan 10 16:17:27 GNUtoo: it was not mickey Jan 10 16:17:36 who was it? Jan 10 16:17:45 I do not remember now Jan 10 16:18:09 GNUtoo Take a look through my blog Jan 10 16:18:18 It was some Spanish guy, I think Jan 10 16:18:29 And thesing who is actually doing the work now Jan 10 16:24:06 Is it true that collie needs zImage and the others a zImage.bin? Jan 10 16:24:45 Laibsch: others need what zaurus-updater need Jan 10 16:25:02 hrw: and that being? Jan 10 16:25:16 other Zaurus is what I am talking about Jan 10 16:25:21 it can be zImage, zimage.bin or even "this-is-my-favorite-kernel-please-flash-it" if you change updater Jan 10 16:25:42 hrw: I think you are the master of updater.sh Jan 10 16:25:52 How about a uniform zImage? Jan 10 16:26:01 at least for the Z Jan 10 16:28:04 Laibsch: impossible Jan 10 16:28:25 Laibsch: you still can not get kernel for 25x/27x in one image Jan 10 16:28:25 why? Jan 10 16:28:36 and we are hard to fit in 1.2M Jan 10 16:28:44 hrw: I am talking only about Z Jan 10 16:29:00 Can we change that at least to simplify things? Jan 10 16:29:07 Laibsch: poodle/tosa/c7x0 are pxa25x. akita/spitz are pxa27x Jan 10 16:29:12 And what does the size have to with the filename? Jan 10 16:29:23 ah.. file name. Jan 10 16:29:25 sorry Jan 10 16:29:27 yes Jan 10 16:29:38 Laibsch: zaurus-updater allow to use zImage for flashing Jan 10 16:29:47 uniform name of zImage Jan 10 16:30:05 OK, I will make a corresponding change to the installkit code, then Jan 10 16:31:45 ok Jan 10 16:47:13 anyone know what distro I should use to build maemo4 packages? Jan 10 16:50:07 maemo-1.0 ? Jan 10 16:51:47 maemo? what it is? Jan 10 16:52:49 hrw: hi Jan 10 16:53:04 the linux os on the nokia n series tablets Jan 10 16:53:49 ant|work: hi Jan 10 16:54:00 hrw: about updater.sh Jan 10 16:54:03 scruggs: ah.. the mistic Finnish tablets Jan 10 16:54:06 ant|work: yes? Jan 10 16:54:20 hrw: CoreDump last commit implemented u-boot? Jan 10 16:54:35 hrw: yep :) Jan 10 16:54:36 ant|work: yes. but we will not support using uboot Jan 10 16:54:40 hrw: I read your blog... Jan 10 16:54:56 hrw: anyway I will try on my c7x0 Jan 10 16:55:10 hrw: with the help of serial cable ;) Jan 10 16:55:18 ant|work: remember that you will have old u-boot Jan 10 16:55:26 hrw: question is: would it actually boot? Jan 10 16:55:49 ant|work: if you are familiar with uboot then you will boot Jan 10 16:56:00 hrw: you talked about missing W100 driver, but not for console I suppose ... Jan 10 16:56:04 out-of-box it will rather not Jan 10 16:56:12 ant|work: c7x0 display is w100 Jan 10 16:56:33 hrw: mean doesn't u-boot support a generic vga? Jan 10 16:56:41 hrw: X does Jan 10 16:56:51 ant|work: there is no such thing as generic vga on !x86 Jan 10 16:56:58 hrw: xserver-kdrive-imageon (2 open bugs btw) Jan 10 16:57:10 ant|work: xserver-kdrive-imageon use w100 Jan 10 16:57:32 hrw: I assume because there is no bios chip? Jan 10 16:57:43 hrw: there is no letsay VESA Jan 10 16:57:57 hrw: in "hardware" Jan 10 16:58:09 ant|work: forget about vesa/vga if you talk about !x86(-64) Jan 10 16:58:29 hrw: other bootloaders support it... Jan 10 16:58:42 hrw: grub / lilo (x86) Jan 10 16:59:02 ant|work: no - they do not Jan 10 16:59:03 hrw: ...as you see I'm novice for arm Jan 10 16:59:22 ant|work: lilo and grub do not support vga - they take what bios set Jan 10 16:59:38 hrw: "support" is a big word.... Jan 10 16:59:52 hrw: I support my wife .... Jan 10 17:00:03 ant|work: what you can get in u-boot on c7x0 is serial console, flash, mmc, keyboard and maybe pcmcia Jan 10 17:00:14 hrw: fine Jan 10 17:00:35 hrw: but I have to hardcode the CONSOLE string Jan 10 17:00:43 hrw: right? Jan 10 17:00:58 hrw: or is it interactive with serial console? Jan 10 17:01:13 ant|work: you never played with u-boot before? Jan 10 17:01:26 hrw: enough to reflash nand over it Jan 10 17:01:40 hrw: and restore some "params" Jan 10 17:01:57 hrw: fuc*ed by pdaxrom testing..... Jan 10 17:02:11 ah Jan 10 17:02:24 hrw: but then I didn't have serial cable... Jan 10 17:02:54 hrw: If it is like with other bootloaders (redboot bootbase) I know how to start at least.... Jan 10 17:03:45 ant|work: you connect serial and (if u-boot is configured to respond on serial) you get prompt on it Jan 10 17:03:54 hrw: I'll try and cry for your help if lost Jan 10 17:04:06 hrw: thanks in advance Jan 10 17:04:11 hrw: ;-) Jan 10 17:05:19 ant|work: I never used u-boot on Z and do not plan to Jan 10 17:05:47 hrw: if you help me on this I owe you a new battery for Z! Promised! Jan 10 17:07:53 ;D Jan 10 17:18:18 i still have the same error | The external toolchain could not be found in /usr/local/openpma/arm! but i have my toolchain there!!! it's even chmod 777 and chown to my user Jan 10 17:18:38 that is with my own toolchain...the one from sharprom works fine Jan 10 17:29:04 anyone? Jan 10 17:49:21 bb Jan 10 18:23:50 anyone? Jan 10 18:24:08 i even tried trace.py but with no results... Jan 10 18:24:26 and the bin/ is in my PATH Jan 10 18:28:12 bonne nuit Jan 10 19:17:56 Crofton|work: I just noticed a NAND driver for DaVinci (v4) was submitted. You already got that? Jan 10 19:30:39 re Jan 10 19:33:39 i've seen that "Toolchains are not relocatable" at http://free-electrons.com/training/devtools and it also says that it's not true anymore for gcc4 but it's gcc4 for building the toolchain or as toolchain itself? Jan 10 19:38:01 GNUtoo: even if it is supposed to work for the basic toolchain i would not recommend it - libtool will bite you in the end ;) Jan 10 19:38:17 hi pH5, hi pvanhoof Jan 10 19:38:24 ei Jan 10 19:39:07 hej flo_lap Jan 10 19:58:55 bye Jan 10 20:06:16 zecke pasted ".pri change for Laibsch" at http://paste.lisp.org/display/53982 Jan 10 20:06:26 Laibsch: ^^ Jan 10 20:26:32 how do i filter flags? because i have cc1: Invalid option `-frename-registers' Jan 10 20:34:02 GNUtoo: Do you want to filer? I think you just want to set a different optimisation? Find out where in conf/ -frename-registers get set Jan 10 20:34:15 GNUtoo: and then set that key to something else in your local.conf Jan 10 20:34:35 ok thanks i'll look Jan 10 20:38:18 zecke, it is set nowhere in conf grep -r -i frename-registers * only show me the content of conf/tmp Jan 10 20:40:45 the other conf Jan 10 20:42:18 timtimred, in org.openembedded.dev? Jan 10 20:43:48 yeah Jan 10 20:45:51 timtimred, thanks a lot it was in bitbake's configuration Jan 10 20:46:10 np ;) Jan 10 20:47:33 GNUtoo: which dirs did you grep? Jan 10 20:48:12 GNUtoo: I find at least 10 hits in my conf/ directory of OE Jan 10 20:49:19 zecke, /oe/org.openembedded.dev/conf Jan 10 20:49:42 i had others march bit they were distribs Jan 10 20:50:33 ??? Jan 10 20:51:51 zecke, what don't you understand? Jan 10 21:27:59 i have a strange faillure for nano | nano.h:76: curses.h: No such file or directory...normaly oe should handle the dependencies... Jan 10 21:28:20 and how do i clean all what i've built? Jan 10 21:28:57 (because i want a clean rebuilt on my external toolchain and not on sharprom's(the one i used at the beguining)) Jan 10 21:29:38 and what happend to the deps? do they get into the ipkg? Jan 10 21:32:15 GNUtoo: yes to both, the safest would be to wipe the TMPDIR excluding the configuration subdir Jan 10 21:33:10 flo_lap, ok and is there a mode for generating an ipkg.conf and the dependences ipkg? Jan 10 21:34:07 GNUtoo: the dependencies are generated from the library dependencies and the RDEPENDS information in the bb. you need to specify what you can't generate Jan 10 21:35:00 flo_lap, no i mean for instance if i want openssh i want openssl and openssh ipks not an ipk with both inside... Jan 10 21:35:15 GNUtoo: if you build an image it will contain a matching conf according to your distributions settings. Jan 10 21:35:32 ok Jan 10 21:36:52 thanks a lot Jan 10 21:37:06 GNUtoo: ah, right... that's what you usually get. things are split up properly. if you build some target such as openssh you will get all the depending ipkg packages as well. Jan 10 21:57:06 flo_lap, i cleaned my temps dirs and the curses.h is still missing should i bugrepot? http://pastebin.com/m3f64dab2 Jan 10 21:59:30 GNUtoo: that's strange indeed.... after the staging step of ncurses it should be there Jan 10 22:03:14 GNUtoo: did this happen with the other toolchain too? Jan 10 22:03:37 flo_lap, don't know...i'll check Jan 10 22:04:02 * flo_lap guesses that ncurses is too easy to break ;) Jan 10 22:04:43 flo_lap, i also have a strange problem my old ipkg can't use the ipk made by oe: http://pastebin.com/m17db6cb2 Jan 10 22:05:50 GNUtoo: ah i remember this, we had somethgin like this in Familiar a long time ago Jan 10 22:06:15 flo_lap, should i upgrade the ipkg of my device? Jan 10 22:06:55 GNUtoo: no... iirc that's caused by an old busybox tar or an old tar on your build machine. Jan 10 22:07:17 * flo_lap feels old Jan 10 22:07:26 pb_: do you remember? :) Jan 10 22:11:57 flo_lap, where do i put my busybox config file? Jan 10 22:14:27 GNUtoo: check the subdirs in packages/busybox Jan 10 22:16:01 GNUtoo: you can have things like busybox-1.7.2//defconfig Jan 10 22:16:14 wow thanks a lot Jan 10 22:16:32 or the device instead if distro Jan 10 22:17:09 it picks the most specific one Jan 10 22:17:30 that's one of the really nice things in oe Jan 10 22:18:39 flo_lap: I think it's a too-old tar on the build machine, but I don't remember for sure Jan 10 22:20:49 GNUtoo: what distro do you use on your build box? Jan 10 22:20:57 flo_lap, gentoo Jan 10 22:21:19 * flo_lap wonders if there could be such an old tar in there Jan 10 22:22:05 1.18-r2 Jan 10 22:23:56 flo_lap, gentoo : tar-1.18-r2,gzip-1.3.12 busybox:v1.1.3 Jan 10 22:24:53 i upgrade gentoo's tar Jan 10 22:25:15 I have 1.19 on my main build box... that runs Debian Lenny Jan 10 22:33:58 flo_lap, nmap_3.81-r2_arm.ipk: Debian binary package (format 2.0) and iptables_1.3.5_arm.ipk: gzip compressed data (iptables was made by hand using the ipkg_build script) Jan 10 22:42:07 GNUtoo: ok the first one is the correct imho. Jan 10 22:42:53 GNUtoo: the ipk itself is in fact an ar archive instead of a tar archive. Jan 10 22:58:48 flo_lap, | arm-linux-gcc: unrecognized option `-static-libgcc' Jan 10 23:08:07 flo_lap, here a more precise compilation faillure: http://pastebin.com/m38f4eb8c Jan 10 23:10:18 GNUtoo: is that the default crosscompiler oe built? Jan 10 23:10:27 flo_lap, no Jan 10 23:12:17 GNUtoo: which version of gcc is that? Jan 10 23:12:37 very old one...2.95 Jan 10 23:13:10 2.95.3-10 Jan 10 23:14:42 eeks Jan 10 23:14:49 yes, that's old indeed Jan 10 23:15:18 * flo_lap wonders if one of the Collie gurus is here... they sufferend from the same Jan 10 23:16:15 flo_lap, there is also sharprom-compatible that has the same kind of old compilers Jan 10 23:17:36 I would try an older busybox Jan 10 23:18:11 Lots of new stuff depends on modern compiler features Jan 10 23:19:11 ok but i also have a problem with tar in my busybox that need to be fixed Jan 10 23:20:06 I would build a non-bb tar Jan 10 23:20:25 We had problems w that as well Jan 10 23:32:51 flo_lap, i upgraded tar and nothing changed Jan 10 23:33:01 flo_lap, mabe i need to reboot? Jan 10 23:33:22 GNUtoo: hmm no, I don't think that makes a difference Jan 10 23:33:48 flo_lap, mabe i need to update ipkg as ar and tar works on the device Jan 10 23:34:42 GNUtoo: yes right, that's a good idea... Jan 10 23:34:43 yes Jan 10 23:34:49 it is! Jan 10 23:35:39 GNUtoo: wow... you have an ancient ipkg - a loong time ago there was one that expected the ipk files to be tar archives instead of ar! Jan 10 23:36:00 that must be that Jan 10 23:37:19 flo_lap, bye i must go Jan 10 23:45:16 n8 Jan 10 23:55:37 zecke: gn **** ENDING LOGGING AT Fri Jan 11 02:59:56 2008