| 00:01.27 | _Psybeers | ;) |
| 00:08.39 | *** join/#openembedded _stigger (~jeff@p110-n1-gos.nlc.net.au) |
| 00:13.27 | masticator | kergoth: laproaig is really peaty boggy stuff |
| 00:13.35 | masticator | tastes like a friggin swamp |
| 00:13.49 | masticator | some people like it |
| 00:13.52 | masticator | not my style |
| 00:19.45 | ljp | swamp |
| 00:21.22 | ljp | my head feels like a swamp |
| 00:21.58 | pb_ | that's unfortunate. |
| 00:27.49 | *** join/#openembedded mithro (~tim@dsl1-83.gw1.adl1.airnet.com.au) |
| 00:28.52 | *** join/#openembedded ictyl (~bradford@public-7.rti.org) |
| 00:47.06 | ljp | http://www.cnn.com/2003/TECH/science/12/11/prime.number.ap/index.html |
| 00:55.19 | *** join/#openembedded stigger (~jeff@p110-n1-gos.nlc.net.au) |
| 01:02.44 | treke | wtf All the sudden my fonts look more antialiased |
| 01:05.55 | *** join/#openembedded sdiop (~kobby@rameau-3-82-66-68-6.fbx.proxad.net) |
| 01:12.35 | ljp | hmm |
| 01:12.41 | ljp | maybe you are just drunk |
| 01:12.58 | treke | not yet. |
| 01:16.31 | ljp | man I was tanked last night |
| 01:16.50 | *** join/#openembedded gawel ([Sb+HPYvwk@nowhere.tetrapyloctomy.org) |
| 01:18.33 | chuckr | any more 760 images? |
| 01:22.25 | *** part/#openembedded ictyl (~bradford@public-7.rti.org) |
| 01:25.27 | treke | no |
| 01:49.30 | ljp | a760? |
| 01:52.26 | chuckr | I should spend more time on the building. I'll go back to it this weekend, but now I'm doing build issues for work |
| 01:52.44 | W8TVI | w00t! ircing on this display, and watching a DVD on the other |
| 01:59.31 | Titeuf | bn |
| 02:27.59 | *** join/#openembedded Bala (~Bala@hsdbsk142-165-117-86.sasknet.sk.ca) |
| 02:39.29 | *** join/#openembedded chouimat|sick (~chouimat@r2351064.cidc.net) |
| 02:39.35 | *** join/#openembedded mmelbert|laptop (~mmelbert@h-141-153-101-53.ckb.meer.net) |
| 02:46.28 | *** join/#openembedded treke|laptop (~ggilbert@lsanca2-ar29-4-41-065-242.lsanca2.elnk.dsl.genuity.net) |
| 03:16.59 | *** join/#openembedded gawel ([Hqn2xRLc8@nowhere.tetrapyloctomy.org) |
| 03:25.12 | chouimat|Zzzz | night |
| 04:11.29 | *** join/#openembedded NomNes (~nn@p213.54.205.176.tisdip.tiscali.de) |
| 04:12.15 | NomNes | can ipks of oe be used on zaurus with cacko Linux? |
| 04:15.17 | treke|home | ask in #openzaurus |
| 04:17.23 | *** join/#openembedded bisho (~bisho@136.Red-213-97-191.pooles.rima-tde.net) |
| 04:35.29 | *** join/#openembedded KeyserSoze (~ksoze@64.235.142.48) |
| 05:07.18 | mithro | treke|home: did you ever release/do a 2.95.x image for c7x0? |
| 05:08.33 | treke|home | Mickey did |
| 05:08.38 | treke|home | it's on the server |
| 05:08.52 | treke|home | I did one, but never officially released it |
| 05:09.12 | treke|home | actually, now that I think of it, that could be myimage on the site since I pointed mickeyl at it :p |
| 05:10.42 | mithro | :) |
| 05:16.58 | *** join/#openembedded Disconnect (~dis@arent.you.special.you.gotontheinter.net) |
| 05:17.04 | *** join/#openembedded gawel ([Levm0B4Qg@nowhere.tetrapyloctomy.org) |
| 06:14.40 | treke|home | btw, I got the compaq package working |
| 06:15.57 | treke|home | cool. my desk looks like it's occupied by a human again |
| 06:19.05 | ljp | and it was occupied by a borg before? |
| 06:19.30 | treke|home | borg are more organized |
| 06:19.37 | treke|home | more like trolls |
| 06:21.42 | ljp | hmm |
| 06:27.19 | *** join/#openembedded gawel ([RIQc1ONnO@nowhere.tetrapyloctomy.org) |
| 07:54.05 | W8TVI | hmmm |
| 07:54.42 | W8TVI | must of jared loose my usb 2.0 pci card the last time I hit the computer case with my knee |
| 07:54.54 | W8TVI | its not showing up in lspci |
| 07:58.20 | *** join/#openembedded lsmith (~vandal@p213.54.241.173.tisdip.tiscali.de) |
| 08:27.26 | *** join/#openembedded gawel ([+c0nvyDV6@nowhere.tetrapyloctomy.org) |
| 09:18.52 | *** join/#openembedded lsmith (~vandal@port-212-202-39-121.reverse.qsc.de) |
| 09:26.35 | *** join/#openembedded Sherlock (jdh41@student.cusu.cam.ac.uk) |
| 09:29.11 | *** join/#openembedded BadCode (~sumwid@dsl-082-082-166-036.arcor-ip.net) |
| 09:56.46 | *** join/#openembedded LinuxDoc_ (~chrissy@pD9E01AAA.dip.t-dialin.net) |
| 10:27.38 | *** join/#openembedded gawel ([GNfQ6rPna@nowhere.tetrapyloctomy.org) |
| 10:34.56 | *** join/#openembedded bisho (~bisho@diez.eurielec.etsit.upm.es) |
| 12:19.49 | *** join/#openembedded ar_work (~ar_work@mail.iks-web.de) |
| 12:27.51 | *** join/#openembedded gawel ([ZhlLAYW+8@nowhere.tetrapyloctomy.org) |
| 12:29.42 | *** part/#openembedded ar_work (~ar_work@mail.iks-web.de) |
| 12:44.40 | *** join/#openembedded W8TVI (~W8TVI@userd90.torchlake.com) |
| 12:44.45 | CIA | repository=03oe user=11mickeyl@gandalf.tm.informatik.uni-frankfurt.de |
| 12:44.46 | CIA | 07ChangeSet@1.324 start a FAQ - this will grow eventually |
| 13:11.14 | *** join/#openembedded Toi_ (~pleemans@oia-gw1.be.jnj.com) |
| 14:06.50 | *** join/#openembedded ToyKeeper (gimkra@c-24-9-95-183.client.comcast.net) |
| 14:28.03 | *** join/#openembedded gawel ([siBXRNWsi@nowhere.tetrapyloctomy.org) |
| 14:33.17 | *** mode/#openembedded [+o chouimat] by ChanServ |
| 14:33.39 | chouimat | morning |
| 14:38.13 | *** join/#openembedded pb_ (~pb@dsl-62-3-66-204.zen.co.uk) |
| 14:41.48 | *** join/#openembedded schrd (~schrd@froeschi.csn.tu-chemnitz.de) |
| 14:42.08 | schrd | hi |
| 14:42.39 | schrd | is it possible to flash a zaurus from a cf with fat32 filesystem on it or must it be vfat 16? |
| 14:44.17 | schrd | oops, sorry, wrong chan |
| 14:44.23 | *** part/#openembedded schrd (~schrd@froeschi.csn.tu-chemnitz.de) |
| 15:00.20 | *** join/#openembedded Virusmaster (~Virus@pD9E045EB.dip.t-dialin.net) |
| 15:05.40 | *** join/#openembedded xai (~pasta@user-0vvdbe1.cable.mindspring.com) |
| 15:06.48 | xai | Do i need an initrd to get the rootfs into ramfs? Although I didnt compile tmpfs ( i didnt see the option) into the kerel, my /proc/filesystems shows a tmpfs and ramfs support. |
| 15:07.27 | xai | My kernel is monolithic, and I think tmpfs is really /dev/hda2.. |
| 15:08.13 | Twiun | xai: more details please, kernel? which kernel? using OpenEmbedded, OpenZaurus, Familiar, something else? |
| 15:11.40 | xai | 2.4.23 kernel.. using a debian embeded system i created .. The lilo i used couldnt boot unless i passed root=/dev/hda2 to it, but I think that mounts the actual partition .. |
| 15:12.04 | xai | It still looks like tmpfs in df and mount: |
| 15:12.48 | xai | I'd like to get this to work w/o an initrd file.. So I guess I just want readonly / |
| 15:13.32 | Twiun | ahh, I haven't worked in that config before - but I think others on the chan have |
| 15:13.50 | xai | I'm somewhat unfamiliar with details of booting and / mounts... I still think what this reports as tmpfs is not tmpfs.. because the size is identical to the physical parition. Im using disk-on-module. |
| 15:16.06 | xai | It would not boot with root=ram0 in lilo.conf, I'm guessing that ram0 is where an initrd would have put it. |
| 15:20.23 | *** join/#openembedded bach (~ciaolinux@MSTU83.CIMS.NYU.EDU) |
| 15:34.07 | xai | Do i need cramfs support in the kernel to use a cramfs initrd file? |
| 15:34.35 | Twiun | that would make sense |
| 15:36.39 | xai | So all kernels that use initrd have cramfs support? (assuming the initrd is cramfs format). |
| 15:37.40 | xai | I was thinking the initrd had its own fs suppor inside it so all it needs to do is expand in memory, and then copy that to the ram device. |
| 15:39.57 | Twiun | you'd still need code to do the actual expanding |
| 15:40.15 | Twiun | and afaik cramfs isn't too big |
| 15:40.29 | xai | I just forgot to put it in my kernel.. |
| 15:41.04 | Twiun | hi drw |
| 15:41.20 | drw | hey Twiun |
| 15:45.00 | gb2 | initrd can be other filesystem types too |
| 15:45.24 | CosmicPenguin | It can be whatever you want |
| 15:47.22 | *** join/#openembedded bisho (~bisho@138.100.21.68) |
| 15:59.14 | xai | heh. i don't even have tmpfs support, but the kernel reports tmpfs support in /proc/filesystems. |
| 15:59.32 | *** join/#openembedded FPCat (~FPCat@CPE000625760564-CM013010001639.cpe.net.cable.rogers.com) |
| 16:00.17 | xai | probably becuase they threw in an mtab that states tmpfs.. confusing. |
| 16:03.00 | xai | can I use /dev/hdc2 as a root device at boot, but later have it go to tmpfs or cramfs? |
| 16:03.12 | xai | i.e. without having initrd? |
| 16:04.28 | xai | Is there a good generic doc or howto on booting aspect of embedded? |
| 16:07.31 | CosmicPenguin | booting an embedded device is the same as booting an regular devcie |
| 16:10.42 | gb2 | xai: you can't fully disable tmpfs support |
| 16:10.51 | gb2 | xai: the kernel uses it internally |
| 16:12.28 | xai | ah.. ok.. I'm trying to understand this. I just dont know what to do to get this to not use the Disk-on-module.. But maybe it is working right .. I just dont know. |
| 16:12.49 | xai | /etc/mtab claims to be using tmpfs, but the size is much larget than my available ram. |
| 16:13.41 | xai | I removed the 'append="hda=flash" ' stuff, but maybe I should not. |
| 16:13.48 | xai | that was out of lilo. |
| 16:14.28 | xai | I see this in lilo.conf: |
| 16:14.34 | xai | geometric |
| 16:14.41 | xai | boot=/dev/hda |
| 16:14.52 | xai | root=/dev/ram0 |
| 16:15.06 | xai | ... normal stuff .... |
| 16:15.29 | xai | And then a clause for my kernel without any initrd (which i removed since i have no modules). |
| 16:16.34 | xai | But that image won't boot unless i pass "root=/dev/hda2" to the bootline... |
| 16:16.55 | xai | Sorry to barf in this channel. |
| 16:19.48 | xai | I get this: VSS: Cannot open root device "100" or 01:00 |
| 16:25.48 | *** join/#openembedded FPCat (~FPCat@CPE000625760564-CM013010001639.cpe.net.cable.rogers.com) |
| 16:28.13 | *** join/#openembedded gawel ([BLVwn502X@nowhere.tetrapyloctomy.org) |
| 16:47.56 | kergoth | morning |
| 16:55.37 | kergoth | ibot: nslookup handhelds.org |
| 16:57.55 | *** join/#openembedded TimRiker (~timr@proxyle01.ext.ti.com) |
| 17:03.50 | treke | kergoth: heard anything from hh.o? |
| 17:04.24 | *** join/#openembedded gawel ([3DNWm9NtS@nowhere.tetrapyloctomy.org) |
| 17:04.26 | kergoth | treke: they okayed it, but our vhost is commented out in apache's config |
| 17:04.30 | kergoth | heh |
| 17:04.46 | treke | hehe |
| 17:04.49 | kergoth | need to pester em to uncomment it, then confirm its happy and start moving content over |
| 17:05.32 | treke | cool |
| 17:05.41 | W8TVI | oh? |
| 17:05.55 | gawel | has anybody been able to build busybox ipk under oe successfully? my build is fine until i try to create ipk out of the build (on RH9 system with python 2.3.2) |
| 17:06.28 | treke | found that I can fit some random one I bought now that I lost a part :) |
| 17:07.33 | treke | I ought to open the thing back up and see if I can figure out where the part goes |
| 17:07.54 | kergoth | gawel: did you build ipkg-utils-native? |
| 17:08.11 | kergoth | gawel: its in the deps.. oebuild showdata busybox/busybox_1.00-pre2.oe|grep DEPENDS.*= |
| 17:08.46 | gawel | shows only libc and patcher 8> |
| 17:09.05 | gawel | DEPENDS="patcher virtual/libc" |
| 17:09.19 | gawel | did a pull on package repository this morn |
| 17:09.47 | kergoth | ah, our package_ipk oeclass isnt adding it to deps for some reason |
| 17:09.51 | kergoth | its commented out in the oeclass |
| 17:17.00 | kergoth | oh, probably due to recursive deps. ipkg-utils-native depends on auto*, but if we set that dep, auto* will end up depending on ipkg-utils-native |
| 17:17.04 | kergoth | heh |
| 17:17.27 | kergoth | guess we need our native and cross classes to actually _remove_ the packaging tasks, rather than just not packaging any files |
| 17:18.08 | kergoth | err no, itd have to explicitly remove package_* from INHERIT.. dont think thats easily done atm |
| 17:21.39 | gawel | is the stuff (where stuff=packages being transitioned from buildroot to oe) you are working in still located in nonworking? |
| 17:25.59 | kergoth | yeah. obviously there are things that havent been started yet, but those are the ones that are at least in progress |
| 17:28.36 | gawel | cool. i've started working on migrating of difficult packages like grep... ;> |
| 17:28.52 | kergoth | hehe. thats cool |
| 17:29.10 | kergoth | you'll notice i added a bunch of the gnu equivalents to busybox commands, mostly since GNU apps tend to be quick to add |
| 17:29.41 | gawel | yeah |
| 17:30.41 | kergoth | mallum pushed the freetype and fontconfig .oe's :) |
| 17:33.16 | gawel | so if i actually build something that works should i just send you a patch, or will you get me write access to packages repository? (mickey was of opinion that folks should email patches first to make sure that folks don't break repository 8> |
| 17:34.33 | kergoth | mickey's correct. first send a patch to openembedded-devel |
| 17:34.44 | kergoth | after a patch or two, then we'll see |
| 17:35.07 | gawel | ah ok. openembedded-devel it is |
| 17:38.08 | kergoth | ugh, i'm going to end up making glibc a 3 pass build |
| 17:38.10 | kergoth | how annoying |
| 17:39.48 | pb_ | oh dear. why's that? |
| 17:39.59 | kergoth | pb_: for the toolchain build anyway |
| 17:40.06 | kergoth | first pass is the headers, saves having to inhibit libc |
| 17:40.18 | kergoth | last pass is only needed if i want to build nptl |
| 17:40.24 | pb_ | ah, I see. |
| 17:40.25 | kergoth | since iirc it wont build with the first pass gcc, only last |
| 17:40.36 | Disconnect | ugh. so if I do a flash, then an erase-all, it should wipe flash, right? so I shouldn't be seeing all these leftover bits? |
| 17:41.16 | kergoth | hehe |
| 17:41.24 | kergoth | in theory you shouldnt.. |
| 17:41.44 | Disconnect | I saw other people reporting that sort of thing, figured it was a fluke (or a lack of clear/delete-all) .. |
| 17:43.47 | Disconnect | well.. umount /mnt/card;umount /mnt/cf first ;) |
| 17:48.28 | Disconnect | ibot: updater.sh |
| 17:48.32 | | Disconnect: KCI error, or a problem with the Keyboard-Chair Interface. |
| 17:51.49 | Disconnect | ibot: updater.sh is the shell script used to reflash a zaurus. its "encrypted" |
| 17:51.51 | | Disconnect: okay |
| 17:51.55 | Disconnect | ibot: updater.sh is also http://cacko.biz/~sash/slc7x0/encsh.c |
| 17:51.56 | | Disconnect: okay |
| 17:54.51 | Disconnect | ibot: updater.sh |
| 17:54.54 | | from memory, updater.sh is the shell script used to reflash a zaurus. its "encrypted", or http://cacko.biz/~sash/slc7x0/encsh.c |
| 18:00.03 | Disconnect | btw rm -rf / and then reflash did bad things. |
| 18:29.52 | *** join/#openembedded darkschneider (~gab@213.140.22.64) |
| 18:38.49 | kergoth | Twiun[away]: you end up picking up an mm10? |
| 18:49.19 | kergoth | gb2: fyi, a glibc toolchain build fails in like 3 ways, depending on how many fixes i push |
| 18:50.03 | *** join/#openembedded gortsleigh (~chatzilla@64.254.123.107) |
| 18:52.33 | *** join/#openembedded andersee (~andersee@codepoet.org) |
| 18:53.40 | gb2 | kergoth: glibc toolchain is in need of fixups |
| 18:53.46 | kergoth | i noticed |
| 18:53.46 | gb2 | i know |
| 18:53.49 | kergoth | glibc requires libgcc_eh |
| 18:53.53 | kergoth | which doesnt exist until libgcc_s gets built |
| 18:54.08 | kergoth | have to symlink the static libgcc to it as a workaround, then remove the link before the final pass gcc build |
| 18:54.11 | kergoth | heh |
| 18:54.18 | kergoth | or do it in the glibc build |
| 18:54.18 | gb2 | yuck |
| 18:54.20 | kergoth | yeah |
| 18:54.24 | kergoth | ugly |
| 18:54.31 | gb2 | does it produce a viable glibc? |
| 18:54.40 | kergoth | then if we want those glibc apps to actually be linked against the shared libgcc, we'd have to build glibc again after the final gcc |
| 18:54.49 | kergoth | picture 3 pass glibc. blech. |
| 18:54.52 | kergoth | yes, its usable |
| 18:55.01 | gb2 | hm. will it work if the gcc-initial builds a shared libgcc? |
| 18:55.21 | kergoth | can it? i have a feeling thatd fail due to dependency on glibc |
| 18:55.33 | kergoth | shared will fail, _eh isnt shared, but isnt built until _s is |
| 18:55.34 | gb2 | why does it depend on glibc? |
| 18:55.43 | kergoth | hell if i know, i just remember it failing when i tried |
| 18:55.46 | gb2 | hrm |
| 18:56.01 | kergoth | feel free to play with it, i'll push a workaround for now |
| 18:56.29 | gb2 | ok. |
| 18:56.34 | gb2 | i'll look tonight probably |
| 18:56.49 | kergoth | mallum pushed freetype :) wonder how many more items we need for a full x build |
| 18:57.49 | gb2 | heh |
| 18:58.12 | kergoth | hmm |
| 18:58.20 | kergoth | the glibc toolchain build still fails |
| 18:58.21 | kergoth | the 2nd pass gcc explodes |
| 18:58.22 | gb2 | kergoth: i was also thinking i may take a stab at augmenting the graph type to do loop detection in walkup/down |
| 18:58.27 | kergoth | thatd be cool |
| 19:04.51 | *** join/#openembedded pb_ (~pb@2002:5160:45ef:0:240:95ff:fe30:aa84) |
| 19:07.32 | kergoth | pb_: re |
| 19:07.42 | pb_ | kergoth: re |
| 19:10.04 | kergoth | hm, why do people use popt instead of getopt, i wonder |
| 19:11.02 | pb_ | Dunno. I'm not sure why Red Hat even wrote popt in the first place. |
| 19:11.16 | kergoth | weird |
| 19:12.36 | gb2 | getopt wasn't flexible enough |
| 19:13.13 | gb2 | popt lets you extend stuff with popt config files for the apps |
| 19:13.21 | gb2 | a lot of rpm's options are implemented this way |
| 19:14.40 | kergoth | ah |
| 19:16.41 | gb2 | mostly they're stuff like turning: rpm -qp --scripts into the appropriate rpm -qp --qf '...' command |
| 19:16.46 | gb2 | qf == query format |
| 19:19.01 | gb2 | kergoth: hrm.. yeah, --enable-shared for gcc requires the startfiles |
| 19:19.07 | gb2 | which are part of libc |
| 19:19.11 | kergoth | yeah, thats it |
| 19:19.26 | gb2 | wonder how glibc is supposed to be bootstrapped then |
| 19:19.39 | kergoth | i found a ppc toolchain build steps on a site that recommended the symlink hack |
| 19:19.45 | kergoth | it was like a fucking 5 pass build |
| 19:19.57 | gb2 | i ran into something that said it was a test program in the glibc build |
| 19:20.17 | kergoth | yeah i'm sure it is |
| 19:20.25 | gb2 | hrm. |
| 19:20.31 | kergoth | maybe we'd be best off building just hte core glibc lib in pass 2, and add a 3rd glibc build pass. |
| 19:20.36 | gb2 | more reasons to move to uclibc |
| 19:20.40 | kergoth | hehe |
| 19:21.38 | gb2 | if it is infact a test program, why's it being built while cross compiling? |
| 19:22.04 | kergoth | tests arent always runtime. |
| 19:22.09 | pb_ | we build the tests so that you can run them on the target. |
| 19:22.16 | gb2 | i guess yeah |
| 19:22.19 | kergoth | ah |
| 19:22.29 | pb_ | I think it should only happen if you "make check", though. |
| 19:24.28 | *** join/#openembedded bipolar (~bflong@dhcp-103-630-1.msns.sm.ptd.net) |
| 19:24.39 | kergoth | hey bipolar |
| 19:24.43 | gb2 | wonder if you can do: glibc-headers, gcc-static, glibc-startfiles, gcc-shared, glibc-final, gcc-full |
| 19:26.36 | gb2 | kergoth: if you edit glibc's configure to not add -lgcc_eh, does it break? |
| 19:29.53 | gb2 | I wonder if uclibc gets built differently if it's got a gcc with libgcc_s/_eh available to it during build |
| 19:30.40 | kergoth | well, itd work if we had it -lgcc.. i dont know what the diff is between gcc and gcc_eh |
| 19:30.55 | kergoth | pb_: do you know what the difference is? between libgcc.a and libgcc_eh? |
| 19:32.17 | pb_ | I think libgcc_eh is exception handling code. I'm not sure exactly why it's separate, but iirc you generally link with both. |
| 19:32.26 | pb_ | Might have something to do with shared libgcc_s.so. |
| 19:32.26 | kergoth | ahh okay |
| 19:32.40 | kergoth | whatever in glibc that needs it appears to build fine without it, seemingly |
| 19:32.56 | pb_ | which platform? |
| 19:33.05 | kergoth | arm-linux |
| 19:33.12 | pb_ | I don't think the arm glibc needs any eh support from gcc, but others might differ. |
| 19:33.14 | gb2 | kergoth: the reasoning behind it has to do with multilib |
| 19:33.41 | pb_ | I think sysdeps/generic/framestate.c is the culprit. |
| 19:34.03 | bipolar | kergoth: yo |
| 19:34.21 | pb_ | ah, yes. |
| 19:34.26 | kergoth | i think we're touching an empty arm framestate.c. is that needed? |
| 19:34.26 | pb_ | <PROTECTED> |
| 19:34.26 | pb_ | <PROTECTED> |
| 19:34.26 | pb_ | <PROTECTED> |
| 19:34.28 | kergoth | bipolar: how goes? |
| 19:34.39 | pb_ | kergoth: yes. the generic framestate.c won't compile on arm. |
| 19:34.51 | kergoth | k |
| 19:35.08 | pb_ | it's dependent on dwarf2-based eh, and arm is still on sjlj. |
| 19:35.23 | kergoth | ah |
| 19:36.57 | gb2 | kergoth: ok.. shared libgcc results in libgcc.a libgcc_s.so libgcc_eh.a, static ends up with the symbols from libgcc_eh in libgcc.a |
| 19:37.21 | kergoth | yep, which explains why we can just symlink it over to libgcc.a and have it work |
| 19:37.42 | gb2 | or create an empty archive and it'd work too |
| 19:37.49 | kergoth | yep |
| 19:38.08 | gb2 | make the gcc-initial-cross .oe create it as an empty archive |
| 19:38.10 | *** join/#openembedded gawel ([5TObPv65P@nowhere.tetrapyloctomy.org) |
| 19:38.18 | gb2 | when the gcc-cross builds, do_stage will overwrite it |
| 19:39.11 | gb2 | that glibc should be viable, its tools will have a static libgcc instead of the dynamic libgcc_s |
| 19:39.31 | gb2 | so a third pass build of glibc would result in more-dynamic tools |
| 19:39.35 | gb2 | but otherwise ought to be the same |
| 19:40.20 | gb2 | make sense? |
| 19:41.16 | kergoth | right |
| 19:41.25 | kergoth | thats basically what i have here, just using a symlink instead of an empty archive |
| 19:42.44 | gb2 | empty archive is probably safer |
| 19:43.04 | kergoth | yep, dont have to bother rm -f'ing the link and stuff |
| 19:43.11 | kergoth | ill change that and push it |
| 19:43.16 | gb2 | or create a hard link rather than a symlink |
| 19:43.23 | gb2 | well |
| 19:43.24 | gb2 | hrm |
| 19:44.23 | gb2 | nothing will ever want -lgcc_eh without -lgcc |
| 19:45.04 | kergoth | why didnt they just make gcc_eh explicit linking mandatory and build it when building libgcc.a as well? |
| 19:45.08 | kergoth | heh |
| 19:46.00 | *** join/#openembedded qip|Virus|da (~Virus@pD9E65ECD.dip.t-dialin.net) |
| 19:46.08 | gb2 | kergoth: yeah, just always do the split |
| 19:46.34 | kergoth | yep |
| 19:46.54 | pb_ | kergoth: well, nobody is meant to know about -lgcc_eh _or_ -lgcc, anyway. |
| 19:47.31 | gb2 | pb: yeah.. the gcc specs are supposed to handle that for you |
| 19:47.37 | kergoth | heh. well still, why isnt it split all the time |
| 19:48.08 | ljp | oohhh. secret linking, hmmm? |
| 19:48.08 | gb2 | kergoth: did your glibc build work fine if you did this trick? |
| 19:48.10 | gb2 | you said something about a problem before. |
| 19:48.24 | kergoth | it gets through the gcc-cross, then fails there |
| 19:48.31 | gb2 | with what error? |
| 19:48.35 | kergoth | the xgcc fails to find ld-linux.so.2 for soem reason |
| 19:48.40 | pb_ | gb2: yup, exactly. |
| 19:48.47 | kergoth | almost like /lib isnt ending up in its rpath or something |
| 19:48.48 | gb2 | kergoth: now that's weird. |
| 19:49.03 | kergoth | yeah. now get this |
| 19:49.12 | kergoth | if i build my chain using my other tree of .oe's, life is good |
| 19:49.23 | kergoth | has slightly differing configure options for gcc, both passes |
| 19:49.29 | gb2 | what changes? |
| 19:49.38 | kergoth | i dont recall offhand, i'll diff |
| 19:49.42 | gb2 | did i break a glibc thing when i tweaked them? |
| 19:49.46 | kergoth | its using the options used in the crosstool shell script |
| 19:49.52 | kergoth | dunno |
| 19:57.52 | *** join/#openembedded MK500 (~anderex@sys7.cthub.com) |
| 20:03.50 | *** join/#openembedded MK500 (~MK500@sys7.cthub.com) |
| 20:04.51 | *** join/#openembedded MK500 (~MK500@sys7.cthub.com) |
| 20:18.25 | *** join/#openembedded edhodges (~edhodges@modem-3798.gorilla.dialup.pol.co.uk) |
| 20:22.12 | *** join/#openembedded gawel ([HspOILED+@nowhere.tetrapyloctomy.org) |
| 20:25.11 | *** join/#openembedded ar_oszine (ar@port-ip-213-211-231-12.reverse.mdcc-fun.de) |
| 20:25.17 | ar_oszine | hi |
| 20:25.24 | kergoth | hey |
| 20:25.47 | ar_oszine | i have trouble to build patcher. oebuild can't find oe.conf |
| 20:27.00 | kergoth | then you didnt set OEDIR to the oe dir |
| 20:27.10 | kergoth | most likely anyway |
| 20:27.15 | ar_oszine | oedir is correct |
| 20:27.30 | kergoth | ls $OEDIR/conf/oe.conf |
| 20:28.34 | bipolar | I'm an hour late, but, it's going... |
| 20:28.42 | kergoth | heh |
| 20:29.11 | ar_oszine | ls works. |
| 20:29.33 | ar_oszine | here comes parts of strace... |
| 20:29.39 | kergoth | strace wont help |
| 20:29.39 | ar_oszine | access("./conf/oe.conf", R_OK) = -1 ENOENT (No such file or directory) |
| 20:29.39 | ar_oszine | access("/mnt/data/opie/oe/packages/conf/oe.conf", R_OK) = -1 ENOENT (No such file or directory) |
| 20:29.40 | ar_oszine | access("/home/ar/.oe/conf/oe.conf", R_OK) = -1 ENOENT (No such file or directory) |
| 20:29.40 | ar_oszine | access("/usr/share/oe/bin/conf/oe.conf", R_OK) = -1 ENOENT (No such file or directory) |
| 20:29.41 | ar_oszine | access("/usr/share/oe/conf/oe.conf", R_OK) = -1 ENOENT (No such file or directory) |
| 20:29.42 | ar_oszine | access("/usr/share/oe/conf/oe.conf", R_OK) = -1 ENOENT (No such file or directory) |
| 20:30.12 | ar_oszine | hm... |
| 20:30.15 | kergoth | and whats OEDIR set to? |
| 20:30.15 | gb2 | what is your OEDIR set to? |
| 20:30.18 | gb2 | echo $OEDIR |
| 20:30.20 | kergoth | did you override OEPATH? |
| 20:30.21 | gb2 | it can't be . |
| 20:30.42 | ar_oszine | echo $OEDIR |
| 20:30.49 | gawel | heh |
| 20:31.00 | gb2 | ar: type that in a shell, not irc :P |
| 20:31.04 | ar_oszine | mom ;) echo $OEDIR /opt/opie/oe/oe |
| 20:31.11 | ar_oszine | it set correct. |
| 20:31.17 | gb2 | did you export it? |
| 20:31.17 | ar_oszine | s/it/is |
| 20:31.24 | ar_oszine | yes |
| 20:31.39 | gb2 | env |grep OE |
| 20:32.08 | gb2 | what's that give you? |
| 20:32.08 | ar_oszine | i have read the wiki: http://openembedded.org/phpwiki/index.php/GettingStarted |
| 20:32.37 | ar_oszine | gb2: nothing. |
| 20:32.42 | ar_oszine | env is empty? |
| 20:32.45 | gb2 | you don't have OEDIR exported then |
| 20:32.54 | gb2 | is this tcsh or bash? |
| 20:33.02 | ar_oszine | bash |
| 20:33.02 | gb2 | which shell do you use |
| 20:33.08 | gb2 | export OEDIR |
| 20:33.10 | gb2 | do that |
| 20:33.46 | gb2 | then try the env |grep OE command again |
| 20:33.48 | ar_oszine | great. it's work. |
| 20:33.54 | ar_oszine | thanks :) |
| 20:34.00 | kergoth | gb2: good call |
| 20:34.34 | gawel | btw, if i do oebuild clean on any of the native tools, should stuff from packages/tmp/staging/build be removed? |
| 20:34.53 | kergoth | nope |
| 20:35.00 | kergoth | staging isnt ever cleaned unless you wipe it manually |
| 20:35.11 | kergoth | (thatll also get fixed when we split staging. on the todo) |
| 20:35.36 | gawel | ah ok. cool. that exaplains a lot 8> |
| 20:35.50 | kergoth | rm -rf tmp/staging tmp/stamps/*stage*; may be useful |
| 20:35.52 | kergoth | :) |
| 20:35.58 | kergoth | then repopulate |
| 20:39.06 | gb2 | kergoth: i updated the wiki so exporting is a little more obvious |
| 20:39.10 | gb2 | OEDIR=/path/to/oe/dir |
| 20:39.10 | gb2 | OEFILES=*/*.oe |
| 20:39.10 | gb2 | export OEDIR OEFILES |
| 20:39.10 | gb2 | PATH=$PATH:$OEDIR/bin |
| 20:39.13 | gb2 | separate step |
| 20:39.45 | kergoth | cool |
| 20:39.46 | gb2 | also added tcsh instructions for those people |
| 20:39.57 | gb2 | i had a sudden urge to be helpful, weird. |
| 20:40.01 | kergoth | hehe |
| 20:41.18 | kergoth | i need to patch a version of aclocal that also drops all the component .m4's that supplied a macro in the generated aclocal.m4, into a sepecified dir |
| 20:41.23 | kergoth | heh |
| 20:41.41 | gawel | gb: how about a wiki page that would describe how to build a toolchain 8> |
| 20:42.12 | gb2 | gawel: once we figure out how to build toolchains reliably. |
| 20:42.13 | gb2 | ;) |
| 20:42.21 | kergoth | doh, thats the problem |
| 20:42.30 | kergoth | freetype2 was relying on the old method of config.h generation |
| 20:42.32 | kergoth | hmm |
| 20:42.51 | gawel | *har* 8> must be dreaded libgcc_eh problem, right? |
| 20:43.10 | kergoth | oh, the autoheader call is blowing away the ftconfig.in .. |
| 20:43.13 | kergoth | doh |
| 20:43.58 | gb2 | gawel: one of the main problems is toolchains don't like to coexist at the moment, but oemake isn't smart enough to avoid building both. |
| 20:44.03 | gb2 | this is very close to being solved |
| 20:44.21 | gawel | cool |
| 20:46.23 | *** join/#openembedded andersee (~andersee@codepoet.org) |
| 20:49.23 | gawel | btw, when i override do_install method inherited from autotools for a package, am i in directory tmp/base/<packagename>-ro/<packagename> at the time of do_intall being called? |
| 20:49.26 | Disconnect | anyone know where cacko got the w100 driver for mplayer? |
| 20:50.51 | kergoth | gawel: -ro? |
| 20:51.10 | kergoth | you're in ${S} or, optionally, ${B} if its defined |
| 20:51.19 | kergoth | ${S} is ${WORKDIR}/${PN}-${PV}-${PR} |
| 20:51.25 | kergoth | er |
| 20:51.29 | gawel | ah, i meant r0 |
| 20:51.29 | kergoth | s/-${PR}// |
| 20:51.33 | kergoth | r0 is revision |
| 20:51.43 | gb2 | ${S} is the "source dir" which is your unpacked source |
| 20:51.49 | kergoth | workdir is ${TMPDIR}/${CATEGORY}/${PN}-${PV}-${PR} |
| 20:51.57 | kergoth | default S is ${WORKDIR}/${PN}-${PV}, but that varies |
| 20:52.06 | kergoth | just read oe.conf |
| 20:52.13 | gb2 | ${B} if you define it is a "build dir" for doing objections separated from the source. often toolchains are built this way |
| 20:52.26 | kergoth | ugh, freetype is a bitch |
| 20:52.54 | gawel | ker: thanks |
| 20:52.58 | kergoth | cant call autoheader against it, it blows up |
| 20:53.07 | kergoth | guess i'd better override do_configure |
| 21:01.24 | kergoth | woot, got it. |
| 22:49.40 | *** join/#openembedded Quantum (~Quantum__@67.104.83.11.ptr.us.xo.net) |
| 22:58.32 | *** join/#openembedded gawel ([KnyJ6a25g@nowhere.tetrapyloctomy.org) |
| 23:05.44 | *** join/#openembedded lsmith (~vandal@p213.54.241.173.tisdip.tiscali.de) |
| 23:14.18 | *** join/#openembedded KSoze (~ksoze@64.235.142.48) |
| 23:27.01 | *** join/#openembedded drw (~drw@c-24-0-141-17.client.comcast.net) |
| 23:29.34 | *** join/#openembedded stigger (~jeff@p079-n1-gos.nlc.net.au) |
| 23:35.41 | *** join/#openembedded SozeK (~ksoze@64.235.142.48) |