| 00:00.39 | mickeyl | OT: what FPU performance gain do you expect from soft-float? |
| 00:00.44 | mickeyl | (at average) |
| 00:01.14 | pb_ | I'd expect somewhere around a factor of 10x-20x on average workloads. |
| 00:01.26 | *** join/#oe JameZUK (~none@82-36-85-17.cable.ubr03.harb.blueyonder.co.uk) |
| 00:02.47 | mickeyl | very good. |
| 00:02.57 | hrw | heh.. I forgot what to say to Z to get usbnet working ;) |
| 00:03.04 | mickeyl | heh |
| 00:03.45 | treke|laptop | pb_: what accounts for the performance increase? |
| 00:04.24 | treke|laptop | better code? |
| 00:04.32 | pb_ | treke|laptop: better code, and removal of trap overhead. |
| 00:04.40 | treke|laptop | ag |
| 00:04.41 | treke|laptop | ah |
| 00:05.10 | pb_ | NWFPE is written in C, and not anything close to optimal. You could probably get a factor of at least 2 quite easily by replacing it with a better implementation. |
| 00:05.46 | hrw | "ifconfig usb0 192.168.129.20 netmask 255.255.255.0 gateway 192.168.129.1 broadcast 192.168.129.255" should work |
| 00:06.31 | hrw | uf.. started |
| 00:07.08 | pb_ | Also, emulating instructions is very expensive. When an exception happens, the kernel needs to save a pile of state, go grovelling around in the process image to figure out which instruction faulted, decode the bit pattern to figure out what operation is needed, then finally call into the emulator to actually perform the instruction. |
| 00:07.29 | treke|laptop | sound ugly |
| 00:07.49 | pb_ | If it's a load or store, the kernel has to word-swap between the application (which is expecting big-endian order) and the emulator (which operates in little endian). |
| 00:08.28 | treke|laptop | so what your saying is soft-float is a good thing? |
| 00:08.31 | mickeyl | well, guess a *10 enhancement will convince most gcc 3.x naysayers :D |
| 00:08.43 | pb_ | treke|laptop: that's the executive summary, yeah. |
| 00:09.32 | hrw | mickeyl: can you remote ssh login to Z? |
| 00:09.54 | mickeyl | hrw: hold on, my build machine is remote. i must transfer 14mb via dsl before flashing... |
| 00:09.59 | treke|laptop | pb_: Is there a downside? |
| 00:10.05 | hrw | mickeyl: ouch.. |
| 00:10.17 | mickeyl | what about a size drawback if the FPU code is now inline? |
| 00:10.21 | hrw | mickeyl: how fast dsl you have? |
| 00:10.28 | mickeyl | 786mbit/s |
| 00:10.34 | pb_ | mickeyl: it isn't inline; it's a function call. |
| 00:10.38 | hrw | not bad |
| 00:10.44 | hrw | mickeyl: 640/100kbps here |
| 00:10.46 | pb_ | treke|laptop: loss of backwards compatibility, but I guess C++ folks are used to that. |
| 00:10.55 | treke|laptop | hehe |
| 00:10.58 | treke|laptop | that we are |
| 00:10.58 | mickeyl | heh, we are indeed |
| 00:12.58 | hrw | uf.. wget on Z is good thing |
| 00:13.26 | pb_ | one other issue is that the old hard float system is incapable of coexisting with Intel's new wMMX stuff because the instruction encodings overlap. So, if you want to take advantage of the new xscales, you need to be using soft-float. |
| 00:13.41 | mickeyl | ah that's cool stuff |
| 00:14.06 | treke|laptop | this is good, so trolltech might have to switch to gcc3 :p |
| 00:14.13 | mickeyl | hehe |
| 00:14.17 | hrw | so we are ready for newer models |
| 00:14.30 | pb_ | treke|laptop: hehe, right. well, we can hope |
| 00:14.37 | zecke | so pxa27x is the stock xscale nowadays? |
| 00:15.01 | pb_ | it's getting that way, yes. at least some of the new iPAQs are pxa27x based. |
| 00:15.03 | mickeyl | ok, rcS runs... |
| 00:15.20 | mickeyl | btw., haven't heard much about 2.6 on the Zs lately |
| 00:15.36 | pb_ | yah, I haven't heard anything either. |
| 00:15.46 | hrw | after reflash I will push kernel changes |
| 00:15.46 | treke|laptop | pb_: You happen to know who's making the new ipaqs? |
| 00:15.55 | zecke | pb_: so floating point operation is an operation and softfloat a method. And this is why these two modes are incompatible |
| 00:15.58 | hrw | treke|laptop: HTC? |
| 00:16.02 | pb_ | treke|laptop: nope, no idea. |
| 00:16.07 | treke|laptop | hrw: Hoping not. |
| 00:16.29 | treke|laptop | hrw: There isnt exactly a stunning history of Linux support on HTC built ipaqs |
| 00:16.44 | hrw | booting... |
| 00:17.04 | treke|laptop | what was causing it? |
| 00:17.13 | mickeyl | a glitch in update-rc.d |
| 00:17.30 | treke|laptop | ah |
| 00:17.30 | mickeyl | ok, apart from a couple of '[ : init undefined number] the bootup looks normal |
| 00:17.45 | hrw | mickeyl: and you got "update-modules force" thing |
| 00:18.06 | hrw | locale-base-* postinst are errored |
| 00:18.25 | pb_ | oh dear. what's the error? |
| 00:18.30 | hrw | w8 |
| 00:19.03 | hrw | have to reboot - my opie-image is without big part of opie |
| 00:19.53 | hrw | Configuring locale-base-fr-fr |
| 00:20.10 | hrw | cannot create temporary file: No such file or directory |
| 00:20.43 | pb_ | odd. okay, I'll take a look at that. |
| 00:20.45 | mickeyl | hmm |
| 00:20.49 | mickeyl | pressing cancel closes opie-console |
| 00:20.54 | mickeyl | i thought we didn't want that ? |
| 00:21.19 | hrw | pb_: it;s really odd - that package doesn't have postinst |
| 00:21.28 | treke|laptop | mickeyl: sounds like a bad thing to me |
| 00:22.22 | mickeyl | oh oh |
| 00:22.34 | mickeyl | and texteditor has problems with tabs in files |
| 00:23.12 | pb_ | hrw: ah, maybe /usr/lib/locale is missing. |
| 00:23.18 | pb_ | could you create that directory and see if it helps? |
| 00:23.18 | zecke | mickeyl: need to capture the closeEvent |
| 00:23.23 | treke|laptop | hehe, I've never liked texteditor much :p |
| 00:23.30 | hrw | pb_: ok |
| 00:24.10 | hrw | pb_: cannot map archive header: Invalid argument |
| 00:24.17 | pb_ | ugh |
| 00:24.20 | hrw | instead of temporary file |
| 00:24.21 | pb_ | I guess that's progress of a sort. |
| 00:24.46 | pb_ | sounds like jffs2 disagrees with localedef. I'll look at that tomorrow. |
| 00:26.06 | hrw | heh.. I remember time when I splitted modutils to modutils-depmod and modutils to get some more space on collie.. but all zaurus models will need full modutils ;( |
| 00:26.55 | Robot101 | mooo |
| 00:27.00 | CIA-7 | 03hrw 07 * r1.2394 10packages/ (2 files in 2 dirs): switched to 'update-modules' method of handling modules on collie |
| 00:27.16 | Robot101 | someone spoke to me in this channel but it scrolled off by now |
| 00:27.25 | hrw | now collies 1st boot is much cleaner |
| 00:27.58 | hrw | 01:11 < Cwiiis> Robot101: Could I just remove the RDEPENDS entirely then? |
| 00:29.18 | hrw | hmm.. it's from yesterday log.. |
| 00:29.25 | pb_ | heh |
| 00:29.50 | Robot101 | I'm not much of an irssi guru yet :) |
| 00:29.59 | Robot101 | and no, Cwiiis shouldn't remove the RDEPENDS |
| 00:30.27 | Robot101 | it needs the -data package, that's why I added the RDEPENDS :P |
| 00:30.45 | hrw | Robot101: /set scrollback_lines BIG_NUMBER |
| 00:30.54 | pb_ | Robot101: I think he just changed the package name for the time being. |
| 00:31.23 | *** join/#oe benjp_ (~ben@reynerie.net1.nerim.net) |
| 00:31.33 | Robot101 | what does scrollback_hours do? |
| 00:32.07 | hrw | Robot101: I'm not irssi guru too |
| 00:35.04 | hrw | ok. I did what I planned - cleaning situation of collie modules |
| 00:36.55 | Robot101 | pb_: oh cool, you did esd |
| 00:37.06 | Robot101 | pb_: what's this "bigscreen" lark? |
| 00:37.13 | treke|laptop | Robot101: Irssi has a feature that will prompt you before pasting large chunks of text |
| 00:37.34 | pb_ | Robot101: just a way to get slightly different package sets in gpe-image. |
| 00:37.53 | Robot101 | pb_: is there support for making jffs2 images as well as tarballs? |
| 00:38.26 | pb_ | at the moment it's kind of an either/or thing, but it'd be easy enough to change that to support multiple output formats. |
| 00:38.43 | pb_ | you'd just need to wrap the oeimage call at the end of image_ipk.oeclass in a for-loop or some kind. |
| 00:38.48 | pb_ | s/or/of/ |
| 00:39.11 | Robot101 | cool. could you do that? :) |
| 00:39.23 | pb_ | yah, I'll do that tomorrow |
| 00:39.35 | Cwiiis | Robot101: What should be done? |
| 00:40.38 | hrw | bye |
| 00:40.39 | Robot101 | Cwiiis: well in Debian land, keeping anything that's not the shared lib in a lib package is broken |
| 00:40.55 | Robot101 | so I split the -data to the seperate package and added the depend |
| 00:41.09 | Robot101 | really it needs a conditional to see if it's doing debian-style package renaming |
| 00:41.30 | Robot101 | or you could add something that sets PKG_foo to foo even if package renaming is not being done |
| 00:41.45 | Robot101 | then ${PKG_foo} will always contain the ultimate name of the package |
| 00:42.19 | Cwiiis | Robot101: I don't really understand the problem in the first place, I'm probably not the right person to go about fixing it... |
| 00:42.35 | Robot101 | I'll add it to my todo list |
| 00:42.43 | Cwiiis | Robot101: Cool, thanks |
| 00:43.06 | Robot101 | pb_: it strikes me that nothing checks the exit status of ipkg-cl in image_ipk - so when stuff fails it's not too surprising it doesn't error the image build? |
| 00:43.18 | pb_ | Robot101: it should be executing under 'set -e', iirc |
| 00:43.32 | Robot101 | does set -x at the top override it? |
| 00:43.40 | pb_ | shouldn't do. |
| 00:43.59 | pb_ | my suspicion is that ipkg-cl just isn't setting its return code properly. |
| 00:45.09 | Robot101 | is it a shell script? |
| 00:45.22 | pb_ | no, it's C code. |
| 00:45.41 | pb_ | somewhat hairy code at that |
| 00:45.55 | treke|laptop | is it being maintained these days? |
| 00:46.01 | pb_ | yeah, by Pigi. |
| 00:46.33 | treke|laptop | cool |
| 00:47.07 | pb_ | oh dear, my simpad image is up to 43MB now. |
| 00:47.20 | Robot101 | pb_: did you merge that busybox patch? |
| 00:47.21 | pb_ | clearly removing firefox wasn't the power for good that I hoped it would be. |
| 00:47.34 | Robot101 | pb_: the later version I sent you (the two liner) tested fine |
| 00:47.36 | pb_ | Robot101: I patched busybox up to the CVS HEAD as of this morning. |
| 00:47.45 | pb_ | hopefully that will take care of it. |
| 00:47.52 | Robot101 | thats... perhaps overkill? :) |
| 00:48.08 | pb_ | perhaps, but I wanted one of the other fixes as well. |
| 00:48.19 | pb_ | makedevs is completely hosed in rc2; it creates everything with major=0. |
| 00:48.25 | Robot101 | yow |
| 00:48.43 | Robot101 | anyway |
| 00:48.59 | pb_ | not that we're actually _using_ busybox makedevs at the moment, but it'd be nice to be able to do that. |
| 00:49.54 | pb_ | night all |
| 00:52.53 | CIA-7 | 03mickey 07 * r1.2393.1.1 10packages/sharp-binary-only/ (2 files in 2 dirs): fix the sdmmc package and add support for ext2 |
| 00:54.10 | *** join/#oe Disconnect (~dis@pcp701187pcs.bowie01.md.comcast.net) |
| 01:17.16 | Cwiiis | samba takes far too damn long to compile... |
| 01:22.13 | *** join/#oe ZincOxide (ZincOxide@082.a.002.sop.iprimus.net.au) |
| 01:23.28 | ZincOxide | is debian the easiest to use for OE development? |
| 01:24.32 | Cwiiis | ZincOxide: I've had about the same amount of problems with OE in Debian as I did with Gentoo |
| 01:25.52 | ZincOxide | Gentoo was the other one i was thinking of too |
| 01:26.29 | Cwiiis | Debian is a lot easier to maintain than Gentoo imo |
| 01:26.53 | Cwiiis | Although I think it'd probably be fair to say that that's more fact than opinion.. hehe |
| 01:30.39 | ZincOxide | debian seems to be easier to install too |
| 01:30.51 | Cwiiis | yes, by far |
| 01:31.01 | Cwiiis | But Gentoo isn't as hard as a lot of people would make out |
| 01:31.15 | Cwiiis | And there's tons of support on the forums/IRC |
| 01:32.45 | ZincOxide | yeh, i usually notice that the gentoo forum has the most users ... which to me means most problems |
| 01:37.24 | ZincOxide | can i use fdisk from the Gentoo liveCD to create new partitions in unused space WITHOUT damaging other existing partitions? |
| 01:38.53 | Cwiiis | ZincOxide: Yes |
| 01:39.10 | ZincOxide | tnx |
| 01:47.34 | *** join/#oe qfh (~qfh@ppp-62-245-160-105.mnet-online.de) |
| 01:50.12 | *** join/#oe Cwiiis (~chris@82-43-43-168.cable.ubr02.croy.blueyonder.co.uk) |
| 01:52.59 | *** join/#oe camcam (~camrein@d213-101-207-72.cust.tele2.fr) |
| 02:00.51 | Cwiiis | er, looks like OE isn't producing valid collie images anymore :/ |
| 02:02.59 | Cwiiis | oh no, my bad... |
| 02:03.36 | Cwiiis | great, another rebuild-from-scratch... |
| 02:13.57 | *** join/#oe treke|laptop (~ggilbert@pool0304.cvx34-bradley.dialup.earthlink.net) |
| 02:15.50 | *** join/#oe LordVan (~lordvan@LordVan.developer.gentoo) |
| 02:31.56 | CIA-7 | 03chris 07 * r1.2367.1.16 10packages/gpe-package/ (4 files in 2 dirs): Add gpe-package 0.0.10 |
| 02:32.35 | *** join/#oe mikelp (~phillim@pool-141-150-246-115.pskn.east.verizon.net) |
| 02:48.20 | *** join/#oe benjp_ (~ben@reynerie.net1.nerim.net) |
| 02:53.35 | *** join/#oe ralph__ (~ralph@pD95E8956.dip.t-dialin.net) |
| 02:56.05 | *** join/#oe treke|laptop (~ggilbert@pool0320.cvx35-bradley.dialup.earthlink.net) |
| 04:18.16 | *** join/#oe mithro (~tim@dsl1-83.gw1.adl1.airnet.com.au) |
| 04:18.18 | *** join/#oe LordVan (~lordvan@LordVan.developer.gentoo) |
| 04:49.02 | *** join/#oe benjp_ (~ben@reynerie.net1.nerim.net) |
| 05:02.01 | *** join/#oe bug1 (~bug1@203-217-33-229.dyn.iinet.net.au) |
| 05:29.27 | *** join/#oe mithro_ (~tim@dsl1-83.gw1.adl1.airnet.com.au) |
| 05:39.16 | *** join/#oe Timelord (~TL@65.59.240.28) |
| 05:51.42 | *** join/#oe linuxwhore (~johnh@d10-119.rb.lax.centurytel.net) |
| 06:12.21 | *** join/#oe Rustre (~Rustre@lns-th2-7-82-64-109-180.adsl.proxad.net) |
| 06:12.36 | Rustre | hi |
| 06:21.06 | *** join/#oe LordVan (~lordvan@LordVan.developer.gentoo) |
| 06:24.10 | *** join/#oe Disconnect (~dis@3ffe:bc0:91:0:203:6dff:fe1a:c5b6) |
| 06:49.40 | *** join/#oe benjp_ (~ben@reynerie.net1.nerim.net) |
| 06:59.26 | *** join/#oe synth (~cb@pcp858839pcs.ptchar01.fl.comcast.net) |
| 07:10.29 | *** join/#oe camcam (~camrein@d213-101-207-72.cust.tele2.fr) |
| 08:11.14 | Bernardo | 'morning |
| 08:11.18 | *** join/#oe CIA-7 (~CIA@to.je.spocco.com) |
| 08:34.45 | *** join/#oe wincent (~wincent@void-109.pmnet.uni-oldenburg.de) |
| 08:44.13 | *** join/#oe synth (~cb@pcp858839pcs.ptchar01.fl.comcast.net) |
| 08:50.15 | *** join/#oe benjp_ (~ben@reynerie.net1.nerim.net) |
| 09:10.00 | *** join/#oe reenoo_ (~r@p50867DD0.dip.t-dialin.net) |
| 09:10.04 | *** join/#oe CIA-5 (~CIA@to.je.spocco.com) |
| 09:10.23 | reenoo_ | morning |
| 09:13.45 | Rustre | hi |
| 09:15.12 | *** join/#oe CoreDump__ (~mhentges@xdsl-213-196-250-24.netcologne.de) |
| 09:18.12 | *** join/#oe siti (~user@210-55-242-126.dialup.xtra.co.nz) |
| 09:39.22 | camcam | hi reenoo_ |
| 09:39.34 | camcam | oops. |
| 09:40.22 | Rustre | hi camcam ;) |
| 09:40.27 | camcam | hi reenoo_ hi |
| 09:40.55 | camcam | I'm figthing with 'tab' and R to have Rustre displayed. |
| 09:40.56 | Rustre | try again ;) |
| 09:41.01 | Rustre | lol |
| 09:41.51 | Rustre | will you modify the script that i gave you back or leave the one you gave me on handhelds.org/~christophe ? |
| 09:42.08 | Cwiiis | It seems I can only compile one image at a time for some reason :/ When I make an image afterwards, it doesn't get generated? |
| 09:42.29 | Cwiiis | ah, need to clean the image target... |
| 09:42.33 | Cwiiis | That's new... |
| 09:42.34 | camcam | I will leave it like this. You remember my 7 comments about the modified script... |
| 09:42.39 | Rustre | yup |
| 09:42.48 | Rustre | ok so i can change the wiki page to add a l ink to it ;) |
| 09:43.45 | camcam | No problemo. |
| 09:46.05 | Rustre | still don't have any answer to my mail |
| 09:46.16 | Rustre | asking if it's ok for me to modify the wiki about the helloworldapp |
| 09:48.25 | Rustre | do you want me to add you name to the wiki? |
| 09:48.45 | Rustre | for the moment i said the script was provided by "camcam" |
| 09:53.12 | *** join/#oe ar_ (~ar@port-ip-213-211-231-77.reverse.mdcc-fun.de) |
| 09:54.01 | *** join/#oe Cwiiis (~chris@82-43-43-168.cable.ubr02.croy.blueyonder.co.uk) |
| 09:55.18 | Cwiiis | mmm, the GPE bootsplash is lovely :) |
| 10:03.54 | camcam | Rustre: for the name, you don't need to but you can if you want. People will be able to ask for me on Opie (less stress for other devs to support something they haven't rigth) |
| 10:04.41 | Rustre | then i need you name |
| 10:04.44 | Rustre | i don't know it ;) |
| 10:04.49 | camcam | Note: christophe Amrein-Marie |
| 10:05.09 | camcam | .) |
| 10:05.44 | Rustre | done |
| 10:05.49 | Rustre | tell me if it's ok with you |
| 10:07.18 | camcam | Not ok. The email shoudn't be displayed |
| 10:07.36 | CIA-5 | 03cwiiis 07 * r1.2398 10packages/gpe-package/gpe-package_0.0.10.oe: Mark gpe-package as broken (will fix soon) |
| 10:07.37 | CIA-5 | 03cwiiis 07 * r1.2397 10packages/meta/meta-gpe.oe: Remove gpe-filemanager from gpe-image (was making resulting image too large on collie) |
| 10:07.39 | Rustre | oh? |
| 10:07.46 | Rustre | ok i'll take it out |
| 10:07.50 | camcam | You should remove the .org and @ and replace them by 'point' and at |
| 10:08.23 | Rustre | oh |
| 10:08.24 | camcam | If you put my name, yours should be there too. |
| 10:08.32 | Rustre | where should i put it? |
| 10:08.46 | camcam | christophe at handhelds point org |
| 10:08.48 | Rustre | it's already at the bottom, as the last person that modified the wiki |
| 10:09.00 | camcam | ok |
| 10:09.02 | camcam | :) |
| 10:09.27 | Rustre | ok |
| 10:09.28 | Rustre | is that ok? |
| 10:10.41 | Rustre | maybe we should talk about all that on #opie :p |
| 10:10.54 | camcam | yes and yes. |
| 10:36.09 | CIA-5 | 03cwiiis 07 * r1.2399 10packages/mbmerlin/ (mbmerlin-0.8.1-2/fix_makefile.patch mbmerlin_0.8.1-2.oe): Fix mbmerlin |
| 10:38.30 | Cwiiis | hmm, don't have time to fix gpe-package right now... Will do it tomorrow evening |
| 10:43.02 | *** join/#oe LordVan (~lordvan@LordVan.developer.gentoo) |
| 10:50.53 | *** join/#oe benjp_ (~ben@reynerie.net1.nerim.net) |
| 10:53.59 | *** join/#oe pb_ (~pb@2002:5168:d330:1:20a:95ff:fe76:cd6c) |
| 11:06.09 | hrw | re |
| 11:07.33 | pb_ | hi all |
| 11:08.08 | hrw | pb_: I'm inspecting collie modules more |
| 11:08.18 | pb_ | cool |
| 11:08.40 | reenoo_ | hey pb_ |
| 11:08.50 | pb_ | yo reenoo_ |
| 11:09.52 | hrw | pb_: they suxx |
| 11:10.04 | pb_ | heh |
| 11:10.27 | hrw | usbdcore + sa1100_bi is halt, but with net_fd is ok (but this needs full modutils :( |
| 11:11.44 | hrw | I will remove modutils from rc2.d and retest it (after flash) |
| 11:12.52 | pb_ | yeah, that stuff is all very bad. |
| 11:13.09 | pb_ | can I talk you into working on the 2.6 kernel for collie? |
| 11:17.15 | Cwiiis | pb_: What exactly needs to be done on the 2.6 kernel for collie? |
| 11:19.16 | hrw | pb_: I work on getting 2.4 works better in current OZ/OE - I don't know anything about kernel developing |
| 11:19.39 | pb_ | Cwiiis: drivers need to be written. |
| 11:20.21 | pb_ | I'm not sure exactly what the current status is. Best to check with mickey or kergoth for that. |
| 11:20.35 | hrw | current status was sent to oz-devel |
| 11:20.49 | pb_ | ah, right. I'm not on that list. |
| 11:21.03 | hrw | The good news is that I have pushed some of the changes that were in my |
| 11:21.03 | hrw | jl patch upstream. Notibly the locomo code and the collie base code |
| 11:21.03 | hrw | are now included in the main kernel tree. |
| 11:21.09 | hrw | it's qoute |
| 11:21.13 | pb_ | sounds good |
| 11:21.25 | hrw | kernel 2.6.8-rc2 broke MTD |
| 11:21.36 | pb_ | ah, not quite so good |
| 11:21.41 | hrw | The problem is, 2.6.8 changed around the api that the mtd code uses. |
| 11:21.41 | hrw | Since sharp.c was obsolete, it was not upgraded to use the new api. |
| 11:21.41 | hrw | Thus when trying to compile for collie we get a bunch of compile |
| 11:21.42 | hrw | errors. |
| 11:21.57 | pb_ | oh, I see. so it's not that MTD is broken, just that the collie code is outdated. |
| 11:22.24 | hrw | yep |
| 11:23.11 | pb_ | someone should try to talk John into using a source control system. |
| 11:25.43 | mickeyl | morning folks |
| 11:25.49 | pb_ | mickeyl: good afternoon |
| 11:25.53 | reenoo_ | hi mickeyl |
| 11:29.32 | hrw | mickeyl: can I ask you about collie 2.4 kernel modules? |
| 11:29.44 | mickeyl | go ahead. first coffee already in |
| 11:30.07 | hrw | mickeyl: sa1100_bi depends on usbdcore - but it also need net_fd? |
| 11:30.17 | mickeyl | no |
| 11:30.25 | mickeyl | net_fd is on top of the others |
| 11:31.35 | mickeyl | for collie, we don't need to rely on order |
| 11:31.40 | mickeyl | it's only on pxa |
| 11:31.44 | mickeyl | where it crashes |
| 11:31.54 | mickeyl | if the order isn't the one the lineo engineers had in mind |
| 11:31.55 | mickeyl | *cough* |
| 11:33.14 | mickeyl | hmm |
| 11:33.44 | hrw | ah |
| 11:34.05 | mickeyl | usbdcore bases on sa1100_bi and net_fd |
| 11:34.12 | pb_ | brb |
| 11:34.15 | mickeyl | c'ya |
| 11:34.17 | hrw | mickeyl: so I can load net_fd and then sa1100_bi? |
| 11:34.30 | mickeyl | yes. net_fd and sa1100_bi are on collie totally unrelated |
| 11:34.36 | mickeyl | at least in theory |
| 11:34.37 | mickeyl | let me try |
| 11:34.41 | hrw | mickeyl: now I set sa1100_bi loads net_fd and usbdcore |
| 11:35.12 | mickeyl | ah wait |
| 11:35.17 | mickeyl | ok, that's tricky |
| 11:35.31 | mickeyl | symbol reference is not exactly represented in usage statistics |
| 11:35.44 | mickeyl | ok |
| 11:35.47 | mickeyl | it's like that |
| 11:35.52 | mickeyl | usbdcore must be in first |
| 11:36.25 | hrw | atleast before sa1100_bi |
| 11:36.52 | mickeyl | on collie though we can just rely on modprobe resolving the dependencies |
| 11:36.56 | mickeyl | no need to worry |
| 11:37.15 | hrw | ah |
| 11:41.22 | CIA-5 | 03mickey 07 * r1.2400 10packages/orinoco/ (5 files in 2 dirs): add orinoco modules 0.15 rc2 |
| 11:42.06 | hrw | mickeyl: what licence are sharp modules? (except sdmmc) |
| 11:42.50 | mickeyl | i think we are safe to say they're GPL |
| 11:42.56 | mickeyl | so we could add it into the MODULE code |
| 11:43.00 | mickeyl | to prevent warnings |
| 11:43.04 | hrw | I just thought about it |
| 11:43.15 | hrw | <PROTECTED> |
| 11:43.20 | hrw | it say GPL even |
| 11:43.26 | mickeyl | pretty good |
| 11:43.28 | mithro | hey mickeyl |
| 11:43.28 | mickeyl | can you make a patch? |
| 11:43.31 | mickeyl | hi mithro |
| 11:43.43 | mickeyl | hrw: it's easy, just adding MODULE_LICENSE("GPL") macro |
| 11:43.44 | mickeyl | (or so) |
| 11:43.44 | hrw | probably |
| 11:44.01 | mithro | mickeyl: any chance you could make it to the linux.conf.au if there was a miniconference about Embedded Linux? :) |
| 11:44.26 | mickeyl | mithro: d'oh. australia is pretty far away from here. |
| 11:44.43 | mickeyl | and my amount of vacation is nearly empty :( |
| 11:44.48 | mithro | mickeyl: yeah - i know :( |
| 11:45.01 | mithro | mickeyl: can't get work to fund your trip? ;) |
| 11:45.05 | mickeyl | hehe |
| 11:45.17 | mithro | lots of good PR ;) |
| 11:46.11 | mithro | i would like to run one - but i doubt i can get enough people |
| 11:48.15 | mithro | i think i could get the trolls to come - but i doubt that would make a very intresting conference ;) |
| 11:48.25 | mickeyl | hehe |
| 11:48.27 | mickeyl | true that |
| 11:49.24 | mithro | you could come an heckly ljp ;) |
| 11:49.35 | mickeyl | that alone would be worth it |
| 11:49.36 | mickeyl | hahah |
| 11:49.38 | mickeyl | brb |
| 11:49.45 | reenoo_ | heh |
| 11:49.56 | mithro | so i'll have to get ljp to come and give a talk to get everyone else to come :) |
| 11:56.34 | *** join/#oe pb_ (~pb@dsl-62-3-66-204.zen.co.uk) |
| 11:59.15 | mithro | hey pb_ |
| 11:59.28 | mithro | any chance you could get yourself to linux.conf.au 2005? :) |
| 12:00.43 | *** join/#oe zecke (~ich@83.169.133.125) |
| 12:01.04 | mithro | by get yourself i mean get your employer to pay you to go ;) |
| 12:01.46 | pb_ | mithro: probably. |
| 12:02.06 | pb_ | hi zecke |
| 12:02.33 | zecke | moin |
| 12:04.04 | mithro | pb_: i'm trying to find speakers and attendee's for a possible OpenEmbedded miniconference at linux.conf.au 2005 |
| 12:06.47 | hrw | ERROR: strange, with --dry-run the patch worked, but without it does not work ... |
| 12:06.49 | hrw | nice |
| 12:07.41 | hrw | found |
| 12:18.01 | hrw | ok - reboot to check module_licence |
| 12:19.06 | hrw | works |
| 12:20.59 | CIA-5 | 03hrw 07 * r1.2401 10packages/linux/openzaurus-sa_2.4.18-rmk7-pxa3-embedix20030509.oe: collie: set some modules licence to GPL |
| 12:21.00 | CIA-5 | 03hrw 07 * r1.2400 10packages/ (3 files in 2 dirs): collie modules clean up - modutils-collateral is not more needed |
| 12:22.03 | CIA-5 | 03hrw 07 * r1.2403 10packages/linux/openzaurus-sa-2.4.18-rmk7-pxa3-embedix20030509/module_licence.patch: collie: set some modules licence to GPL (forgotten patch) |
| 12:28.01 | mickeyl | woot |
| 12:28.22 | mickeyl | hmm |
| 12:28.28 | mickeyl | flakey things |
| 12:28.44 | mickeyl | spectrum_cs is much better dealing with firmware uploadable cards than hostap_cs |
| 12:28.51 | mickeyl | then again, they just started to support that at all |
| 12:29.53 | hrw | hmm.. found two more tainted modules... |
| 12:30.08 | Rustre | what does tainted means? |
| 12:30.10 | mickeyl | which onees ? |
| 12:30.19 | hrw | usbdserial serial_fd |
| 12:30.27 | mickeyl | these are gpl, too |
| 12:30.36 | hrw | sdmmc will stay tainted, registers doesn't have licence info |
| 12:30.38 | mickeyl | Rustre: non-GPL |
| 12:30.52 | hrw | mickeyl: I fixed module_licence patch already for them |
| 12:30.55 | Rustre | mickeyl : ok thx |
| 12:31.29 | hrw | Rustre: tainted == "non-free, w/o source" |
| 12:32.01 | Rustre | w/o ? |
| 12:32.10 | reenoo_ | without |
| 12:32.11 | Rustre | write only ? |
| 12:32.15 | Rustre | ok :p |
| 12:32.21 | Rustre | ok i understand |
| 12:32.42 | Rustre | software that are non gpl but used by software or libraries that are gpl |
| 12:32.54 | Rustre | err not gpl but opensource |
| 12:34.23 | mithro | Rustre: talking about the kernel |
| 12:34.47 | Rustre | oh |
| 12:34.47 | Rustre | ok |
| 12:35.36 | Rustre | is it allowed (regarding to licence) to load non gpl modules in a gpl kernel? |
| 12:37.29 | mickeyl | allowed yes. however the kernel is then regarded "tainted" and you won't get support from the developers |
| 12:37.34 | mickeyl | that's the whole thing about being tainted |
| 12:37.55 | Rustre | ok |
| 12:38.01 | Rustre | understood ;) |
| 12:39.00 | reenoo_ | mickeyl: legally it's probably not allowed. the point is that Linus just doesn't care. |
| 12:39.46 | Rustre | oh |
| 12:40.09 | mickeyl | well why should he |
| 12:40.22 | mickeyl | it's the users with validate the license |
| 12:40.25 | mickeyl | not linus |
| 12:40.29 | mickeyl | s/with/who/ |
| 12:42.07 | reenoo_ | nm |
| 12:43.11 | mickeyl | really, it's a can of worms anyway |
| 12:43.19 | mickeyl | probably the whole sharp rom is not legal at all |
| 12:43.42 | Rustre | lol |
| 12:43.46 | Rustre | that's scary |
| 12:44.10 | reenoo_ | arguably the vanilla sources are not legal at all. sad, but true. |
| 12:44.13 | *** join/#oe niklash (~niklas@h164n2fls33o828.telia.com) |
| 12:44.25 | Rustre | mmm why is that? |
| 12:45.07 | reenoo_ | Rustre: read debian-legal or wherever that was discussed by the debian people |
| 12:45.12 | hrw | /lib/modules/2.4.18-rmk7-pxa3-embedix/kernel/arch/arm/mach-sa1100/registers.o only one left ;) |
| 12:45.57 | zecke | mickeyl: but only Linus and other kernel developers can complain about licensing |
| 12:46.54 | mithro | actually users can't voliate a license |
| 12:47.29 | mithro | only distributers |
| 12:48.10 | Rustre | fdhdutrrytrfkkulku |
| 12:48.12 | reenoo_ | users can. e.g. by linking to non-gpl-compatibly licensed stuff |
| 12:48.31 | mithro | reenoo_: users can't |
| 12:48.34 | Rustre | (this last contribution to the channel was my girlfriend's :p° |
| 12:48.52 | mithro | a user is free to do what ever they want to with the stuff |
| 12:49.02 | mithro | the GPL does not restrict users in anyway |
| 12:49.17 | mithro | it does however stop people from distributing non-gpl-compatible lincese stuff |
| 12:49.29 | mithro | it i can't give it to you without the GPL kicking in |
| 12:50.01 | mithro | i have long chats with the FSF and Debian-Legal about it |
| 12:50.24 | mithro | and thats both their consenses |
| 12:51.16 | mithro | so can people please remeber that next time somebody claims violation of the GPL? |
| 12:52.02 | zecke | mithro: who claimed otherwise? |
| 12:52.24 | mithro | the only difference that Debian-Legal and FSF is there definition of "distribution" |
| 12:52.36 | mithro | [22:20] <reenoo_> users can. e.g. by linking to non-gpl-compatibly licensed stuff |
| 12:53.09 | *** join/#oe benjp_ (~ben@reynerie.net1.nerim.net) |
| 12:53.27 | mithro | mickeyl: not to say all microsoft stuff is most probably not legal either :) i wouldn't like to count the number of things which have creeped there way into MS source trees |
| 12:53.57 | mickeyl | no kidding |
| 12:54.28 | mithro | it just shows how much licensing, patents and other crap suck when it comes to software |
| 12:55.03 | CIA-5 | 03hrw 07 * r1.2404 10packages/linux/openzaurus-sa-2.4.18-rmk7-pxa3-embedix20030509/module_licence.patch: collie: all GPL modules have licence set |
| 12:56.39 | hrw | enough kernel work |
| 12:56.50 | mickeyl | do you have a wlan card ? |
| 12:56.58 | mithro | hrw: you be able to get yourself to linux.conf.au 2005? :) |
| 12:57.41 | reenoo_ | mithro: would you agree with me that's it's not legally possible to link GPL'ed code against non-GPL-compatible code? |
| 12:58.03 | mithro | reenoo_: that is incorrect |
| 12:58.27 | mithro | it's not legal to link GPL'ed code against non-GPL-compatible code and DISTRIBUTE IT |
| 12:58.27 | hrw | mickeyl: no other cards then 2 cf memory cards |
| 12:58.36 | mickeyl | hrw: ok. |
| 12:58.40 | mithro | the little end bit is the most important bit |
| 12:58.47 | hrw | mithro: who will sponsor me? :) |
| 12:59.06 | mithro | hrw: sadly noone :( |
| 12:59.37 | mithro | hrw: although if you do a talk you'll get free entry to the conference i belive |
| 13:00.45 | hrw | mithro: .au is on opposite point of earth for me |
| 13:00.56 | mithro | hrw: dang :( |
| 13:01.23 | zecke | hrw: just dig a hole then |
| 13:01.27 | mickeyl | hehe |
| 13:01.29 | mithro | hey zecke |
| 13:01.35 | mithro | zecke: anyway you could make it? |
| 13:01.58 | hrw | zecke: you are from .de as mickeyl - maybe you start digging and I will help? |
| 13:01.59 | zecke | mithro: traveling expenses are too high |
| 13:02.06 | mickeyl | heh |
| 13:02.25 | mickeyl | zecke is not in the shape to dig and me neither. harlekin is in good shape, he can do |
| 13:02.31 | mickeyl | :D |
| 13:03.08 | mithro | mickeyl and zecke are both in germany? |
| 13:03.14 | mickeyl | ya |
| 13:03.24 | mickeyl | and harlekin, ar, groucho, alwin, coredump |
| 13:03.25 | mithro | get a group booking :) |
| 13:03.34 | mithro | and bring the whole crowd |
| 13:04.02 | mickeyl | that'd be nice but we seem unable to even arrange a local meeting let alone an international one... |
| 13:04.12 | CoreDump|home | mickeyl: :) |
| 13:11.02 | hrw | lufthansa doesn't give price for tickets from berlin to sydney... |
| 13:13.57 | zecke | mickeyl: how can I eject CONFIG -= thread again? |
| 13:14.04 | mickeyl | inject ? |
| 13:14.18 | mickeyl | EXTRA_QMAKEVARS_POST += "CONFIG-=thread" |
| 13:14.23 | zecke | :) |
| 13:15.05 | reenoo_ | mithro: hmm, ok |
| 13:15.11 | mickeyl | mithro: scummvm 0.6.1 is out |
| 13:15.16 | mickeyl | the patches don't apply |
| 13:15.21 | hrw | EXTRA_QMAKEVARS_POST is veery powerfull thing ;) |
| 13:15.27 | mickeyl | hehe |
| 13:15.57 | pb_ | mithro: what date is lca 2005? |
| 13:17.25 | mithro | it's in april i belive |
| 13:17.37 | mithro | mickeyl: i'll do scummvm 0.6.1 sometime this week :) |
| 13:17.45 | mickeyl | mithro: ok, cool |
| 13:18.03 | zecke | mithro: it is a shame that the PockePC GUI for ScummVM is not generally available |
| 13:18.20 | mithro | zecke: what do you mean? |
| 13:20.00 | zecke | mithro: ScummVM the PocketPC Version is quite adjusted to the small screen (including virtual keyboard) |
| 13:20.16 | zecke | mithro: and the SDL/X11 versions are not meant for handhelds. Or did this change? |
| 13:20.30 | mithro | zecke: the PocketPC version now uses the SDL backend |
| 13:20.52 | mithro | so I may be able to get to use the virtual keyboard stuff in the zaurus version |
| 13:21.08 | mickeyl | someone should start feeding sdl patches upstream :( |
| 13:21.17 | *** join/#oe LordVan (~lordvan@LordVan.developer.gentoo) |
| 13:21.18 | mickeyl | we're diverging more andmore |
| 13:21.22 | mickeyl | that's not good |
| 13:21.31 | zecke | mithro: wow, I started porting the PocketPC stuff sometime ago as well |
| 13:21.33 | mithro | mickeyl: i know - i havn't had the time :( |
| 13:21.44 | zecke | mithro: as I really love to play Day of Tentacles |
| 13:21.50 | mithro | zecke: the SDL/X11 stuff works fine on my zaurus however |
| 13:22.03 | mithro | (with the qtopia patches |
| 13:22.04 | mithro | ) |
| 13:22.17 | mithro | been playing BrokenSword 1 on it lately |
| 13:22.34 | zecke | mithro: but not on iPAQ. Generally it would be awesome to toggle a virtual keyboard generally in SDL/QPE |
| 13:22.52 | mithro | bah - i don't have a linux based iPAQ :) |
| 13:23.03 | mithro | zecke: i'll look into it |
| 13:23.10 | mithro | but i'm kinda pressed for time at the moment |
| 13:23.27 | mithro | trying to release MiniSec http://www.thousandparsec.net/tp/ |
| 13:24.09 | zecke | mithro: is there a nice car game :} |
| 13:24.23 | mithro | zecke: ? |
| 13:24.38 | zecke | mithro: auto racing game (SDL based) |
| 13:24.45 | mickeyl | there isn't |
| 13:25.09 | mithro | zecke: not that i know off |
| 13:26.24 | *** part/#oe jmau (~jmau@pD9E5DFE7.dip.t-dialin.net) |
| 13:28.00 | hrw | someone tried journaling FS on CF card? |
| 13:29.21 | mickeyl | i think there are more interesting ways to damage cf cards |
| 13:29.51 | hrw | EXT2-fs: Unrecognized mount option user |
| 13:29.53 | hrw | hmm |
| 13:30.14 | mickeyl | need to patch fstab |
| 13:30.51 | hrw | I thought that ext2 recognise this |
| 13:46.39 | *** join/#oe joshin (~joshin@VDSL-130-13-185-112.PHNX.QWEST.NET) |
| 14:14.51 | hrw | heh.. have to rebuild whole opie ;( |
| 14:16.14 | chouimat | morning |
| 14:17.52 | *** join/#oe hcarty (~hcarty@dsl093-100-214.wdc2.dsl.speakeasy.net) |
| 14:18.28 | *** part/#oe hcarty (~hcarty@dsl093-100-214.wdc2.dsl.speakeasy.net) |
| 14:19.25 | CIA-5 | 03mickey 07 * r1.2404.1.1 10packages/ethload/ (ethload-native_1.2.oe ethload_1.2.oe): add ethload, a minimal tftp server program |
| 14:20.45 | hrw | hi drw |
| 14:20.52 | drw | hi hrw |
| 14:21.01 | hrw | mickeyl: opie-networksettings_cvs.oe are 1.1.4 not 1.1.5 |
| 14:21.17 | mickeyl | ? |
| 14:21.35 | mickeyl | ah |
| 14:21.38 | hrw | PV=1.1.4-${CVSDATE} not 1.1.5-${CVSDATE} |
| 14:21.47 | mickeyl | can you upgrade ? must have been slipped through my replacement skript |
| 14:21.54 | hrw | ok |
| 14:21.56 | mickeyl | thanks |
| 14:24.09 | hrw | np |
| 14:26.00 | hrw | mickeyl: how to get opie started from CF? CF is mounted by /etc/init.d/pcmcia but opie doesn't start by default |
| 14:26.49 | mickeyl | probably need to adjust paths in /etc/init.d/opie |
| 14:27.02 | CIA-5 | 03hrw 07 * r1.2407 10packages/opie-networksettings/opie-networksettings_cvs.oe: updated version to the same as upstream |
| 14:27.18 | hrw | mickeyl: it work when I login and start by hand |
| 14:27.23 | hrw | will look into later |
| 14:27.31 | CIA-5 | 03mickey 07 * r1.2406.1.1 10packages/cetools/ (cetools-native_0.3.oe cetools_0.3.oe): add cetools, which converts between various rom image formats |
| 14:27.57 | hrw | "various rom image formats" means? |
| 14:28.34 | mickeyl | linux-sh formats, tftp rom headers |
| 14:28.40 | mickeyl | non-zaurus things |
| 14:32.01 | hrw | ok |
| 14:38.59 | niklash | I've just converted a homemade little program to use auto*, and added it to my local packages dir. |
| 14:39.11 | niklash | When I try to build it, I get an error during autoreconf. |
| 14:39.22 | niklash | AM_PATH_SDL not found in library. |
| 14:39.33 | niklash | It works when I run it on my local computer. |
| 14:40.11 | niklash | I'm not an auto* guru, but I'd guess this error wouldn't arise if OE didn't call autoreconf, but just configure. |
| 14:40.25 | niklash | How can I fix this? |
| 14:40.54 | mithro | niklash: it caused by SDL not been found :) |
| 14:41.00 | mithro | try adding SDL to your depends |
| 14:41.10 | *** join/#oe Dessimat0r (Dessimat0r@spc1-birk1-3-0-cust203.bagu.broadband.ntl.com) |
| 14:41.14 | *** join/#oe qfh (~qfh@ppp-62-245-160-105.mnet-online.de) |
| 14:41.33 | niklash | I have set DEPENDS="virtual/libsdl mesa". |
| 14:41.51 | hrw | niklash: build via oemake? |
| 14:42.05 | niklash | It built mesa and libsdl-x11 (after I changed local.conf to prefer x11 over qte). |
| 14:42.10 | niklash | hrw: Yes. |
| 14:44.04 | hrw | niklash: you builded libsdl-qpe before building libsdl-x11? |
| 14:44.25 | *** join/#oe andersee (foobar@codepoet.org) |
| 14:44.27 | niklash | I Ctrl-C:d the build of libsdl-qpe. |
| 14:44.59 | hrw | in which stage? |
| 14:45.10 | hrw | maybe it left some things in staging |
| 14:45.14 | niklash | I don't know. |
| 14:45.40 | mithro | niklash: hrm... |
| 14:45.49 | mithro | i know whats causing this - i just don't know how to fix :) |
| 14:46.07 | mithro | it caused by your local.ac no longer having AM_PATH_SDL in it |
| 14:46.19 | mithro | which happens because OE does a reconfigure |
| 14:46.28 | mithro | check out how others do it |
| 14:46.40 | mithro | i think they just patch it to include the thing again |
| 14:47.10 | mithro | ask mickeyl when he gets back |
| 14:47.16 | mithro | or as kergoth if he turns up |
| 14:49.48 | niklash | What is the fastest way to make sure a package is completely rebuilt. Increase the revision? (How?) oebuild clean? |
| 14:50.26 | hrw | niklash: oebuild clean package.oe;oebuild package.oe |
| 14:50.45 | hrw | increase by bumping PR |
| 14:51.13 | niklash | OK. I got the impression the other day that adding -r1 or something is preferable or something, or maybe that it avoids problems with ipkg:s available metadata. |
| 14:55.18 | niklash | I copied do_configure from scummvm. That seems to have done the trick. |
| 14:55.48 | mithro | niklash: ekk! using my do_configure :) |
| 14:58.28 | mjc | doom3 :) |
| 14:59.31 | hrw | mickeyl: opie-camera is from cvs but still PV at 1.0.1? |
| 14:59.43 | mickeyl | yes |
| 15:00.08 | mickeyl | we will eventually use individual versions for all programs |
| 15:00.15 | mickeyl | i started with camera and wellenreiter |
| 15:00.16 | hrw | opie-handwriting has PR="2" not "r2" |
| 15:00.35 | mithro | does anyone have a zaurus camera? |
| 15:04.38 | *** join/#oe chii (chii@chii.utility.freenode) |
| 15:04.39 | *** join/#oe CIA-9 (~CIA@to.je.spocco.com) |
| 15:09.09 | *** join/#oe benjp_ (~ben@reynerie.net1.nerim.net) |
| 15:10.56 | *** join/#oe pb__ (~pb@cpc2-cmbg6-5-0-cust48.cmbg.cable.ntl.com) |
| 15:12.00 | pb__ | re |
| 15:12.21 | hrw | re pb |
| 15:25.04 | niklash | I'm having a problem with modprobe on my gpe-image. It doesn't automatically load depended upon modules. |
| 15:25.55 | hrw | niklash: can you run depmod and update-modules and retry? |
| 15:26.01 | niklash | I try to modprobe <whatever module>, and get unresolved symbols. |
| 15:26.15 | niklash | If I load those modules manually, it works then. |
| 15:26.19 | niklash | hrw: I have. |
| 15:26.49 | niklash | hrw: I also removed the ipsec module, as that caused an error in depmod -a. (Unresolved symbols.) |
| 15:27.11 | niklash | I see there's a new version of busybox, that I'm building now. |
| 15:27.19 | hrw | niklash: can you look what permissions that modules have? |
| 15:27.31 | niklash | So if there are recent changes that should fix this, that may fix it. |
| 15:29.45 | niklash | hrw: -rw-r--r-- |
| 15:30.08 | hrw | niklash: ok |
| 15:31.46 | niklash | I've noticed that a few modules are in /lib/modules/*hh36.12/ |
| 15:32.07 | niklash | While most are in /modules/*hh36/ |
| 15:33.38 | pb__ | sounds like a bug. |
| 15:34.47 | niklash | All modules that are required for h3600_asic are in hh36/ though, so a modprobe of this should work. |
| 15:34.55 | niklash | pb__: I agree. |
| 15:43.30 | *** join/#oe SuN (unices@195-241-12-117-mx.xdsl.tiscali.nl) |
| 15:48.36 | *** join/#oe synth (~cb@pcp858839pcs.ptchar01.fl.comcast.net) |
| 15:50.37 | *** join/#oe LordVan (~lordvan@LordVan.developer.gentoo) |
| 16:11.18 | CIA-9 | 03pb 07 * r1.2393.2.1 10packages/classes/rootfs_ipk.oeclass: s/do_images/do_rootfs/ |
| 16:18.11 | *** join/#oe jmau (~jmau@pD9E5DFE7.dip.t-dialin.net) |
| 16:19.47 | CIA-9 | 03pb 07 * r1.2410 10packages/sysvinit/ (sysvinit-2.85/inittab sysvinit_2.85.oe): don't run getty on tty1 if USE_VT=0 |
| 16:25.08 | pb__ | niklash: it's possible that busybox modprobe is just broken. You might need to take that up with the busybox folks. |
| 16:26.08 | *** join/#oe florian_ (~fuchs@pD9E5FD25.dip.t-dialin.net) |
| 16:26.11 | florian_ | hi all |
| 16:26.18 | pb__ | hi florian_ |
| 16:26.23 | florian_ | This looks like a bug: |
| 16:26.25 | CIA-9 | 03pb 07 * r1.2411 10packages/ (classes/image_ipk.oeclass conf/oe.conf): allow multiple IMAGE_FSTYPES |
| 16:26.27 | florian_ | NOTE: the following files were installed but not shipped in any package: |
| 16:26.27 | florian_ | NOTE: /usr/share/application-registry/gnome-vfs.applications |
| 16:26.36 | pb__ | That does sound like a bug. |
| 16:26.48 | florian_ | This gnome-mime-data |
| 16:26.50 | niklash | pb__: Yeah. I'll try using the modprobe from familiar unstable to see if modprobe is the culprit. |
| 16:27.01 | pb__ | Good idea. |
| 16:27.08 | florian_ | eh this is... |
| 16:28.31 | CIA-9 | 03pb 07 * r1.2412 10packages/classes/gnome.oeclass: add ${datadir}/application-registry to FILES_${PN} |
| 16:28.33 | pb__ | florian_: that might help, if the offending package is using gnome.oeclass. |
| 16:30.53 | florian_ | pb__: it doesn't |
| 16:31.42 | pb__ | doesn't help, or doesn't use the class? |
| 16:32.33 | florian_ | pb__: it only uses autotools and pkgconfig only... |
| 16:32.48 | pb__ | ah. |
| 16:33.09 | pb__ | maybe we should fix it to use gnome. |
| 16:33.13 | pb__ | is this gnome-vfs in question? |
| 16:34.46 | florian_ | oops... gnome-mime-data |
| 16:35.00 | pb__ | ok, I'll take a look |
| 16:35.21 | florian_ | may thanks! |
| 16:36.22 | pb__ | three? crumbs. |
| 16:36.23 | hrw | niklash: you can also use modprobe from modutils |
| 16:38.30 | florian_ | pb__: Yes... we have a pile of IBM Netvista thinclients at kc, just for fun... then there is the Hoeft and Wessel Integral and the SkeyePad. |
| 16:38.40 | pb__ | ah, very cool. |
| 16:39.01 | florian_ | indeed... |
| 16:39.37 | pb__ | which one is the Integral? |
| 16:39.48 | pb__ | is that the ruggedised handheld thing? |
| 16:39.58 | florian_ | Yep |
| 16:41.08 | florian_ | I think we will finish linux port to it soon, it is almost perfect now... |
| 16:41.39 | pb__ | excellent. |
| 16:41.57 | florian_ | But it needs two jffs2 images, it has nor flash for kernel/bootloader and NAND for a filesystem. |
| 16:42.23 | pb__ | ah, right. |
| 16:42.40 | pb__ | does it actually want jffs2 for the NOR, or just a pair of binaries glued together? |
| 16:44.11 | pb__ | This morning I was working on the port for a new s3c2410-based machine, which has only NAND flash - no NOR at all. |
| 16:44.22 | CIA-9 | 03pb 07 * r1.2413 10packages/classes/gnome.oeclass: also add ${datadir}/mime-info to FILES_${PN} |
| 16:44.39 | CIA-9 | 03pb 07 * r1.2414 10packages/gnome/gnome-mime-data_2.4.1.oe: use gnome.oeclass |
| 16:44.53 | pb__ | heh |
| 16:44.55 | pb__ | -r--r--r-- 1 pb pb 15 Aug 1 17:34 gnome-mime-data_2.4.1.oe |
| 16:45.04 | pb__ | I think this must be one of the smallest .oe files we have in the repository :-) |
| 16:45.17 | florian_ | really jffs2 :-) so we should be able to mount it to /boot :-) |
| 16:45.23 | pb__ | cute. |
| 16:45.55 | florian_ | :-) |
| 16:47.25 | hrw | pb__: 15 bytes?? wow... |
| 16:47.52 | florian_ | huh... what is this for a cpu? how does this work? relying on an error-free first block? |
| 16:48.42 | niklash | cd |
| 16:48.49 | niklash | Sorry, wrong window. |
| 16:49.53 | *** join/#oe treke|laptop (~ggilbert@pool0395.cvx30-bradley.dialup.earthlink.net) |
| 16:50.06 | pb__ | florian_: yes. there is some special boot logic in the chip which reads the first 4k of flash into an internal SRAM. The CPU starts executing there, turns on the SDRAM array, then copies more code out of NAND and branches to the new image in SDRAM. |
| 16:51.26 | pb__ | it does depend on the first block not being bad, but this seems to be something that's commonly guaranteed by flash vendors anyway. |
| 16:51.41 | florian_ | pb__: oh... sounds nasty. |
| 16:51.52 | florian_ | yes, right... |
| 16:52.31 | pb__ | florian_: it's kind of funky, but it obviously avoids both the cost and footprint of a NOR chip. |
| 16:53.53 | florian_ | pb__: yes, right... |
| 16:54.37 | pb__ | and being able to have it boot directly off a smartmedia card is rather cool. no need to use jtag for reflashing :-) |
| 16:55.31 | florian_ | indeed, this would be really cool! |
| 16:55.33 | pb__ | hrw: heh. actually, it should only have been 14 bytes, but a stray newline seems to have crept in to the file :-} |
| 16:58.20 | hrw | heh.. some guys want oinclude ROX-Filer in debian... I wish them luck ;) |
| 17:05.27 | *** join/#oe Bernardo (~jbbrs@213.58.75.33) |
| 17:07.15 | *** join/#oe Disconnect (~dis@3ffe:bc0:91:0:203:6dff:fe1a:c5b6) |
| 17:08.12 | treke|laptop | evening pb |
| 17:08.19 | pb__ | morning treke |
| 17:08.25 | pb__ | afternoon chouimat |
| 17:08.47 | chouimat | pb__: the grocery and the liquor store (lcbo) are far away :( |
| 17:08.56 | pb__ | doh |
| 17:08.58 | treke|laptop | hehe aren't timezones grand :p |
| 17:09.04 | pb__ | heh |
| 17:09.15 | chouimat | treke|laptop: hi |
| 17:11.04 | Bernardo | any special reason for opie-networksettings-wlanplugin being out of the default opie image? |
| 17:11.46 | treke|laptop | probably not |
| 17:12.23 | treke|laptop | if there are wireless drivers included, feel free to add it |
| 17:12.24 | Bernardo | possibly I am the only guy here with a wlan card... :) |
| 17:12.45 | Bernardo | I don't have bk write access |
| 17:12.55 | *** join/#oe synth (~cb@pcp858839pcs.ptchar01.fl.comcast.net) |
| 17:13.01 | pb__ | send a patch to oe@handhelds.org and someone will check it in. |
| 17:13.07 | Bernardo | ok |
| 17:13.55 | pb__ | don't suppose anyone here knows about setting up Windows clients to talk to freeswan ipsec? |
| 17:14.33 | *** join/#oe PRoSPeRe (~prospere@APuteaux-151-1-53-246.w82-124.abo.wanadoo.fr) |
| 17:15.06 | *** part/#oe PRoSPeRe (~prospere@APuteaux-151-1-53-246.w82-124.abo.wanadoo.fr) |
| 17:15.48 | florian_ | pb__: Using x.509? Define Windows... |
| 17:16.11 | pb__ | florian_: Windows XP. I think I need to use a pre-shared key, rather than X.509. |
| 17:16.13 | pb__ | at least to start with. |
| 17:16.53 | pb__ | seems that XP expects to use "l2tp" on top of IPSEC. I guess I need to install an l2tp server of some kind on my router. |
| 17:18.04 | florian_ | pb__: just ipsec should work fine, there is a small tool that eats a config file that is similar to freeswan. |
| 17:18.54 | Bernardo | where would be the most apropriate place for opie-networksettings-wlanplugin, in OPIE_BASE_SETTINGS as is opie-networksettings; or OPIE_EXTRA_SETTINGS? |
| 17:19.40 | florian_ | pb__: http://vpn.ebootis.de/ |
| 17:20.51 | florian_ | pb__: but i'm not sure if this works with preshared keys... |
| 17:22.04 | florian_ | i think it does... there is some setting in the config example... |
| 17:22.54 | benjp_ | l2tp is not mandatory |
| 17:23.32 | *** join/#oe reenoo_ (~r@p50867DD0.dip.t-dialin.net) |
| 17:23.40 | reenoo_ | re |
| 17:24.33 | benjp_ | you can even force tunneling mode (http://support.microsoft.com/default.aspx?scid=kb;en-us;252735&sd=tech) |
| 17:24.35 | pb__ | benjp_: ah. how do I persuade XP to use IPSEC without l2tp? |
| 17:24.41 | pb__ | aha, thanks |
| 17:26.03 | pb__ | hm, I fear that the MMC configuration there will be too complicated for my users. |
| 17:27.06 | florian_ | huh? |
| 17:27.13 | pb__ | seems that there's an l2tpd in Debian, so I'll try that. |
| 17:27.14 | benjp_ | well, so use the vpn.ebootis.de's « VPN tools » |
| 17:27.28 | pb__ | florian_: my users are very simple minded. |
| 17:27.37 | florian_ | rhythmbox is part of gpe-image? |
| 17:27.48 | pb__ | on big-screen devices, yeah |
| 17:28.12 | benjp_ | in my experience, l2tpd was quite buggued 6 month ago |
| 17:28.21 | pb__ | ah. |
| 17:28.24 | florian_ | pb__: i use this for univerity wlan, works fine - if you ship a working ipsec.conf :-) |
| 17:28.36 | florian_ | oh cool |
| 17:28.43 | pb__ | florian_: aha :-) |
| 17:29.58 | pb__ | mm, looks like the vpn.ebootis.de stuff is what I need. |
| 17:30.25 | pb__ | I'll try that out tomorrow. |
| 17:30.27 | florian_ | :-) |
| 17:30.35 | benjp_ | else you still can do a dump vpn with openvpn (the beta 2.0 version is more windows friendly ...) |
| 17:30.42 | benjp_ | s/dump/dumb/ |
| 17:31.02 | florian_ | phew... this took ages: |
| 17:31.05 | florian_ | NOTE: package gpe-image-1.0-r0: task do_rootfs started |
| 17:32.43 | hrw | Bernardo: opie-networksettings-wlanplugin? |
| 17:33.36 | pb__ | florian_: yeah, ipkg-make-index is very slow the first time you run it. |
| 17:33.46 | pb__ | it should be faster on subsequent builds. |
| 17:34.50 | Bernardo | hrw: yes |
| 17:34.58 | hrw | Bernardo: ppp plugin also? |
| 17:35.06 | Bernardo | the first thing I install after hand configuring my wlan carf |
| 17:35.11 | Bernardo | carf=card |
| 17:35.18 | Bernardo | hrw: possibly, yes |
| 17:35.46 | florian_ | pb__: ah, ok - i started this new build yesterday. this one is for simpad... i'll try to do something for netvista now :-) |
| 17:36.23 | pb__ | cool |
| 17:36.52 | hrw | Bernardo: will push it when repo will be unlocked |
| 17:37.22 | Bernardo | ok. btw, you were looking at the locale-base-* postinst errors, right? |
| 17:37.29 | reenoo_ | florian_: netvista? |
| 17:37.48 | florian_ | reenoo_: This is an old IBM machine... |
| 17:38.17 | reenoo_ | florian_: ah, ok |
| 17:38.35 | hrw | Bernardo: nope |
| 17:39.01 | Bernardo | ok, my confusion |
| 17:39.29 | florian_ | reenoo_: a small pc (300mhz media-gx or similar) and a small dstn display. |
| 17:42.02 | florian_ | oh, the root image is 30MB... |
| 17:43.55 | pb__ | ah, maybe I added a bit too much stuff |
| 17:44.14 | florian_ | hehe |
| 17:44.36 | florian_ | pb__: what did you add? ok rhytmbox and minimo... |
| 17:44.43 | pb__ | rhythmbox, minimo and GAIM |
| 17:44.56 | florian_ | ah okay |
| 17:45.13 | pb__ | I'm not sure how much space each of those takes up. |
| 17:46.20 | pb__ | iirc, minimo is about 8MB, which I imagine makes it the largest single package. But rhythmbox probably drags in half of GNOME, so its total impact is likely to be large as well. |
| 17:46.45 | florian_ | for the Simpad SL4 we could try to us a larger browser... |
| 17:47.18 | pb__ | yeah. I did have firefox in there for a short time yesterday, but the rootfs was coming out at 34MB even without GAIM. |
| 17:47.19 | florian_ | yes, i remember it built many gnome packages... |
| 17:47.36 | florian_ | oh... |
| 17:47.47 | pb__ | maybe we should try removing rhythmbox, and swapping firefox for minimo. |
| 17:48.14 | pb__ | firefox is definitely a better browser for the SL4. |
| 17:48.37 | pb__ | it might also be possible to shrink firefox's binary size a bit. |
| 17:48.57 | chouimat | <PROTECTED> |
| 17:49.07 | florian_ | this sounds like a good idea... rhythmbox is more or less useless on a simpad |
| 17:49.15 | pb__ | yeah, true |
| 17:56.35 | pb__ | -rw-r--r-- 1 pb pb 29097984 2004-08-01 18:55 gpe-image-simpad-20040801175518.rootfs.jffs2 |
| 17:56.54 | pb__ | that's with firefox, but without gaim, rhythmbox and linphone |
| 17:58.56 | florian_ | okay... maybe should create some targets for different purposes |
| 17:59.38 | reenoo_ | would be nice if atlantis was free. drags in gnome and gtkhtml but is only ~50k on x86. |
| 18:00.18 | pb__ | florian_: yeah |
| 18:00.44 | florian_ | reenoo_: what is this? |
| 18:00.53 | reenoo_ | http://www.akcaagac.com/index_atlantis.html |
| 18:04.23 | hrw | reenoo_: wow.. gtkhtml2 is much nicer then gtkhtml1 |
| 18:06.34 | mickeyl | pb__: would this code be relevant to my problem? |
| 18:06.43 | mickeyl | #ifdef CONFIG_SA1100_BEAGLE |
| 18:06.44 | mickeyl | static void beagle_set_vpp(int vpp) |
| 18:06.44 | mickeyl | { |
| 18:06.44 | mickeyl | <PROTECTED> |
| 18:06.44 | mickeyl | <PROTECTED> |
| 18:06.44 | mickeyl | <PROTECTED> |
| 18:06.45 | mickeyl | <PROTECTED> |
| 18:06.48 | mickeyl | <PROTECTED> |
| 18:06.50 | mickeyl | <PROTECTED> |
| 18:06.52 | mickeyl | <PROTECTED> |
| 18:07.05 | pb__ | mickeyl: ah, yes, looks like it |
| 18:07.16 | pb__ | so that would suggest that Vpp is on GPIO16. |
| 18:07.49 | mickeyl | now i just need to find out which sequence of commands i give the jtag tools to set GPIO16 high |
| 18:10.44 | pb__ | right |
| 18:11.13 | mickeyl | i guess you're not familiar enough with that to give me a hint right? |
| 18:11.32 | *** join/#oe CIA-6 (~CIA@to.je.spocco.com) |
| 18:12.31 | reenoo_ | florian_: basically it's a lightweight gnome browser based around gnome-vfs and gtkhtml2. not sure how likely it is that I could talk the author into releasing it under any license acceptable for us though... |
| 18:12.44 | *** join/#oe joshin (~joshin@VDSL-130-13-185-112.PHNX.QWEST.NET) |
| 18:16.11 | pb__ | mickeyl: I don't know much about jtag-tools specifically. But it should just be a question of setting the appropriate bit in the value that gets scanned in. |
| 18:16.12 | florian_ | reenoo_: looks cool... i just mailed him... |
| 18:16.23 | pb__ | mickeyl: let me try to find the source and have a look |
| 18:16.29 | mickeyl | pb__: oh great. thanks! |
| 18:16.37 | mickeyl | openwince.sf.net |
| 18:17.04 | mickeyl | i'm pretty lost on that. the first time i bricked a PDA ever... and especially not mine :( |
| 18:17.07 | florian_ | pb__: To build gettext-native just rename /usr/bin/javac to something else :-) |
| 18:17.22 | pb__ | florian_: heh, right |
| 18:17.25 | mickeyl | bbl, short walk with wifey |
| 18:20.40 | pb__ | mickey|bbl: take a look at src/bus/sa1110.c. it looks like you need to call part_find_signal("GP16") to locate the pin, then use part_set_signal() to set it high. |
| 18:30.43 | pb__ | Robot101: the multiple-image-formats change that you requested is in now. |
| 18:30.46 | hrw | opie-aqpkg segs :( |
| 18:31.02 | CIA-6 | 03hrw 07 * r1.2415 10packages/meta/opie-collections.oe: add network settings for WLAN and PPP to opie-image |
| 18:31.04 | *** join/#oe zecke (~ich@83.169.133.125) |
| 18:31.08 | hrw | hi zecke |
| 18:31.20 | zecke | hey |
| 18:31.35 | florian_ | pb__: I have the new image on my Simpad. dbus doesn't start: |
| 18:32.17 | florian_ | oh.. in fact it does... |
| 18:32.41 | florian_ | i wonder where this error message im my log comes from |
| 18:34.44 | florian_ | pb__: did i send you the oe file for gpe-contacts 0.23? |
| 18:36.56 | Robot101 | pb__: rock |
| 18:37.07 | Robot101 | pb__: did you get anywhere with firefox? or does minimo work now? :) |
| 18:40.17 | florian_ | pb__: Looks liek you didn't check in the inittab/simpad/devices patch... |
| 18:45.20 | pb__ | florian_: oh, whoops |
| 18:45.59 | pb__ | Robot101: dunno. I was hoping Chocky would send me his error messages, but he hasn't done that. |
| 18:46.09 | *** join/#oe niklash (~niklas@h164n2fls33o828.telia.com) |
| 18:46.37 | pb__ | florian_: I don't think you did send me that version of gpe-calendar.oe. |
| 18:48.56 | pb__ | Robot101: I guess I can try it out on the SIMpad now. |
| 18:49.11 | niklash | I switched from using modprobe in busybox to using modprobe in modutils. This version works for me. |
| 18:49.47 | *** join/#oe jmau (~jmau@p5480CEDA.dip.t-dialin.net) |
| 18:50.28 | pb__ | niklash: Okay. Can you talk to the busybox people and find out if they will fix it? |
| 18:52.02 | niklash | I'll drop a mail on their list. |
| 18:52.05 | pb__ | cool |
| 18:52.30 | CIA-6 | 03pb 07 * r1.2414.1.1 10packages/initscripts/ (initscripts_1.0.oe initscripts-1.0/simpad/devices): apply patch from Florian Boor to make /dev/apm_bios symlink |
| 18:52.35 | pb__ | florian_: there it is :-) |
| 18:53.01 | pb__ | hrm, I wish the shepherd didn't have this stupid flash partitioning. |
| 18:54.46 | niklash | Maybe I have the same error florian has. I get "Cannot open APM device: No such file or directory". |
| 18:55.02 | pb__ | What platform are you on? |
| 18:55.09 | niklash | h3800. |
| 18:55.25 | pb__ | hm, yes, you might have the same issue. |
| 18:55.26 | niklash | I had to "modprobe apm" to get even that far. |
| 18:55.31 | pb__ | does /dev/misc/apm_bios exist? |
| 18:55.59 | pb__ | I wonder why apm was not loaded automatically. Is it in /etc/modules? |
| 18:56.04 | niklash | No. Where should it point? |
| 18:56.17 | pb__ | It isn't a symlink. That's meant to be the actual device. |
| 18:56.35 | niklash | Nope. |
| 18:57.00 | niklash | There doesn't seem to be a MAKEDEV script. |
| 18:57.09 | pb__ | There shouldn't be. The h3800 uses devfs. |
| 18:58.31 | niklash | But the device isn't created automatically. |
| 18:58.41 | CIA-6 | 03pb 07 * r1.2417 10packages/linux/ (2 files): add apm autoloading |
| 18:58.42 | pb__ | That would be a bug. |
| 18:58.51 | niklash | So how do I get it? |
| 18:58.56 | niklash | In what? |
| 18:58.57 | pb__ | I guess you need to fix the bug. |
| 18:58.59 | pb__ | In the kernel. |
| 18:59.25 | niklash | Isn't the kernel the same as is used in the ordinary familiar? |
| 18:59.32 | pb__ | Yes. |
| 18:59.54 | niklash | APM works there, but maybe some scripts are different. |
| 19:01.03 | pb__ | Possibly. I don't see anything obvious at a first glance. |
| 19:01.05 | *** join/#oe treke|laptop (~ggilbert@pool0338.cvx34-bradley.dialup.earthlink.net) |
| 19:01.12 | pb__ | You're sure the module is loaded correctly? Does it show up in /proc/devices? |
| 19:01.35 | niklash | By the way, h3800 needs an "alias ipaq_hal_3800 h3600_asic" in handhelds_sa*.oe. |
| 19:02.46 | niklash | No, but it does show up in /proc/modules. |
| 19:05.16 | niklash | I should have a second h3800 so I could compare familiar with OE. |
| 19:05.53 | pb__ | How about /proc/misc? |
| 19:06.19 | niklash | "134 apm_bios" |
| 19:06.40 | pb__ | that seems ok |
| 19:06.50 | pb__ | do you have anything at all in /dev/misc/? |
| 19:08.05 | niklash | Yep. If I "cd /dev; ln -s misc/apm_bios", I can then suspend. |
| 19:08.38 | pb__ | eh. |
| 19:08.52 | pb__ | <pb__> does /dev/misc/apm_bios exist? |
| 19:08.58 | pb__ | <niklash> No. Where should it point? |
| 19:09.22 | niklash | Oh, sorry. I must have missed that "/misc". |
| 19:09.44 | pb__ | Ah. Okay. In that case, yes, your issue is indeed the same as Florian's. |
| 19:10.00 | pb__ | I'll make the same change to the h3800 devices script. |
| 19:16.10 | CIA-6 | 03pb 07 * r1.2419 10packages/initscripts/initscripts-1.0/h3600/devices: apply /dev/apm_bios fix for h3600 as well |
| 19:16.10 | CIA-6 | 03pb 07 * r1.2418 10packages/initscripts/initscripts-1.0/h3900/devices: add devices script for h3900 |
| 19:17.50 | CIA-6 | 03pb 07 * r1.2420 10packages/linux/handhelds-sa_2.4.19-rmk6-pxa1-hh36.12.oe: fix ipaq_hal aliases for h3600, h3800 |
| 19:17.51 | pb__ | niklash: right, try that |
| 19:18.19 | pb__ | brb, reflashing simpad |
| 19:26.44 | florian_ | pb__: okay, thanks! |
| 19:27.13 | *** join/#oe Teletubbie (~Sam@wprt-d9bae746.pool.mediaWays.net) |
| 19:35.10 | florian_ | oh well |
| 19:35.35 | florian_ | Is there a known issue with startup-notification? I get this error: |
| 19:35.40 | florian_ | | checking whether realloc (NULL,) will work... configure: error: cannot run test program while cross compiling |
| 19:40.41 | niklash | Has anyone run mesa_6.0.1.oe on X11? |
| 19:41.13 | pb__ | florian_: odd, I haven't seen that. what platform is that on? |
| 19:41.29 | pb__ | sounds like you might be missing a site entry. |
| 19:41.36 | florian_ | pb__: i386 |
| 19:41.41 | niklash | "strings libGL.so" says it's linked to libX11, but my program says "X11 driver not configured with OpenGL". |
| 19:42.03 | pb__ | florian_: search for "startup-notification" in site/arm-linux, and copy those bits into the i386-linux file |
| 19:43.52 | florian_ | pb_: okay, thanks... if it works i'll send you a patch |
| 19:44.08 | pb_ | cool |
| 19:48.10 | hrw | heh.. my last build had to be softfloat... but opie-sysinfo says another thing.. |
| 19:52.17 | florian_ | okay... next one |
| 19:56.15 | florian_ | looks like it might be a good idea to wait with the patch until the build is complete... |
| 20:18.49 | *** join/#oe zecke (~ich@83.169.133.125) |
| 20:28.08 | florian_ | hmm.. noe matchbox-panel fails... |
| 20:30.28 | florian_ | s/noe/now |
| 20:40.22 | *** join/#oe ar_ (~ar@port-ip-213-211-231-77.reverse.mdcc-fun.de) |
| 20:41.27 | hrw | bye |
| 20:45.18 | florian_ | hmm... a broken wireless.h... strange |
| 20:47.28 | florian_ | Are there any known issues with libgcrypt? |
| 20:51.27 | CIA-6 | 03pb 07 * r1.2421 10packages/classes/ (autotools.oeclass gnome.oeclass): move gnome_stage_includes() to autotools_stage_includes() |
| 20:51.43 | pb_ | florian_: not that I know of. what kind of issues? |
| 20:52.51 | *** join/#oe CoreDump|home (~mhentges@xdsl-213-196-250-24.netcologne.de) |
| 20:53.05 | *** join/#oe joshin (~joshin@VDSL-130-13-185-112.PHNX.QWEST.NET) |
| 20:53.38 | florian_ | I get some nasty missing symbols linking the lib: |
| 20:53.42 | florian_ | | ../src/.libs/libgcrypt.so: undefined reference to `_gcry_mpih_addmul_1' |
| 20:53.42 | florian_ | | ../src/.libs/libgcrypt.so: undefined reference to `_gcry_mpih_rshift' |
| 20:53.57 | florian_ | (and some more of the same kind...) |
| 20:54.46 | florian_ | Same effect with 1.2.0... |
| 20:54.52 | pb_ | ugh. |
| 20:55.02 | pb_ | I've never seen that before. |
| 20:57.57 | florian_ | hrm... googs does not know much about this |
| 20:59.54 | florian_ | This looks like MPI stuff... |
| 21:01.37 | *** join/#oe nightfir_ (night@Sherbrooke-HSE-ppp3610802.sympatico.ca) |
| 21:02.15 | florian_ | What's going on here?! |
| 21:02.19 | florian_ | $ grep -Ir "_gcry_mpih_addmul_1" * |
| 21:02.20 | florian_ | grep: invalid option -- t |
| 21:02.34 | nightfir_ | a file in that directory probably starts with -t |
| 21:02.40 | nightfir_ | add a -- before the * |
| 21:03.09 | *** join/#oe stoyan (~stoyan@ns.burdenis.com) |
| 21:03.12 | florian_ | nightfir_: good thought... indeed |
| 21:03.30 | nightfir_ | look out for files named -rf ;) |
| 21:03.32 | stoyan | are there any plans of including mplayer in the packages tree |
| 21:04.05 | florian_ | hehe |
| 21:04.42 | nightfir_ | if anyone's interested i fixed Xqt 0.0.9's DPI setting and qcop stuff (so it can raise() properly) |
| 21:04.55 | nightfir_ | looks like the project's been abandoned |
| 21:05.44 | nightfir_ | i can't wait till pda's have enough power just to move to kde and be done with it :P |
| 21:07.03 | pb_ | nightfir_: you can already run X11 on a handheld just fine. You only need to convince the opie dudes to let go of qt/e and use qt/x11 instead. |
| 21:10.57 | nightfir_ | god that would be great |
| 21:11.29 | nightfir_ | but i mean enough horsepower for KDE itself.. the vfs, desktop, full browser, wm, etc. |
| 21:11.59 | nightfir_ | out of curiosity, why *wouldn't* they move to qt/x11? |
| 21:14.19 | florian_ | pb_: This seems to be an asm problem, if i build with --disable-asm it links. |
| 21:15.36 | *** part/#oe jmau (~jmau@p5480CEDA.dip.t-dialin.net) |
| 21:16.27 | pb_ | florian_: ah, strange. |
| 21:18.17 | florian_ | pb_: this seems not to happen on i686->arm crosscompilation, only with my i686->i386 setup. |
| 21:19.17 | niklash | florian_: If its asm, maybe the asm routines are not available for arm, so a C fallback is used. |
| 21:21.08 | florian_ | nightfir_: indeed, asm arm is missing... |
| 21:26.39 | CIA-6 | 03pb 07 * r1.2412.1.1 10packages/classes/gconf.oeclass: add gconf.oeclass |
| 21:30.26 | *** part/#oe Rustre (~Rustre@lns-th2-7-82-64-109-180.adsl.proxad.net) |
| 21:33.55 | mickeyl | nightfir_: please send that patch to opie-devel@handhelds.org, we perhaps can use it |
| 21:40.44 | *** join/#oe CoreDump_ (~mhentges@xdsl-213-196-241-18.netcologne.de) |
| 21:45.08 | *** join/#oe CIA-6 (~CIA@to.je.spocco.com) |
| 21:48.01 | florian_ | pb_: would you please copy gnome/libart/arm-linux to gnome/libart/i386-linux in your oe repository? |
| 21:48.07 | mickeyl | pb_: thanks for that pointer! i think i'll be able to go on from here. Will add that tomorrow and keep you posted |
| 21:49.29 | mickeyl | pb_: oh just one last question while we're there.... in http://sourceforge.net/mailarchive/forum.php?thread_id=3191245&forum_id=12457 there is someone who has the same problem with flashing an ipaq 3600. It works after he issues "poke 0x49000000 1". is that also a method for setting the egpio on ipaq which you talked about? |
| 21:49.44 | pb_ | florian_: your wish is my command |
| 21:49.53 | florian_ | pb_: wait a second! |
| 21:49.54 | pb_ | mickeyl: yep. egpio is at 0x49000000 in cpu address space, so that poke will set bit 1 |
| 21:49.58 | florian_ | hehe |
| 21:50.11 | pb_ | florian_: ah, too late. I already committed the change |
| 21:50.15 | CIA-6 | 03pb 07 * r1.2421.1.1 10packages/gnome/libart-lgpl/i386-linux/art_config.h: copied from arm version |
| 21:50.51 | florian_ | pb_: for some reason it is "i386-oe-linux" |
| 21:50.53 | pb_ | mickeyl: er, bit 0, I mean |
| 21:50.55 | pb_ | florian_: ah |
| 21:51.06 | pb_ | I guess you have a TARGET_VENDOR set. |
| 21:51.16 | mickeyl | pb_: aha. cool. i understand that bit now. so on ipaq the egpios are mapped into cpu space which makes that easier. is that an ipaq specific thing or are all (e)gpio mapped in cpu address space? |
| 21:51.21 | florian_ | ehhh... |
| 21:52.29 | pb_ | mickeyl: they're all mapped. on-chip GPIOs are controlled by internal registers in the SA-1100, which show up in the SFR space somewhere. The EGPIO is just an external latch. |
| 21:52.54 | pb_ | mickeyl: however, in JTAG terms, the way you access them is different. you can't write to internal registers using boundary scan, but you can flip the pin state directly. |
| 21:53.08 | pb_ | florian_: okay. I'll rename the file. |
| 21:53.44 | florian_ | pb_: if there is a chance to have i386-linux as well it might be a good ide to add both. |
| 21:54.01 | joshin | <PROTECTED> |
| 21:54.09 | joshin | oops. Sorry |
| 21:54.40 | mickeyl | pb_: ah. i guess since the mapping is probably different from board to board there's no direct method to find out which poke address would correspond to a certain GPIO |
| 21:55.04 | pb_ | mickeyl: the onboard GPIOs are always at the same addresses. |
| 21:55.30 | CIA-6 | 03pb 07 * r1.2424 10packages/gnome/libart-lgpl/ (i386-linux/art_config.h i386-oe-linux/art_config.h): copy i386-linux art_config.h to i386-oe-linux |
| 21:56.20 | pb_ | mickeyl: the GPIO data register is at 0x90040004, and each bit corresponds to one gpio pin. |
| 21:56.39 | pb_ | but, of course, the usage of GPIOs varies from one board to another. |
| 21:56.53 | florian_ | pb_: for some strange reason i had a broken wireless.h in my netvista staging directory. |
| 21:57.08 | pb_ | odd |
| 21:57.23 | mickeyl | florian_: broken in how far ? |
| 21:57.36 | pb_ | florian_: you might want to make tactful enquiries of mickey. he has been changing some stuff with wireless.h fairly recently. |
| 21:58.36 | florian_ | mickeyl: In the struct iw_point the member "pointer" had the type "void __user". |
| 21:58.47 | mickeyl | ah |
| 21:58.52 | mickeyl | that's how it is |
| 21:58.58 | pb_ | argh, I need to use bk citool again. |
| 21:59.08 | mickeyl | i think there has been something missing in the hh.org kernel port |
| 21:59.12 | florian_ | mickeyl: on x86 this should be caddr_t imho... at least on my pc it is. |
| 21:59.26 | mickeyl | treke stumbled over that, too. |
| 21:59.40 | mickeyl | it compiles fine on kernel 2.6.x on x86, but not on a hh.org 2.6 |
| 21:59.49 | mickeyl | so i think there's a header file not up to date or so |
| 22:00.21 | florian_ | pb_: you see, i'm working on it :-) |
| 22:00.28 | pb_ | florian_: :-) |
| 22:00.45 | mickeyl | the only thing i did btw. is to remove the glibc provided wireless.h and instead link to the kernel provided one. |
| 22:00.48 | florian_ | mickeyl: "on" means the host system or the target environment? |
| 22:01.06 | mickeyl | florian_: target |
| 22:01.28 | CIA-6 | 03pb 07 * r1.2422.1.1 10packages/classes/gnome.oeclass: gnome.oeclass implies gconf |
| 22:01.42 | florian_ | mickeyl: hmm... i have 2.6.7 as my target... |
| 22:01.52 | mickeyl | florian_: hh.org 2.6.7 ? |
| 22:02.07 | florian_ | mickeyl: stock |
| 22:02.13 | mickeyl | oh |
| 22:02.16 | mickeyl | interesting |
| 22:02.34 | florian_ | i build i686->i386 |
| 22:02.40 | CIA-6 | 03pb 07 * r1.2426 10packages/ (4 files in 4 dirs): |
| 22:02.40 | CIA-6 | gnome-games_2.6.1.oe: |
| 22:02.40 | CIA-6 | <PROTECTED> |
| 22:02.40 | CIA-6 | rhythmbox_0.8.5.oe, gnumeric_1.3.1.oe, galculator_1.2.3.oe: |
| 22:02.40 | CIA-6 | <PROTECTED> |
| 22:02.51 | mickeyl | i see. then kernel 2.6.7 changed something which is not compatible with the rest of the glibc headers |
| 22:03.01 | florian_ | oh |
| 22:03.04 | florian_ | nasty |
| 22:09.01 | florian_ | good night all |
| 22:09.20 | CIA-6 | 03pb 07 * r1.2427 10packages/gstreamer/gst-plugins_0.8.1.oe: inherit gconf |
| 22:09.32 | mickeyl | g'night florian_ |
| 22:18.02 | CIA-6 | 03pb 07 * r1.2428 10packages/gnome/gnome-vfs_2.6.0.oe: use gnome.oeclass |
| 22:18.35 | Pendalar | offtopic I guess, but does anyone know anything about putting linux on a Asus A620 |
| 22:21.21 | CIA-6 | 03pb 07 * r1.2430 10packages/gnome/libgnome_2.6.1.1.oe: use gnome.oeclass |
| 22:21.21 | CIA-6 | 03pb 07 * r1.2429 10packages/libgcrypt/libgcrypt_1.1.91.oe: break out gcrypt-pthread bits |
| 22:23.26 | CIA-6 | 03pb 07 * r1.2424.1.1 10packages/site/i386-linux: add missing i386-linux site bits, courtesy Florian Boor |
| 22:23.54 | Pendalar | I read on the handhelds.org moin but I was wondering how far away from being able to use it for a full unit |
| 22:25.05 | Pendalar | it sounds like you can get into a console and do some minor things |
| 22:37.56 | CIA-6 | 03pb 07 * r1.2432 10packages/gnome-themes/gnome-themes_2.6.2.oe: add depends on intltool-native |
| 22:44.25 | CIA-6 | 03ral 07 * r1.2422.2.1 10packages/ (3 files in 2 dirs): keyz-config and zkbapplet made to one project |
| 22:45.28 | pb_ | joshua_: good morning |
| 22:45.43 | joshua_ | morning pb_ |
| 22:46.15 | joshua_ | I guess I should stop using the _Sleep tag. |
| 22:46.34 | joshua_ | pb_: what did I miss in the wold while I was away at camp? |
| 22:47.00 | joshua_ | I guess, more relevantly, in the handhelds.org world. |
| 22:48.10 | pb_ | hm, good question. I guess you've been away for a while, but not all that much has happened. I've been focussing most of my energies on OE of late (as you might expect me to say in this channel). |
| 22:48.30 | joshua_ | indeed. |
| 22:48.42 | pb_ | ipaq-wise, it seems that datasheets are available for the asic3 sd{io} cores, and Brad Campbell is planning to write a driver. so, rock. |
| 22:48.49 | joshua_ | cool. |
| 22:48.57 | pb_ | and of course there's this massive slew of new ipaqs, none of which we have ports for yet. |
| 22:49.13 | joshua_ | last night I got word that aquadran managed to pop open the rom format for h1910. |
| 22:49.20 | pb_ | cool |
| 22:49.34 | joshua_ | so, I've got it all dumped out at handhelds.org/~joshua |
| 22:49.39 | joshua_ | which new iPAQs? |
| 22:50.25 | pb_ | I can't remember the model numbers. There's a whole set of them. |
| 22:50.38 | joshin | Quick question. To build a 64M Memory & 0M Ramdisk kernel, are the following lines correct to add to the local.conf? |
| 22:50.41 | joshin | COLLIE_MEMORY_SIZE=64 |
| 22:50.41 | joshin | COLLIE_RAMDISK_SIZE=0 |
| 22:50.42 | joshua_ | 1715 I see.. 6315 phone.. 4705.. |
| 22:50.50 | mickeyl | looks about right |
| 22:50.53 | pb_ | yah, those sound like the ones. |
| 22:50.54 | joshin | Thanks! |
| 22:51.08 | joshua_ | excellent. yet another s3c2410. thanks HP |
| 22:51.11 | pb_ | the 6315 is OMAP-based. |
| 22:51.26 | pb_ | 1715 is s3c2410? cute. |
| 22:51.27 | joshua_ | so I see. |
| 22:51.48 | joshua_ | beautiful. OMAP uses arm isa or no? |
| 22:51.55 | pb_ | yes |
| 22:52.05 | pb_ | OMAP is an ARM9 (in that incarnation) with a DSP strapped on the side. |
| 22:52.17 | joshua_ | aha |
| 22:52.26 | joshua_ | the 4705 looks like it's going to really scream |
| 22:52.39 | pb_ | is that the pxa270 one? |
| 22:52.57 | joshua_ | looks like it |
| 22:52.57 | joshua_ | 624mhz clock |
| 22:52.57 | pb_ | cool |
| 22:53.11 | joshua_ | have to see if it's going to outperform my epia ;) |
| 22:53.39 | pb_ | heh |
| 22:54.07 | joshua_ | did I show you the photo of my epia project? |
| 22:54.12 | pb_ | no, I don't think so |
| 22:54.14 | joshua_ | http://paul.bluereboot.net/monstertruck.jpg |
| 22:54.34 | joshua_ | woah! |
| 22:54.42 | joshua_ | 4705 appears to be 640x480 |
| 22:54.50 | pb_ | yeah |
| 22:55.09 | joshua_ | only 64mbsdram, though. |
| 22:55.37 | pb_ | mm, too bad |
| 22:56.11 | joshua_ | looks like it's going to be pretty competitive with the zauri, considering that you can actually get it someplace other than japan. |
| 22:56.21 | pb_ | yup |
| 22:56.27 | joshua_ | also looks like they're going to have a 3600mah extended battery for it, which will be pretty cool. |
| 22:56.29 | pb_ | and faster, obviously |
| 22:56.44 | pb_ | btw, I started porting LAB to one of our s3c2410 boards this morning. |
| 22:56.57 | joshua_ | sounds cool. |
| 22:57.41 | pb_ | didn't get very far with it yet, though. I need to debug the kernel and/or hardware a bit more, I think. |
| 22:59.18 | joshua_ | ah. |
| 22:59.47 | joshua_ | I'm considering making lab a linuxbios payload and seeing if I cang et that going on monstertruck to get the boot time down |
| 23:00.37 | pb_ | sounds like a good plan |
| 23:05.16 | joshua_ | I aim to get it to have a nice quick reboot time so that if it rolls or something, it'll go back to doing what it was doing within 15sec or so (see photo) |
| 23:07.20 | *** join/#oe Timelord0 (~TL@65.59.240.28) |
| 23:16.21 | *** part/#oe Teletubbie (~Sam@wprt-d9bae746.pool.mediaWays.net) |
| 23:35.50 | CIA-6 | 03pb 07 * r1.2432.1.1 10packages/glibc/glibc-package.oe: create ${libdir}/locale |
| 23:44.12 | CIA-6 | 03pb 07 * r1.2435 10packages/glibc/glibc-package.oe: add more locale bashing bits to work around lack of mmap(PROT_WRITE, MAP_SHARED) on jffs2 |
| 23:44.13 | CIA-6 | 03pb 07 * r1.2436 10packages/glibc/ (glibc_2.3.2+cvs20040726.oe glibc_2.3.2.oe glibc_cvs.oe): bump PR |
| 23:52.27 | *** join/#oe treke|laptop (~ggilbert@pool0029.cvx29-bradley.dialup.earthlink.net) |
| 23:52.49 | *** join/#oe BillShovel (~BillShove@192.111.105.96) |
