=== dyfet__ is now known as dyfet | ||
lbt | can I get prelink for Ubuntu arm? | 13:49 |
---|---|---|
lbt | http://packages.ubuntu.com/jaunty/prelink says not ... just wondering why? | 13:49 |
ogra | packages.u.c is really no reliable source for ports architectures, rather use launchpad | 13:50 |
lbt | ogra: OK... | 14:04 |
ogra | https://launchpad.net/ubuntu/jaunty/+source/prelink | 14:04 |
ogra | but looks like nobody bothered to make it available on armel, you might need to build it yourself | 14:04 |
lbt | it should help chroot builds I think | 14:04 |
Stskeeps | lbt: afaik prelink doesn't work on arm :/ i had limited success with it | 14:25 |
suihkulokki | it does | 14:25 |
suihkulokki | the stock n8x0 images are all prelinked | 14:25 |
Stskeeps | hmm - i thought that was in the past | 14:26 |
lbt | do binaries need to be compiled/linked in any special way? | 14:32 |
ogra | no | 14:33 |
lbt | ah, I read about -fPIC somewhere | 14:33 |
ogra | its used in the default settings everywhere, yes | 14:33 |
lbt | OK - I'm currently building a version of prelink that automatically runs when installed | 14:34 |
Sarvatt | anyone interested in testing pixman compiled with neon and simd runtime detection support on karmic armel? | 14:42 |
Sarvatt | qemu-arm-eabi sure doesn't like rebuilding man-db pages :D | 14:44 |
ogra | yeah | 14:45 |
ogra | bug 401782 might fix that | 14:45 |
ubot4 | Launchpad bug 401782 in qemu "please build a static version of qemu-arm 0.11.x in a separate binary deb" [Wishlist,New] https://launchpad.net/bugs/401782 | 14:45 |
ogra | havent tested it yet | 14:45 |
Sarvatt | oh sweet | 14:45 |
ogra | but the 0.11 upstream eabi additions might be more accurate than the patches i used | 14:45 |
Sarvatt | i need to do this same thing for PPC anyway, should look into just building it myself | 14:46 |
=== dyfet__ is now known as dyfet | ||
Sarvatt | darn | 15:09 |
Sarvatt | yeah pixman wont compile in neon runtime detection support unless your march cflag is armv6+ | 15:10 |
Sarvatt | checking whether to use ARM SIMD assembler... yes | 15:10 |
Sarvatt | checking whether to use ARM NEON... no | 15:10 |
ogra | karmic builds will sonn default to armv6+ | 15:11 |
suihkulokki | Sarvatt: if that is really runtime detection (as opposed to compile time) that should be fixed | 15:14 |
Sarvatt | http://sarvatt.com/downloads/pixman.txt | 15:19 |
Sarvatt | updated that pixman.txt with the config.log results | 15:24 |
Sarvatt | hmm maybe if i add -mfloat-abi=softfp | 15:26 |
Sarvatt | that did it, needed -mfloat-abi=softfp added to CFLAGS | 15:35 |
Sarvatt | checking whether to use ARM SIMD assembler... yes | 15:35 |
Sarvatt | checking whether to use ARM NEON... yes | 15:35 |
=== cbrake_away is now known as cbrake | ||
Sarvatt | here's pixman http://sarvatt.com/downloads/armel/ will upload xserver there when its done building if anyone wants to test it out | 16:40 |
Sarvatt | Get:78 http://ports.ubuntu.com karmic/main libgl1-mesa-dev 7.5~rc4-1ubuntu2 [27.1kB] | 16:40 |
Sarvatt | i see mesa didnt build yet again :( | 16:40 |
ogra | we didnt find a fix for the slowness yet | 16:43 |
ogra | i'm just building it with reverted dpkg but it will likely still take a while until i get results | 16:44 |
ogra | if you used softfp in your build your pixman will likely run pretty slow, did you try vfp ? | 16:44 |
Sarvatt | ohhh | 16:49 |
* Sarvatt cheers | 16:49 | |
Sarvatt | Get:1 http://ports.ubuntu.com karmic/main libgl1-mesa-dev 7.5-1ubuntu1 [27.5kB] | 16:49 |
ogra | oh ? | 16:50 |
ogra | indeed, http://qa.ubuntuwire.com/ftbfs/ doesnt have it ! | 16:51 |
Sarvatt | it used -mfpu=neon -mcpu=cortex-a8 -mfloat-abi=softfp for the neon runtime detection stuff, -mcpu=arm1136j-s -mfloat-abi=softfp for the arm SIMD runtime detection support files | 16:56 |
Sarvatt | it was mfloat-abi using softfp, not like i coulda used mfloat-abi=hard there :D | 17:00 |
NCommander | ogra, we still have an issue w/ qt4-x11 though | 17:02 |
ogra | yes | 17:02 |
NCommander | http://launchpadlibrarian.net/29325326/buildlog_ubuntu-karmic-armel.qt4-x11_4.5.2-0ubuntu2_FAILEDTOBUILD.txt.gz | 17:03 |
NCommander | want to haphazard a guess? | 17:03 |
ogra | thats the same issue upstart had | 17:03 |
NCommander | crap | 17:03 |
NCommander | How'd we fix upstart? | 17:03 |
ogra | lool did it ... i think he fiddled with CFLAGS | 17:04 |
ogra | Session terminated, killing shell...make[1]: *** wait: No child processesmake[2]: *** | 17:04 |
ogra | "Don't build the testsuite with -fPIE on armel" | 17:05 |
NCommander | That isn't going to fly for a shared library | 17:05 |
ogra | thats the change from upstart | 17:05 |
NCommander | ogra, I thought you needed to build w/ PIE though when you build shared libraries in general | 17:07 |
NCommander | (well, PIC, but its still position independent code) | 17:07 |
ogra | well, it was only disabled for the testsuite which is an insane pile of really huge sourcefiles | 17:08 |
NCommander | ogra, I think you just defined Qt :-) | 17:08 |
NCommander | ogra, we're not going to be able to use the same fix unfortunately | 17:08 |
Sarvatt | 0ubuntu1 failed too, eww | 17:11 |
Sarvatt | Build log: buildlog_ubuntu-karmic-armel.qt4-x11_4.5.2-0ubuntu1_FAILEDTOBUILD.txt.gz | 17:11 |
Sarvatt | Finished: 2009-07-15 (took 1 day, 4 hours, 45 minutes, 47.6 seconds) | 17:11 |
Sarvatt | speedy! | 17:11 |
NCommander | yeah | 17:12 |
NCommander | Sarvatt, that's an issue with gcc and -fPIC/-fPIE | 17:12 |
NCommander | (upstart ran into that one) | 17:12 |
NCommander | Sarvatt, I just kicked off a build on our porting box to see if it actually finishes given enough time, or if we need to do more indepth debugging to fix that one | 17:13 |
NCommander | sorry about that, X chose a lovely time to ABEND | 17:15 |
lool | NCommander: for libqtwebkit: reproduce the issue and see where CPU time is spent | 17:16 |
lool | Is it using a lot of memory etc. | 17:16 |
lool | Then try removing some opts | 17:16 |
NCommander | lool, how do I do that per say? | 17:16 |
lool | it's doing a link and using -Wl,-O1; perhaps we need to drop that | 17:16 |
NCommander | oh right | 17:16 |
NCommander | time | 17:16 |
NCommander | lool, I don't get warm feeling from changin qt's config flags, and we don't want to loose -O1 | 17:17 |
lool | ogra: upstream moved the gnome-keyring bug to NEEDINFO now; could you please update them? | 17:17 |
ogra | yes, i have backported packages around | 17:18 |
ogra | but still no jaunty anywhere | 17:18 |
lool | ogra: Get another SD card, install jaunty on it with rootstock? :) | 17:20 |
ogra | nah, i'll use a jaunty live image and will quickly also test amitk's .31 kernel on my B1 | 17:21 |
lool | Upstream requested access | 17:21 |
lool | That will be hard with a live image | 17:21 |
ogra | access ? | 17:21 |
ogra | that will be very hard with my network setup :/ | 17:22 |
lool | Offer to debug then | 17:22 |
ogra | i'm also sure it is fixed in the karmic version | 17:22 |
ogra | just didnt have the time to verify | 17:23 |
ogra | (on jaunty that is) | 17:23 |
lool | Yeah please do take it | 17:24 |
amitk | ogra: I think USB should be fixed now, let me know if it works for you | 17:28 |
Sarvatt | libqt4-webkit-dbg_4.5.1-1ubuntu6_armel.deb (55.3 MiB) wow thats lzma compressed too | 18:07 |
ogra | yeah, they are tiny :P | 18:08 |
NCommander | lool, I might be able to give upstream access. I've requested my superattention powercycle my apartment (:-)) | 18:19 |
NCommander | lool, assuming my powermac isn't completely fried, it should auto-restart once power reapplied, and I can remote in and give access | 18:19 |
NCommander | *superintendent | 18:19 |
ogra | superattending superintended ? | 18:21 |
ogra | boo, why dont we have a redboot-imx for i386 b | 18:22 |
ogra | s/b// | 18:22 |
* ogra grabs the armel binary and copies the bits in place | 18:22 | |
NCommander | ogra, :-P | 18:36 |
NCommander | ogra, the only issue is the machine in question has a karmic base OS, but I think I can do some chroot foo to fix that ... | 18:37 |
NCommander | (I hope) | 18:37 |
=== Stskeepz is now known as Stskeeps | ||
ogra | beyossi, hey ... | 20:16 |
beyossi | hey | 20:16 |
ogra | beyossi, http://elinux.org/BeagleBoardUbuntu has instructions for running ubuntu on beagle | 20:16 |
ogra | hmm, that was a short visit | 20:17 |
beyossi | sorry.. disconnected | 20:22 |
ogra | well, i said http://elinux.org/BeagleBoardUbuntu has instructions for running ubuntu on beagle | 20:22 |
ogra | not sure you catched that before being disconnected | 20:22 |
beyossi | actually i was following the link http://gebaar.blogspot.com/2009/06/ubuntu-on-beagleboard.html but then when booting it stucks after " * Starting GNOME Display Manager... [OK] " | 20:23 |
beyossi | it sucked after " * Starting GNOME Display Manager... [OK]", i could see the desktop but couldn't login as it didn't recognize neither the mic nor the keyboard | 20:24 |
beyossi | sorry I meant mice and not mic | 20:25 |
ogra | yeah, use the ubuntu kernels, that howto uses a debian lenny one | 20:25 |
ogra | http://elinux.org/BeagleBoardUbuntu is definately the best resource | 20:26 |
ogra | you can keep your rootfs, just replace kernel and uboot setup | 20:26 |
beyossi | i will give a try, thanks | 20:27 |
beyossi | just making sure, shall I keep the rootfs i built using the lxde meta-package? | 20:27 |
ogra | it shouldnt be a rootfs issue at all | 20:28 |
ogra | its pretty sure the kernel | 20:28 |
beyossi | Ogra, I just noted that the link you gave me includes the note: "Note: For USB WIFI adapters, add package "linux-firmware" to the --seed ". This means that I need to rebuild the entire rootfs with those packages, isn't it? | 20:44 |
lbt | when I run prelink I get : Using /lib/ld-linux.so.3, not /lib/ld-linux.so.2 as dynamic linker | 21:12 |
lbt | and -u says "does not have .gnu.prelink_undo section" | 21:12 |
lbt | this suggests that it may not be working... ? | 21:13 |
Sarvatt | there we go -- http://sarvatt.com/downloads/armel/ | 21:19 |
Sarvatt | ogra: if it helps any -- https://edge.launchpad.net/~sarvatt/+archive/bugs/ | 21:58 |
Sarvatt | no patch needed | 21:59 |
ogra | Sarvatt, great to hear that | 22:06 |
ogra | does it still produce man-db errors ? | 22:06 |
Sarvatt | know a package off the top of your head that'd generate them? | 22:07 |
ogra | beyossi, just install linux-firmware | 22:07 |
ogra | Sarvatt, hmm, apt-get build-dep mesa i think | 22:07 |
Sarvatt | root@ubuntu-9:/# dpkg-reconfigure man-db | 22:08 |
Sarvatt | Updating database of manual pages ... | 22:08 |
Sarvatt | root@ubuntu-9:/# | 22:08 |
Sarvatt | getting mesa build-deps now to see | 22:08 |
ogra | dpkg-reconfigure man-db doesnt fail here | 22:09 |
ogra | its rather the packages adding manpages | 22:10 |
Sarvatt | no errors :) | 22:10 |
Sarvatt | ack i take that back | 22:10 |
Sarvatt | Processing triggers for man-db ... | 22:11 |
Sarvatt | qemu: uncaught target signal 13 (Broken pipe) - exiting | 22:11 |
ogra | yeah | 22:12 |
ogra | and in the end it segfaults | 22:12 |
* ogra is off again, 11pm | 22:13 | |
Sarvatt | it segfaults for you? i havent had it segfault yet | 22:17 |
Sarvatt | trying to find a simple little benchmark to run to see if its any faster/slower | 22:19 |
Sarvatt | looks like the segfaults are locale specific? never had that problem with en_US.UTF-8 | 22:23 |
=== Stskeepz is now known as Stskeep | ||
=== Stskeep is now known as Stskeeps | ||
* NCommander sighs | 22:42 | |
NCommander | I think I ejected my SD card on my Babbage1 back home | 22:42 |
NCommander | and forgot to reinsert it | 22:42 |
NCommander | *sighs* | 22:43 |
NCommander | <- *prays the B1 doesn't need a reboot between now and three weeks from now* | 22:43 |
=== pwnguin is now known as jldugger | ||
=== jldugger is now known as pwnguin |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!