**** BEGIN LOGGING AT Sat Jun 18 23:59:56 2005 Jun 19 01:05:02 jacques: Nothing changed that I know of - serial still works fine for me... Jun 19 01:10:41 RP, you still get kernel boot messages ? Jun 19 01:11:10 we don't, but serial console works after boot Jun 19 01:12:02 RP, also, how were you measuring your performance increases? I'd like to see if we got any Jun 19 01:18:12 jacques: The bootup time is much faster and the device just feels generally quicker Jun 19 01:18:22 jacques: Yes, I still get kernel boot messages Jun 19 01:26:58 Morning Jun 19 01:27:16 RP: Something to test ? Jun 19 01:59:02 uv1: -rc6-mm1 won't hang 100% of the time now. I've not looked at the key repeat problem yet though Jun 19 02:25:24 RP: Will waiting for the key repeat problem to be fixed ;-) Or would you like me to test the latest kernel. Jun 19 02:46:57 morning Jun 19 02:51:09 Great, I've totally messed up the w100 driver. Jun 19 02:51:45 * SirFred kicks himself on the head with the husky battery Jun 19 02:54:43 err s/kicks/hits Jun 19 03:19:23 hi reenoo_ Jun 19 03:20:06 hey pb_ Jun 19 03:34:53 hi, whats the future plans with bitkeeper and oe development? Jun 19 03:35:55 morning zecke Jun 19 03:36:32 hey Jun 19 03:46:04 wb bluelightning Jun 19 03:46:09 hi zecke Jun 19 03:46:13 hi all Jun 19 03:50:30 mickey|tv: GStreamer looks good Jun 19 03:50:38 mickey|tv: it is written in C though Jun 19 03:51:17 mickey|tv: the Balzert according to a HP publication claims that even writing in Assembly is more productive than C ;) Jun 19 03:51:34 hah Jun 19 03:52:28 pb_: no kidding Jun 19 03:52:53 pb_: it is a book about Software Management, but it does not try to make a point against using C Jun 19 03:53:29 pb_: it even says that this publication is not representative Jun 19 03:54:05 I see. sounds interesting. Jun 19 03:58:27 OT: Is there an easy tool to test context free grammars? Jun 19 04:01:49 pb_: any progress towards a decision on the replacement for bitkeeper for the OpenEmbedded repository? We will be setting up a new machine soon, and it will be the perfect opportunity for us to host the nslu2-linux clone on that machine if the chosen system groks clones, otherwise we need to talk about whether to continue being an offshoot of OE, or whether our 10 or so key developers apply for w Jun 19 04:01:49 rite access to the main OE repo ... Jun 19 04:01:57 pb_: run or you get abused Jun 19 04:02:08 pb_: how is activation handled with d-bus nowadays? Jun 19 04:11:05 anyone else know about the future of source code management for the OE repository once bitkeeper goes away next month? Jun 19 04:15:15 I tried searching the oe@handhelds.org mailing list, but it seems that the hypermail archives of that at http://www.handhelds.org/hypermail/oe/index.html have not been updated since 31 Oct 2004 ... Jun 19 04:15:42 rwhitby: we do not know yet Jun 19 04:15:54 gmane is now hosting the mailarchive Jun 19 04:15:55 rwhitby: there is nothing that prevents your developer to have direct access Jun 19 04:16:46 zecke: understood. we have up until now had a bk clone, and kergoth would vet any non nslu2-linux-specific changes before he pulled them into OE proper. Jun 19 04:17:45 but I think that our core developers who have write access to the nslu2-linux bk repo now have enough of an undertanding of OE to be given write access to the main OE repo and not stuff things up. Jun 19 04:17:47 rwhitby: we will review the changes as well Jun 19 04:17:56 of course :-) Jun 19 04:18:18 rwhitby: I'm pretty sure our next SCM will not be bk Jun 19 04:18:39 nod. that's why we're interested. Any leading contenders? Jun 19 04:18:51 rwhitby: I guess cogito or monotone Jun 19 04:18:59 rwhitby: I'm pretty sure it won't be svn or cvs Jun 19 04:19:37 rwhitby: maybe we start without history and in this case I would love to run cvs2svn on the stuff we get off bkbits Jun 19 04:19:39 cogito would be very interesting :-) we've already created an ipkg for it for the nslu2 Jun 19 04:20:09 rwhitby: from what I know cogito is not too good for our usage Jun 19 04:20:18 rwhitby: basicly only one can 'push' Jun 19 04:20:54 ah, so it's tuned to Linus's way of working ... Jun 19 04:21:26 so it relies on the consumer pulling from many producers? Jun 19 04:22:12 rwhitby: yupp Jun 19 04:22:41 sounds like I need to read up on monotone then .... Jun 19 04:28:26 ok, so after the first 5 minutes of reading up on monotone, it looks like we could continue to have an nslu2-linux "server" (to which all our authorised nslu2-linux developers push stuff), and then have a couple of nslu2-linux core developers have the authority to push changes to the main OE "server". Jun 19 04:29:12 unless the OE folk would prefer that our 10 or so authorised nslu2-linux firmware developers push directly to the main OE "server" ? Jun 19 04:30:15 I do not know what is preferred by mickeyl,pb_,kergoth Jun 19 04:30:19 rwhitby: but that would work Jun 19 04:31:02 we're happy working either way. whichever is preferred by the OE core team would be what we would do. Jun 19 04:34:05 in a sense, if the nslu2-linux people had write access to our main tree, we could probably get away with svn. Jun 19 04:34:16 afaik, nobody else is actually making use of the peer-to-peer feature of bitkeeper. Jun 19 04:34:39 on the other hand, I know that kergoth is very fond of that feature, and he probably wouldn't want to give it up. Jun 19 04:35:26 zecke: regarding your dbus question, I'm not sure. I know that the Imendio people were doing some work on activation, but I have no idea on the current status. Jun 19 04:37:37 I agree that monotone does seem like the most promising SCM, though. Jun 19 04:37:47 zecke, pb_: thanks for the info. pb_: I personally would be happy enough with svn (or monotone, from what I've read so far) Jun 19 04:37:56 x Jun 19 04:38:27 some of our nslu2-linux developers have two local repos and use the bk peer-to-peer feature to manage local diversions too. Jun 19 04:38:36 right Jun 19 04:38:47 so monotone looks like it would be better for that to be able to continue Jun 19 04:38:58 maybe we can sweet-talk treke into setting up an experimental monotone repository on oesources.org Jun 19 04:43:50 zecke: heh, you should have signed up to work on improving monotone for the summer of code. Jun 19 04:48:31 i try to build OE from the latest snapshot and i got an error while unpacking quilt-native, any ideas whats be wrong here? Jun 19 04:49:03 was it just a general kind of error, or anything specific? Jun 19 04:49:25 NOTE: package quilt-native-0.39-r0: task do_fetch: started Jun 19 04:49:25 NOTE: package quilt-native-0.39-r0: task do_fetch: completed Jun 19 04:49:29 but there is no package there Jun 19 04:49:40 so i get the error while unpacking the not available tar.gz file Jun 19 04:51:34 that does seem strange Jun 19 04:51:58 do you have cvs installed? Jun 19 04:52:06 yes Jun 19 04:53:18 i will run bb -D, perhaps i will get more informations on that Jun 19 04:53:37 hmm, no Jun 19 04:54:42 whats the difference between the native and not-native packages? i want to crosscompile, does native mean to compile directly at the target? Jun 19 04:55:34 native means that the package is compiled to run on your build host. Jun 19 04:55:41 ok Jun 19 04:59:08 hmm, something wrong, i can do a fetch quilt without any errors but i cannot do an unpack quilt Jun 19 05:06:59 morning Jun 19 05:07:22 morning mickeyl Jun 19 05:07:35 moin mickeyl Jun 19 05:07:41 fwiw, i have worked with monotone for some days and my vote is on that. i will setup a preliminary repository next week and import oe into that and then add some keys Jun 19 05:08:05 i hope it works out Jun 19 05:08:21 if not, well, it's open, we can always import the history into the next big thing then Jun 19 05:09:47 although i think it'd be best to continue the seperate branch methodology for nslu2, we don't have problem with giving your people access to ours as well. Jun 19 05:10:00 it's just a matter of what's the most productive and sane setup Jun 19 05:11:20 mickeyl: could you confirm #103 is fixed? Jun 19 05:11:22 ~praise zecke for fixing #103 Jun 19 05:11:24 All hail zecke for fixing #103! Jun 19 05:11:24 heh Jun 19 05:11:25 hehe Jun 19 05:11:48 * mickeyl reads the diff Jun 19 05:12:00 mickeyl: better don't do it Jun 19 05:12:19 * zecke throws the coin to decide to do the assignment or to look discovery channel Jun 19 05:12:45 heh, nice Jun 19 05:12:57 shorter than i've thought Jun 19 05:13:07 mickeyl: excellent Jun 19 05:13:11 (and good morning) Jun 19 05:13:33 damn orinoco and mipv6 is broken with hh-sa.26 Jun 19 05:13:51 oh dear, that sucks. what's wrong? Jun 19 05:14:22 (are you saying that orinoco and mipl are both individually broken, or that they don't work together?) Jun 19 05:14:33 pb_: both are not buildable Jun 19 05:14:45 (might be a problem with the simpad defconfig) Jun 19 05:14:55 but my XDMCP session is virtually dead... Jun 19 05:14:55 ah, now I come to think of it, I remember that the mipl folk are still on kernel 2.6.8.1 Jun 19 05:15:08 at least, they were last time I looked Jun 19 05:15:24 I don't know what would be wrong with orinoco, though. I guess that would be a task for mickeyl. Jun 19 05:15:26 *** Wireless extension not available Jun 19 05:15:41 so my defconfig is likely to be broken Jun 19 05:15:58 could the error be only a problem with the latest snapshot? Jun 19 05:16:02 zecke: sounds like it Jun 19 05:16:02 which upstream version corresponds to hh-sa atm.? Jun 19 05:16:07 IFA_F_HOMEADDR undeclared Jun 19 05:16:07 morning RP Jun 19 05:16:07 and Jun 19 05:16:11 morning all Jun 19 05:16:11 ~hail RP for fixing rmk's bugs Jun 19 05:16:13 * ibot bows down to RP for fixing rmk's bugs and chants, "I'M NOT WORTHY!!" Jun 19 05:16:37 mickeyl: Did you see his post to LKML? He used grumble tags! Jun 19 05:17:00 RP: hah. no, not yet. from when is that? Jun 19 05:17:07 mickeyl: 2.6.11 Jun 19 05:17:08 i only read lkml twice a week Jun 19 05:17:22 heh, I only read lkml twice a year Jun 19 05:18:01 well i only glimpse through the headers until i find something that interests me. I'm glad that kernel traffic exists Jun 19 05:18:02 mickeyl: about midnight last night Jun 19 05:18:44 aaah Jun 19 05:18:44 zecke: I'm not sure what the holdup is in porting the mipl bits to a newer kernel. might be worth checking the usagi mailing lists or website for clues Jun 19 05:18:53 oh 2.6.12 is out? Jun 19 05:18:55 RP: now that's funny. i never have seen him using that Jun 19 05:18:56 cfetch is not working too, it seems the buildsystem is not downloading the sources but tells that the sources were downloaded Jun 19 05:19:02 zecke: yeah Jun 19 05:19:07 RP: i think that's more grumbling against himself Jun 19 05:20:32 mickeyl: I'm not so sure about that. I took it to mean, "why didn't you do this before now?"... Jun 19 05:20:41 Anyway, he's provided a fix which i'll test Jun 19 05:22:04 * mickeyl flashes linux-openzaurus-2.6.12-rc6-mm1-r3 Jun 19 05:22:09 At least the problem was in -mm1 so 2.6.12 should be Zaurus friendly :) Jun 19 05:22:32 mickeyl: You might be interested to try the oprofile code - its quite interesting Jun 19 05:23:02 The only problem is you need unstriped binaries and oe doesn't seem keen to produce them in all cases, even when you set the debug flags Jun 19 05:23:08 RP: can you sum up what oprofile is doing? Jun 19 05:26:02 mickeyl: Have a look at http://www.rpsys.net/openzaurus/temp/op_sample.txt.gz Jun 19 05:26:19 woot Jun 19 05:26:22 that's handy! Jun 19 05:26:32 It gives you a percentage breakdown of which functions the processor spent time in Jun 19 05:26:37 Both user and kernel space Jun 19 05:27:02 and thanks to my latest patch will tell you which functions are calling what percentage of the time Jun 19 05:27:12 very cool Jun 19 05:27:26 thats exactly the same error like my one http://pastebin.ca/13163 Jun 19 05:27:29 need to run opie against that Jun 19 05:27:49 mickeyl: I thought you might find it interesting for opie :) Jun 19 05:27:59 mr_claus: echo 'INHERIT += patcher' >> conf/local.conf Jun 19 05:28:05 mr_claus: then rebuild and forget about quilt Jun 19 05:28:14 RP: defintly, yeah Jun 19 05:28:19 definitly, as well Jun 19 05:28:41 mickeyl: You need oprofile_cvs.bb as I've had them introduce some fixes for arm Jun 19 05:29:02 mickeyl: and the default zaurus kernel doesn't include oprofile by default - the patches are all there though Jun 19 05:29:11 right, i see Jun 19 05:29:18 mickeyl: how would that fix such corruption? Jun 19 05:29:39 add CONFIG_PROFILING and CONFIG_OPROFILE to the defconfig if IIRC... Jun 19 05:29:56 mickeyl: it downloads galculator in quilt-native's do_fetch... Jun 19 05:30:09 reenoo_: oh right. Jun 19 05:30:12 in that case Jun 19 05:30:19 update to latest bitbake and rm -rf tmp/cache Jun 19 05:30:25 reenoo_: use older version of bitbake for older oe trees? Jun 19 05:30:26 then rebuild Jun 19 05:30:49 zecke: dunno. I'm just looking at that pastebin posting Jun 19 05:31:19 reenoo_: if that happen classes/base.bbclass is broken Jun 19 05:34:49 i'm using the latest snapshot and bitbake 1.3.0, will that work or do i need an older bitbake version? Jun 19 05:35:53 mr_claus: what oe tree do you've Jun 19 05:36:16 good afternoon all Jun 19 05:36:17 zecke: the latest snapshot and i try to build openslug Jun 19 05:36:22 hey koen Jun 19 05:36:31 hey mickeyl Jun 19 05:36:43 03mickeyl 07 * r1.3564 10openembedded/conf/distro/openzaurus-3.5.4.conf: oz-3.5.4.conf: bump cvsdate Jun 19 05:36:56 mickeyl: can we select a few new names and remove the others from the new distro wiki? Jun 19 05:37:05 koen: yeah Jun 19 05:37:20 koen: how about every one of us selects 4 names ? Jun 19 05:37:32 or 3? Jun 19 05:39:55 mr_claus: how does line 327 look in openembedded/classes/base.bbclass Jun 19 05:40:19 mickeyl: move the names you like to the top table Jun 19 05:40:36 let us vote who is eligible to vote ;) Jun 19 05:40:42 koen: ok, will do Jun 19 05:40:51 zecke: bb.fetch.init(src_uri.split()) Jun 19 05:41:22 mickeyl: and we'll have to compose a list of eligible voters Jun 19 05:41:27 mr_claus: ok you have an at least one month old oe tree Jun 19 05:41:59 zecke: yes, the latest snapshot, i have to bitkeeper version to get the newest sources Jun 19 05:42:07 mr_claus: http://openembedded.bkbits.net:8080/openembedded/hist/classes/base.bbclass?nav=index.html|src/.|src/classes Jun 19 05:42:21 mr_claus: do not use old oe with brand new bitbake Jun 19 05:43:30 zecke: ok, i would use the newest sources of OE but how i can get it without bitkeeper? Jun 19 05:43:55 mr_claus: using the free client? Jun 19 05:43:57 with the opensource bk client Jun 19 05:45:37 hah Jun 19 05:45:39 koen: cool Jun 19 05:45:45 koen: we agree in 2 names Jun 19 05:45:49 :) Jun 19 05:45:50 that means i only take 2 more Jun 19 05:46:21 chose an odd number of voters :) Jun 19 05:46:30 for sure ;) Jun 19 05:47:24 hmm Jun 19 05:47:30 the rest of the names kind of suck Jun 19 05:47:37 * mickeyl grins Jun 19 05:47:57 if we choose openhand, mallum will sue us :) Jun 19 05:48:04 hehe, yeah Jun 19 05:48:20 what are you voting for? Jun 19 05:48:24 ok, i'll take one more Jun 19 05:48:53 Crofton|laptop: new name for our merged super distro Jun 19 05:49:06 heh Jun 19 05:49:34 People ask me if I am using familiar on my OSK Jun 19 05:49:39 I already have a tagline for our first release: "No more ROMs" Jun 19 05:49:44 It is a confusing answer .... Jun 19 05:50:01 ok Jun 19 05:50:06 What are the names that are being voted on? Jun 19 05:50:19 ok, done Jun 19 05:50:29 BigAl: http://www.handhelds.org/moin/moin.cgi/NewDistroNames Jun 19 05:50:31 lo, isnt0 the name of zImage-poodle-pxa255 a bit confusing considering that my poodle says to have a pxa250 Jun 19 05:50:41 darkschneider: that's just your poodle Jun 19 05:50:48 there are poodles with pxa255 Jun 19 05:50:51 and this is their kernel Jun 19 05:50:55 ok then i'm stupid Jun 19 05:50:58 nm :) Jun 19 05:51:02 let me re read Jun 19 05:51:24 moin Jun 19 05:51:29 moin CoreDump|home Jun 19 05:51:37 hey CoreDump|home Jun 19 05:51:42 the answer came fromt eh fact that the kernel says he is not the rigth kernel Jun 19 05:51:53 when i shutdown the thing Jun 19 05:51:55 hey mickeyl & koen. Jun 19 05:52:52 mmmm in "system info" i still read PXA250 rev4(v5l) that where my confusion come from Jun 19 05:53:18 mickeyl: you may want to have a look at the latest altboot. It now implements a timeout like you suggested. Jun 19 05:53:34 CoreDump|home: excellent. will do Jun 19 05:53:40 darkschneider: yes, of course Jun 19 05:53:44 darkschneider: reread what i told you Jun 19 05:53:49 _your_ poodle has pxa250 Jun 19 05:53:53 there are other poodles around Jun 19 05:53:55 some have pxa255 Jun 19 05:53:57 ahh oki Jun 19 05:54:11 i thouth it was sarcastic.. sorry Jun 19 05:54:13 lol Jun 19 05:54:15 heh Jun 19 05:54:22 ~lart pxa250 for existing Jun 19 05:54:23 * ibot whips out a sword and chops pxa250 in half for existing Jun 19 05:54:26 I'm never sarcastic Jun 19 05:54:27 ok then what kernel i shoudl use? Jun 19 05:54:29 ;) Jun 19 05:54:35 darkschneider: please use the slow one Jun 19 05:54:43 there are all kinds of problems that can occur with the faster one Jun 19 05:54:52 s/slow/appropriate/ Jun 19 05:54:58 s/faster/cache workaround disabled/ Jun 19 05:55:04 ahh oki Jun 19 05:55:28 so i reflash it now Jun 19 05:55:29 random memory corruption being one of them Jun 19 05:55:47 `lart pxa250 again, just because Jun 19 05:55:48 mickeyl: nod on the continuing the separate branch methodology for nslu2-linux. we will request to have a couple of core developers who have a long history of sane contributions to OE (via nslu2-linux) to have direct access (so that we can push batches of changes, but also so we can push individual changes when necessary too). Jun 19 05:56:30 rwhitby: right. Jun 19 06:01:31 much better now with the newest OE sources :) Jun 19 06:03:13 mickeyl: cool, thanks. we will set up an nslu2-linux monotone server to be ready to sync with your new server next week. Can I ask that you leave a message in #nslu2-linux when it's set up, or contact me directly? Jun 19 06:04:28 rwhitby: got your email. i'll contact you when it's ready. Jun 19 06:04:43 ah, we've decided on monotone? Jun 19 06:04:56 koen: decided would be too strong Jun 19 06:05:07 it's my recommendation and i'm willing to do some work to set it up Jun 19 06:05:15 mickeyl: thanks. very much appreciated. and know that OpenEmbedded as a basis for nslu2-linux firmware is also very much appreciated in general. Jun 19 06:05:24 rwhitby :) Jun 19 06:05:40 BTW, is there somewhere where we can make a small donation to OpenEmbedded? Jun 19 06:05:49 koen: i'll set it up next week and we can start learning a bit how to work with it. Jun 19 06:05:55 mickeyl: cool Jun 19 06:06:40 (we're currently running a donation drive in nslu2-linux to buy some hardware, and anything left over is usually donated to other projects (like freenode.org, and CIA) on which we depend) Jun 19 06:06:40 rwhitby: we don't have a group account for various reasons. the main developers receive donations via paypal. see http://oe.handhelds.org/contribution.php and http://openzaurus.org/wordpress/developers/ Jun 19 06:08:39 ok, well we've already sent kergoth an nslu2 a while ago. Can we interest you in one (assuming we get enough to cover that cost) ? Jun 19 06:09:20 rwhitby: for sure. i have been interested in one for quite a while now Jun 19 06:09:31 rebooting Jun 19 06:09:40 mickeyl: ok, I'll see what I can do. Jun 19 06:09:59 rwhitby: excellent. thanks. Jun 19 06:11:57 did anyone else notice that the "buzzer" doesn't work for poodle in recent builds? Jun 19 06:13:08 or does the poodle buzzer require special handling like an additional tool or some driver? Jun 19 06:21:22 CoreDump|home: i just noticed it :) Jun 19 06:23:16 iirc the poodle buzzer is a dsp device Jun 19 06:23:29 hence it should work via /dev/dsp Jun 19 06:23:37 do you mean the opie system sounds or at all? Jun 19 06:24:25 mickeyl: you've been added to the list: http://www.nslu2-linux.org/wiki/Info/Phase3List Jun 19 06:25:28 rwhitby: ah, very good. Jun 19 06:25:33 (that's what we're buying with the proceeds of our third donation drive) Jun 19 06:25:54 http://www.nslu2-linux.org/wiki/Info/DonationList is what we have bought previously Jun 19 06:26:02 (see the bottom of that page) Jun 19 06:26:15 darkschneider: could you try cat'ting a .wav file into /dev/dsp? Jun 19 06:26:35 sure Jun 19 06:26:51 rwhitby: ah that's pretty good for such a niche project. Jun 19 06:26:54 have a wav under hand? or i'll rip a cd Jun 19 06:27:20 mickeyl: niche? 4000 subscribers on the mailing list and 20000 firmware downloads :-) Jun 19 06:27:38 CoreDump|home: allready done, coping it, hold on Jun 19 06:27:44 rwhitby: woah. so much for a niche. that's probably much more than the amount of OZ users Jun 19 06:27:46 * koen should get one of those nslu2's someday Jun 19 06:28:43 we're having to set up ipkg mirrors right now, cause we're getting 100GB/month package downloads. Jun 19 06:29:11 (and it's increasing at the rate of 20GB/month every month) Jun 19 06:29:22 hmm you have a pretty good wiki, btw. Jun 19 06:29:24 congrats Jun 19 06:29:37 but hopefully will level off ... Jun 19 06:29:49 the OZ3.5.3 moved about 150GB in total in that month Jun 19 06:30:04 thanks on the wiki. we have a lot of very helpful community members. Jun 19 06:30:24 OZ3.5.2 release i mean Jun 19 06:30:24 rwhitby: yeah. it looks like there's not such a big gap between users and developers like it's in the PDA scene Jun 19 06:31:39 mickeyl: we believe the combination of Unslung (stays compatible with stock firmware, but adds packages) and OpenSlug (latest linux kernel and OE base system) covers a wide range of users/developers. Jun 19 06:32:40 CoreDump|home: it actually produce sound Jun 19 06:33:01 so opie appears to be bugged then Jun 19 06:33:01 rwhitby: that's a good strategy, for sure. Jun 19 06:33:03 then it's probably a opie problem Jun 19 06:33:16 darkschneider: is opie-qss running ? Jun 19 06:33:19 (both Unslung and OpenSlug firmware are build with OpenEmbedded) Jun 19 06:33:30 s/build/built/ Jun 19 06:33:51 /dev/dsp0 is silent instead Jun 19 06:34:00 rwhitby: ah, so you use OE to also build the compatible packages. that's good, we would like to do that as well (i.e. sharpROM compatible packages), but there's too few developers interested in the SharpROM to do that. Jun 19 06:34:13 *cough*ROM*cough* Jun 19 06:34:19 mickeyl: no idea.. wait i see (the rom is the stock one from /official/unstable/3.5.3/sl5600) Jun 19 06:34:21 mickeyl: actually, no. we build the Unslung firmware with OE, but not the packages. Jun 19 06:34:38 rwhitby: ah. what compiler is the stock system using? Jun 19 06:35:01 tehre is a qss process running Jun 19 06:35:04 koen: SharpROM is actually pretty accurate since it's read only :)) Jun 19 06:35:36 only /home is writable? Jun 19 06:35:53 koen: yeah. and they have a shitload of symlinks to make things like /etc tweakable Jun 19 06:35:55 mickeyl: we use a derivative of OpenWRT (buildroot system) to build the Unslung packages. Those packages have also been ported to run on the Asus WL500g router series. glibc 2.2.5 is the compatibility burden we have in Unslung Jun 19 06:36:04 mickeyl: even uglier as I expected Jun 19 06:36:11 darkschneider: hmm no idea offhand then. I've always wanted to write a troubleshooting application where you can test fb, sound, keys, etc. time *cough* Jun 19 06:36:16 the unholy trinity of sharp, tt and lineo Jun 19 06:36:33 koen: *nod* Jun 19 06:36:36 rwhitby: wait a minute. Which packages work on the w500g? <- owns one Jun 19 06:37:04 openwrt buildroot? Jun 19 06:37:05 if you do you know i will use it.. btw i have sent the beckmarks for sl-5600 at opie@handhelds.org... microdrive look like very slow with that test Jun 19 06:37:17 mickeyl: we also keep the packages in a separate repo from the firmware, and have very loose requirements on giving package developers access to the package repo (we have 80 package developers vs 10 firmware developers) Jun 19 06:37:47 CoreDump|home: http://ipkg.nslu2-linux.org/feeds/unslung/wl500g/ Jun 19 06:38:05 you need to be running the firmware from wl500g.dyndns.org Jun 19 06:38:28 rwhitby: wow very nice. Thanks! And I'm already using this firmware Jun 19 06:39:11 CoreDump|home: there is a thread on wl500g.info on how to set up the latest firmware to download and run those ipkgs. Jun 19 06:39:34 should be a matter of editing ipkg.conf (hopefully) Jun 19 06:40:09 CoreDump|home: "ipkg.sh install ipkg ; ipkg update" Jun 19 06:40:25 I absolutely hate the directory structure of this firmware though :) Jun 19 06:40:28 ah Jun 19 06:40:32 (the binary was too big for the wl500g firmware, so we use the script version to bootstrap) Jun 19 06:41:20 CoreDump|home: but you need to bind mount a writeable /opt directory first (all Unslung packages install under /opt for historical reasons) Jun 19 06:41:26 yeah, the flash size is a joke. Thank god one can add a memory stick Jun 19 06:41:37 rwhitby: I'll have a look, thanks Jun 19 06:41:54 np. feel free to become an Unslung package developer :-) Jun 19 06:42:26 http://www.nslu2-linux.org/wiki/Unslung/AddAPackageToUnslung Jun 19 06:44:21 I'm afraid that OE / OZ is already very time consuming :) Jun 19 06:45:41 thanks for the discussion guys. I must sleep now. Jun 19 06:45:54 'night rwhitby-asleep Jun 19 06:46:03 g'night rwhitby-asleep Jun 19 06:48:33 n8 rwhitby-asleep Jun 19 06:48:51 mickeyl: I'd like to talk about w100 rotation. Jun 19 06:49:35 SirFred: go ahead :) Jun 19 06:49:47 mickeyl: The question is that I'm considering two ways to implement it. One should be to enable transformations on the qte driver, and so, be able to say something like QWS_DISPLAY=W100:Rot90:0 Jun 19 06:50:16 mickeyl: In that way, the driver would tranform coordinates, pixmaps,... to adapt it to the real framebuffer orientation that will remain unchanged. Jun 19 06:50:48 mickeyl: The other way could be to actually rotate the framebuffer. I think there's a "hardware way" to do that. Jun 19 06:51:11 and even a sysfs way Jun 19 06:51:12 SirFred: I'm for the second way Jun 19 06:51:17 that should be faster Jun 19 06:51:25 mickeyl: In that way, the only thing to manage from the driver should be the change of dimensions. Jun 19 06:51:34 watch out for tslib weirdnesses Jun 19 06:51:40 koen: I think the sysfs way is the same I'm talking about as the second way. Jun 19 06:52:12 koen: Something I've found in some patches as start_vga_rotation() or something so, for the w100fb.c kernel driver. Jun 19 06:52:40 mickeyl: Let's talk about the second way. Is that the way the screen is rotated nowadays in openzaurus? Jun 19 06:53:46 SirFred: I'm not sure but i think we use the first way since we just reuse the Transformed buffer Jun 19 06:53:48 mickeyl: I suppose it isnt's as far as the Transformed driver is used. Jun 19 06:53:52 right Jun 19 06:54:39 mickeyl: Well, so the idea should be to implement setTransformation in the w100 driver, and call "something" to actually rotate the framebuffer. Jun 19 06:55:07 mickeyl: It should be faster and simpler. I've spend this morning into implementing the first way, but I'm not happy with it. Jun 19 06:55:37 mickeyl: There's a lot of processing involved for images to rotate them before painting. Jun 19 06:56:04 SirFred: I would think the best way for it would be to not change the external API, i.e. just use QWS_DISPLAY=W100:Rotx:0 and then transfer X to the proper setTransformation call in the w100 driver Jun 19 06:56:24 mickeyl: Yes, that's fine Jun 19 06:56:38 mickeyl: I want to keep that way to specify the rotation. Jun 19 06:56:53 ok, good Jun 19 06:56:56 mickeyl: But I don't want to use some mapping wrappers all the time. Jun 19 06:57:03 mickeyl: To get the actual coordinates, images,... Jun 19 06:57:48 RP: What should be the right way to change the orientation from qt/e itself? Jun 19 06:58:39 RP: Is there some IOCTL operation or something so? Jun 19 07:02:40 RP: I've moved the cvsdate for kdrive forward, so we should benefit from the fb tweaks Jun 19 07:12:19 altboot: Disable fsck of loop-images since fsck isn't installed in a stock OZ image Jun 19 07:15:02 CoreDump|home: could that be solved with a if [ -e /bin/fsck ] ? Jun 19 07:15:57 yeah sure. I was thinking about that, too. But I don't remember an image shipping fsck in the first place :) Jun 19 07:16:35 true Jun 19 07:16:41 does fam come w/ fsck? Jun 19 07:17:03 * koen checks Jun 19 07:17:19 only fsck.minix Jun 19 07:17:24 heh Jun 19 07:17:32 * koen wonders why Jun 19 07:20:21 could someone running a 2.4 kernel w/ sdcontrol _please_ look at bug #47. It is driving me crazy to hand-tune busybox after every build from scratch Jun 19 07:20:50 FYI: This bug is a show-stopper for collie and poodle and maybe more machines Jun 19 07:26:57 does anybody have good contacts with the ubuntu people? Jun 19 07:28:55 ok, deleting /bin/busybox wasn't very clever heh Jun 19 07:34:06 does anyone here have experience with mtx-1? I'm trying to get an ATI Rage XL to init Jun 19 07:39:09 mickeyl: how do vote? 1 vote / person or award points for 1st 2nd and 3rd choices? Jun 19 07:39:57 s/do/to/ Jun 19 07:52:10 :( i still have no way to make the usbnerowrking work Jun 19 07:58:45 moin zecke Jun 19 08:18:54 03CoreDump 07 * r1.3573 10openembedded/conf/distro/openzaurus-3.5.4.conf: openzaurus-3.5.4.conf: s/opfficial/official for the machine feed. Thanks to Broeggle for pointing it out Jun 19 08:37:31 zecke: how's your tinderbox endavour working out? Jun 19 08:45:37 koen|afk: I'm waiting for france Jun 19 08:45:53 I've a tinder.bbclass Jun 19 08:48:53 koen|afk: so once we've tinderbox.hh.org we can start the clients Jun 19 08:49:20 zecke: that's good to hear Jun 19 08:49:30 mickeyl: what version of monotone are you going to use? Jun 19 08:52:10 oops my laptop dies Jun 19 08:54:33 time to get an ibook Jun 19 09:02:59 ok, OE tree (without history) imported Jun 19 09:03:15 mickeyl: are you going to import the bk cvs mirror? Jun 19 09:08:51 mickeyl: I have a monotone repo too, so we can try to merge stuff Jun 19 09:12:37 <_ggilbert> mickeyl: looking at going with monotone? Jun 19 09:14:03 mickeyl: exacly where i can find the old slow kernel, since both kernel in 3.5.3 dir does not look so stable (shutdown in about 2 hours and need hard reset) Jun 19 09:16:20 darkschneider: you may one to try this one for poodle: http://oz.hentges.net/releases/3.5.3/T2/poodle/kernels/ Jun 19 09:17:23 CoreDump|home: sure :) i hope usb contines to work.. finally after 2 hours i can finally ping the Z :) Jun 19 09:17:30 heh Jun 19 09:19:11 i remember right that to flash only the kerneli just need to put only it with updater on the cf, rigth? Jun 19 09:20:32 * CoreDump|home thinks so Jun 19 09:20:45 tx! Jun 19 09:25:31 done..let's see (i allways miss the lack of feedback when flashing.. a message "kernel sucessfuly updated" would like much to me Jun 19 09:26:15 that's sharp Jun 19 09:26:22 lol Jun 19 09:26:25 anything 'usefull' is missing Jun 19 09:26:40 (yes a serial output or a parallel too..) Jun 19 09:26:53 even 1 wire output could help Jun 19 09:26:55 it should have serial Jun 19 09:27:05 aye need cable Jun 19 09:27:06 but you'll need a expensive cable Jun 19 09:27:20 indeed.. real problem is that i'm in italy... can0t have it anyway Jun 19 09:30:08 koen: i'm for every person has 3 choices and then we count the points Jun 19 09:30:19 koen: version 0.19 - versions older are much too slow Jun 19 09:30:45 koen: i won't import the history - even 0.19 would be too slow with that amount of history. the history will be read-only in a seperate cvs repo afterwards Jun 19 09:31:01 darkschneider: no idea about other kernels, sorry. Jun 19 09:31:54 03CoreDump 07 * r1.3574 10openembedded/packages/altboot/ (altboot_cvs.bb altboot_20050620.bb): altboot: Update to latest version Jun 19 09:33:53 mickeyl: ok, I'll try to open up the firewall to allow netsync so we can experiment a bit Jun 19 09:34:04 I'll try to merge bk commit from now on Jun 19 09:36:26 oki used the one Core suggested, let see if it's stable, i'm doing opie benchmarks to see if there is any numeric diference Jun 19 09:46:20 pb_: are you the brains behind blueprobe? Jun 19 09:46:44 yes Jun 19 09:47:08 ok, I thinks I'll have some question for you the next few days Jun 19 09:47:12 think* Jun 19 09:47:24 thanks for the warning Jun 19 09:48:46 SirFred: You can use ioctl's to rotate the display as with any other fraembuffer - look and see how fbset does it Jun 19 09:49:16 SirFred: I'm not familiar with the user side of the fbset code, just what the kernel gets to see... Jun 19 09:58:42 mickey|bbl: I also think we need to make a monotone mini tutorial for OE if we decide to go for it Jun 19 09:58:58 mickey|bbl: or at least a small bk -> monotone cheatsheet Jun 19 09:59:20 koen: THat would be handy :) Jun 19 10:12:30 The main missing thing in monotone is the bk GUI, but it looks like relatively few people use that regularly. Jun 19 10:13:00 yeah, mostly because it's so ghastly Jun 19 10:13:37 I use citool but will happily move to something more sane :) Jun 19 10:14:11 IIRC monotone has the SVN/CVS issue with getting everything checked in - bk citool makes it difficult to forget to check things in. Jun 19 10:14:17 bk vi bla ; bk ci -y"modified bla" bla ; bk commit ; bk push Jun 19 10:15:50 jbowler: yeah, that's true. Jun 19 10:16:12 * koen looks for the monotone version of bk sfiles Jun 19 10:17:41 I run {{bk edit; vi}*; { test; sleep }*}*; {bk citool; wth was I doing... oh yes, discard/commit}; bk push Jun 19 10:18:42 yeah Jun 19 10:18:54 bk citool is a fine idea, it's just a dire implementation Jun 19 10:19:03 agreed Jun 19 10:19:12 0.00000001 > 0 Jun 19 10:19:17 heh Jun 19 10:43:30 hey Pigi Jun 19 10:44:04 hey koen. I'm trying to flash my (secondhand) new h3900.... Jun 19 10:44:51 Pigi: cool Jun 19 10:44:51 but I'm right in the " can't read control file " bug from bootloader :) Jun 19 10:44:57 heh Jun 19 10:45:08 check your cf card formatting: bootldr hates fat32. Jun 19 10:45:36 gosh... what's a good partition type ? Jun 19 10:45:45 fat16/vfat Jun 19 10:46:09 RP: Thanks, I'll take a look. Jun 19 10:46:19 it's a fat16 :( Jun 19 10:46:44 I have one CF card here that bootldr hates Jun 19 10:47:20 this card was really good for the older sa arch ( 3630-3760-3765 and 3870 ) Jun 19 10:47:23 03pb 07 * r1.3575 10openembedded/packages/libgpewidget/libgpewidget_0.102.bb: update libgpewidget to 0.102 Jun 19 10:47:30 going to get another card. Jun 19 10:47:37 ah, that's weird. the pxa shouldn't behave any differently. Jun 19 10:48:02 maybe you have discovered some new, obscure bug. Jun 19 10:48:28 that's really strange, I'm using the same sleeve and card. What has changed is the ipaq :) Jun 19 10:48:37 after the defeating of my 3870 Jun 19 10:52:13 pb_ do you need a bug finder ( not hunter ) for the bugs in bootloader ? Jun 19 10:52:28 It seems I have already discovered two :) Jun 19 10:52:46 now my wife call for dinner Jun 19 10:52:47 brb Jun 19 10:53:06 * Pigi is away: I'm busy Jun 19 10:56:01 SirFred: As a hint, you want to change the resolution, not rotate the display - just request the x and y resolutions you want Jun 19 10:56:56 RP: So, how can I tell between 90 and 270 degrees, for example? Jun 19 10:57:21 RP: I mean, the x and y resolutions are the same, aren't they? Jun 19 10:58:06 RP: I think that this stuff should be implemented in the linuxfb driver, that I'm subclassing with the W100. Jun 19 10:58:24 SirFred: There is no way to handle that through the linux framebuffer as things stand Jun 19 10:58:33 SirFred: Hence the flip sysfs file... Jun 19 10:58:44 Its on my list of things to try and sort out sometime... Jun 19 10:58:51 But a low priority Jun 19 11:00:10 RP: I wonder what is the more educated way to access that from the driver. Jun 19 11:00:20 CoreDump|home: that kernel still crash after heavy use.. Jun 19 11:00:24 RP: Is there any way other than opening the file and writing to it? Jun 19 11:01:51 SirFred: I'm afraid not Jun 19 11:02:04 SirFred: Although opie may have some way of abstracting sysfs files? Jun 19 11:03:34 RP: It's only my problem, sure. I just feel that opening a file and manipulating it is a little dirty. But that is what sysfs is about. Jun 19 11:04:58 RP: I'll try to write something that just work. Later, I will question the implications of it. Jun 19 11:12:16 SirFred: It would be nice to find a better solution and sort out rotation handling in the fb layer Jun 19 11:13:16 RP: Sure. but I'm not sure of what should be the best solution. Is there info related with the flipping state of the framebuffer in the framebuffer standard structures? Jun 19 11:14:18 SirFred: There are some rotation hooks in there. Nobody seems to understand what they do or how to use/implement them Jun 19 11:15:02 I think there were problems with the current implementation but I can't remember what they were :-( Jun 19 11:15:50 how tslib deal with the rotation? Jun 19 11:16:32 koen: No idea. I did offer to make corgi_ts rotate with the screen but the idea sent shivers down poeple's spines :) Jun 19 11:17:01 The weather here is quite spectaular - thunder and lightning and torrential rain, so much so the gutters are overflowing and the street looks like a river... Jun 19 11:17:23 RP: I love storms. Jun 19 11:17:59 It's going to be a really hot and dry summer here. Jun 19 11:18:02 SirFred: So do I. Its just far too hot and humid :-/ Jun 19 11:18:54 We are at 36 degrees just now. Too hot for me. Jun 19 11:19:02 I can't see the next house, the rain is that heavy! Jun 19 11:19:15 33 in this room Jun 19 11:20:40 hot & humid over here Jun 19 11:21:13 warm and dry here Jun 19 11:21:13 heh Jun 19 11:21:41 RP: How should I proceed, something like this? Jun 19 11:21:49 16.5 here Jun 19 11:21:55 open "sys/devices/platform/sharp-scoop/corgi-ssp/w100fb" Jun 19 11:22:01 RP: write "1" ? Jun 19 11:22:21 ljp: crikey Jun 19 11:22:33 heh Jun 19 11:22:47 got down to 6 the other night Jun 19 11:23:22 sounds pretty extreme Jun 19 11:24:03 I prefer that than our 20 minimun - 38 maximum Jun 19 11:26:11 SirFred: you might want to check into libsysfs Jun 19 11:26:23 morning kergoth Jun 19 11:26:26 hey Jun 19 11:26:31 kergoth: Hi and thanks. Jun 19 11:27:04 hi kergoth Jun 19 11:27:34 SirFred: Something like that, yes. libsysfs will probably save you lots of pain :) Jun 19 11:28:35 RP: I expect nobody to get angry if libsysfs is a new dependency for libqte. Jun 19 11:29:29 "fluff" Jun 19 11:31:18 koen: I think there is something wrong with your irc client. Almost every time you try to say something, it seems to come out as "fluff". Jun 19 11:31:45 :) Jun 19 11:31:47 how strange Jun 19 11:35:48 had some incredible weather here in northwich Jun 19 11:36:03 water was *gushing* into the kitchen under the back door. Jun 19 11:36:18 I had to open it (got me soaked) and jam towels into it... Jun 19 11:36:53 wow Jun 19 11:36:59 hm. I better check on my fileserver (underfloor) Jun 19 11:37:02 brb Jun 19 11:39:51 no problem. a bit damp down there but all A-OK water wise Jun 19 11:43:03 * Pigi is back (gone 00:49:57) Jun 19 11:43:16 re Jun 19 11:46:27 pb_ you there ? Jun 19 11:49:27 Pigi: hello Jun 19 11:49:59 I'm trying ( again ) to flash the 3900 via cf. Jun 19 11:50:27 It seems that with the new sleeve there isn't way to read the cf ( tried a couple of cf ) Jun 19 11:50:50 with an old sleeve, it seems a bit better, as at least it read the controlfile Jun 19 11:50:55 strange Jun 19 11:51:04 are the two sleeves the same? Jun 19 11:51:18 no, they aren't. Jun 19 11:51:40 One is the new one, the one with the protecting plastic Jun 19 11:51:48 and it is never read. Jun 19 11:52:08 The other one is the old version, the one that used to be with the h3600 series. Jun 19 11:52:14 ah. Jun 19 11:52:21 I don't think I've seen one of those new sleeves. Jun 19 11:52:36 Is the one with the removable battery Jun 19 11:52:58 maybe bootldr doesn't understand how to talk to them properly Jun 19 11:53:13 it can be. Jun 19 11:53:30 But with the old one it seems that is not able to read the image anyway Jun 19 11:54:28 oh dear Jun 19 11:54:42 No, stdby. Jun 19 11:54:54 Maybe I did a little error in reflash.ctl Jun 19 11:55:59 but with the new sleeve, no way to get reflash.ctl read. Jun 19 11:56:21 The old one seems ok. It was an error in reflash.ctl Jun 19 11:56:45 Aha. Jun 19 11:56:55 Who's maintaing bootloader these days ? Jun 19 11:57:01 Nobody, really. Jun 19 11:57:06 gosh. Jun 19 11:57:14 I think the hx4700 guys probably touched it last. Jun 19 11:57:46 We're kind of in this limbo state where nobody really wants to do any more work on the old bootldr, but nobody really wants to use LAB either. Jun 19 11:58:21 ok. That's sad, because the 38xx bootloader is able to work with that sleeve, that is nice as have the removable battery Jun 19 11:58:46 pn 249704-b21 if interessed Jun 19 11:59:17 unfortunately, it's virtually impossible to debug this kind of thing without having the hardware in front of you. Jun 19 11:59:36 so, you'd need a bootldr maintainer who has one of those sleeves, and I suspect such people are a scarce resource. Jun 19 12:00:11 hmm, I need a new battery for my new sleeve Jun 19 12:00:18 The only way to play with the bootloader is to use jtag, right ? Jun 19 12:00:48 Pigi, you need jtag if you ever make a mistake that causes your device to be unbootable Jun 19 12:01:16 I know, jacques, I have a nice 3870 bricked, ready to go to valhalla Jun 19 12:01:23 in theory it's possible to do bootloader development without jtag, but you'd have to be a 100% perfect coder Jun 19 12:03:30 well, some bits of bootldr are relatively low risk Jun 19 12:03:58 you can work on the sleeve code, for example, with impunity because it is not used in any critical path. no matter how badly you break that stuff, you can still install a new bootldr using serial. Jun 19 12:04:37 the only catch is that setting yourself up to compile the bootldr in the first place is slightly risky. If your toolchain generates duff images, obviously all is lost. Jun 19 12:04:51 yeah Jun 19 12:05:05 eheh. I'm not sure I will risk another machine ( and 200 E again ) Jun 19 12:05:15 heh, right Jun 19 12:05:19 speaking of which, I got wine setup and built haret Jun 19 12:05:32 fitting jtag to the h3900 is relatively straightforward, though. Jun 19 12:05:38 now I need to find time to go through the code and try to determine where the endian bit would be set Jun 19 12:05:39 much easier than on the h3600/h3800. Jun 19 12:05:48 jacques: ah, cool Jun 19 12:06:14 grepping for -i endian didn't find anything obvious :-) Jun 19 12:06:35 yah, it probably doesn't do anything at all with that bit at the moment. Jun 19 12:06:45 I was afraid of that Jun 19 12:06:53 you probably want to turn it on right at the end, just when it jumps into the kernel Jun 19 12:07:14 good to know, thanks Jun 19 12:07:41 I should probably compare with the be bootloaders we have available - redboot and APEX Jun 19 12:07:49 actually, it occurs to me that an easier solution would be to patch the kernel to flip the endian bit in its own bootup code. Jun 19 12:08:01 it can do that?? Jun 19 12:08:11 sure Jun 19 12:08:30 it never occurred to me :-) Jun 19 12:08:34 just put it in head.S before anything complicated happens, and it should be just as good as doing it in haret. Jun 19 12:08:49 well that's a lot easier to do (for me) Jun 19 12:08:52 yeah Jun 19 12:09:26 building haret is always a bit of an adventure, all that fiddling with eVC++ Jun 19 12:09:37 hmm, for me it was exactly as in the wiki Jun 19 12:09:43 emerge wine Jun 19 12:09:56 untar the evc tarball Jun 19 12:10:00 ah, actually, I was thinking of bootblaster Jun 19 12:10:04 patch the haret makefile, and make Jun 19 12:10:07 ah Jun 19 12:10:13 for some reason I could never get that to build right, I had some DLL problem Jun 19 12:10:24 maybe wine is better now Jun 19 12:10:28 I'm not sure it builds for anyone other than Brian, in fact Jun 19 12:10:28 heh Jun 19 12:10:56 I wasn't using wine, this was on a real windows machine. Jun 19 12:11:10 maybe that was my mistake :-) Jun 19 12:11:14 lol Jun 19 12:11:31 BTW, I learned a valuable lesson yesterday about usb card readers Jun 19 12:11:42 some are faster than others Jun 19 12:12:06 my SD cards that were doing 6MB/s went up to 8.85MB/s in my new reader Jun 19 12:12:10 aha Jun 19 12:12:25 and my ultraII cards I just got (on sale at costco) do 9.65MB/s Jun 19 12:12:31 that's cool Jun 19 12:12:42 let's hope someone figures out the speed problem with the h2200 driver soon Jun 19 12:12:43 they onlt did 6.5MB/s on the old reader Jun 19 12:12:55 yeah - that reminds me - I want to bench the h2200 under wince Jun 19 12:13:00 ah, good plan Jun 19 12:13:20 I benched it when I first got it but I might not have even owned any SD cards then Jun 19 12:13:29 heh Jun 19 12:16:10 ah, hm, I guess I should go to a birthday party. Jun 19 12:16:23 * pb_ looks for some suitable bottles of wine to take Jun 19 12:16:40 heh, have a good time Jun 19 12:17:31 pb_, before you leave, what's the state of SD on h3900 ? Jun 19 12:17:51 pb_ don't worry, I found on releasenotes Jun 19 12:18:17 Pigi: aha :-) Jun 19 12:18:28 later all Jun 19 12:18:31 * pb_ -> Jun 19 12:21:11 have fun Jun 19 12:33:30 koen, do you own an 3900 ? Jun 19 12:33:43 no, a 2200 and 5550 Jun 19 12:34:06 and michaelo is lending me his hx4700 Jun 19 12:34:29 ok. It seems that I can't get my sd recognized in 0.8.2 Jun 19 12:34:43 I0m not so happy about that 3900 :( Jun 19 12:34:48 did you rmmod and insmod the mmc modules? Jun 19 12:35:01 and rebooted quite a bunch of times Jun 19 12:36:26 well, after rmmod and insmod it seems to work, but I'm not sure it will survive a reboot. Jun 19 12:42:08 * koen is abit puzzled between the MT/ dir and the db in monotone Jun 19 12:44:11 Pigi: what problem? Jun 19 12:44:31 about the sd card, not recognized on boot. Jun 19 12:45:04 Pigi: hotplug detection not works.. for SD and 39xx (in case you did not know.... :) Jun 19 12:45:16 I have read that :) Jun 19 12:45:42 look in realease notes... for 2.8.0.. there is some commands ... Jun 19 12:45:52 to have it loaded in boot time.. Jun 19 12:45:56 ?2.8.0 ? Jun 19 12:46:04 0.8.1 maybe:) Jun 19 12:51:23 same is for later versions.. (IMHO) Jun 19 12:51:23 ups... Jun 19 12:51:23 or 0.8.0... Jun 19 12:51:23 Search for MMC/SD card doesn't get recognized on H39xx (not h5xxx) Jun 19 12:51:23 in release notes... Jun 19 12:51:38 yes, already seen, thx Jun 19 12:53:22 gints|home, do you own a 39xx ? Jun 19 12:54:23 Pigi: yes.. Jun 19 12:54:41 opie or gpe ? Jun 19 13:19:45 koen, DISTRO=0.8.3 ? Jun 19 13:19:57 familiar-0.8.3 Jun 19 13:20:05 yes indeed Jun 19 13:20:24 Pigi: http://familiar.handhelds.org/snapshots/v0.8.3/20050616/ Jun 19 13:20:32 is ipkg 0.99.151 in ? Jun 19 13:20:42 no idea Jun 19 13:20:47 ok Jun 19 13:24:22 thx for the link, koen Jun 19 13:27:21 03koen 07 * r1.3576 10openembedded/packages/ipkg/ (3 files in 2 dirs): Add ipkg 0.99.151, courtesy PierLuigi Frullani Jun 19 13:27:43 ~hail koen Jun 19 13:27:44 * ibot bows down to koen and chants, "I'M NOT WORTHY!!" Jun 19 13:27:55 ~praise koen Jun 19 13:27:56 All hail koen! Jun 19 13:28:52 mighy! ping crash my Z :) Jun 19 13:29:08 ping_of_death..... Jun 19 13:29:36 eheh the only diffence is that i'm pinging from it :P Jun 19 13:33:43 ping_of_suicide ? Jun 19 13:34:14 yes :) Jun 19 13:34:40 you can file a CERT advisor as "local DOS" Jun 19 13:34:51 heh funny poodles Jun 19 13:34:52 hehe indeed :) Jun 19 13:35:05 depends on the fun...:P Jun 19 13:35:25 sell it and buy a collie. Waaaayyyyyyyyyyyyyyyy better supported ;) Jun 19 13:35:39 want to buy mine :) Jun 19 13:35:45 not really heh Jun 19 13:35:57 then i'm out of customers :) Jun 19 13:36:23 i remember when i bougth it.. igone to oz channel and asked "what i shoudl buy" Jun 19 13:36:37 everyone suggested me the really improved sl-5600 :) Jun 19 13:37:36 anyway.. it's ok... you guys are often great.. expecially mickey and kergoth that helped me much everytime i needed Jun 19 13:38:53 hi all Jun 19 13:39:41 darkschneider: "improved" lies in the eye of the beholder I guess :) Jun 19 13:40:00 :) Jun 19 13:40:27 koen: if you have some free time, could you please help luke_jr build abiword on #penzaurus ? he is stuck with this : http://pastebin.ca/15040 . iirc, you did build abiword for GPE, right ? Jun 19 13:41:03 I did Jun 19 13:41:23 the solution is to use the gcc-3.4.4 compiler instead of the _csl one Jun 19 13:42:25 ok. thanks a lot koen Jun 19 13:43:13 np Jun 19 13:45:13 got to go. Thanks a lot again koen. Jun 19 13:45:16 bye Jun 19 14:02:39 i correct, ANY tentative with any program to access the usb crash my nice poodle :) not just ping :) Jun 19 14:07:55 koen: hey Jun 19 14:08:17 koen: ping Jun 19 14:34:34 I am just reading about the direction the kernel development is taking about the BK debacle. I'd be interested to learn what the state of the discussion is for OE. Jun 19 14:46:23 Laibsch, we are waiting for the users to donate enough to buy licenses Jun 19 14:46:57 sleepin' time Jun 19 14:47:01 night all Jun 19 14:49:59 on that note ... do we even have a donate page? Jun 19 14:52:54 it is possible to donate directly to the devs Jun 19 14:53:25 but that will not solve the bitkeeper problem Jun 19 14:54:27 emte: Are you serious with that comment??? I know that kergoth went through a couple of VC systems but as others was not happy either with performance or feature set. I hear that Linus has developped his own solution for the kernel but most comments say it might not be applicable outside kernel development. Jun 19 14:54:40 well if we come up with 5k for each dev, all is well Jun 19 14:55:07 you dont need a license to pull from bk Jun 19 14:55:46 and i was only semi-serious Jun 19 14:56:05 Laibsch, torvalds is using git and that seems just to be usable by himself :/ Jun 19 14:56:20 but it is an option ... not a feasible one mind you ... but an option Jun 19 14:56:33 tubbie|food, not using, writting Jun 19 14:57:20 at the moment monotone is the most serious candidate Jun 19 14:57:27 there are not real open projects that can handle things like OE and the kernel Jun 19 14:57:30 and git/cogito sucks ass Jun 19 14:57:30 Thanks kone for the comment. Jun 19 14:57:42 Sorry for the misspell. Jun 19 14:57:47 so monotone finally works now? Jun 19 14:57:51 I heard also positive notes from monotone Jun 19 14:57:56 cogito can't push Jun 19 14:57:56 Well, it always worked Jun 19 14:58:01 Just very slow I heard Jun 19 14:58:08 it's reasonable fast Jun 19 14:58:20 Yes, the newer versions are said to be better. Jun 19 14:58:21 on a dual opteron 244 :) Jun 19 14:58:25 ;-) Jun 19 14:58:36 You can use mine ;-) Jun 19 14:58:40 lol Jun 19 14:59:19 I just imported the OE tree into monotone, mickeyl and I are going to play with it Jun 19 14:59:21 I just read about a new machine that kernel.org got donated from HP: quad-opteron with 24G RAM !!! I'd like to take that puppy for a spin Jun 19 15:00:01 That machine has more RAM than my desktop has in hard drives. Jun 19 15:00:09 But I do not need more. Jun 19 15:00:41 koen: Cool, thanks for the update. Jun 19 15:00:50 I hope it works out Jun 19 15:00:51 yeah linus is probably the only one HP would donate anything to with their new managment Jun 19 15:01:14 hehe Jun 19 15:02:20 koen: ping Jun 19 15:26:00 how do I select gcc-3.4.4 for compiler? Jun 19 15:35:07 luke-jr_: REFERRED_VERSION_gcc-cross=3.4.3 Jun 19 15:35:07 PREFERRED_VERSION_gcc-cross-initial=3.4.3 Jun 19 15:35:12 except 3.4.4 :) Jun 19 15:35:34 Both should be PREFERRED... Jun 19 15:37:45 RP: doesn't virtual/ go in there? O.o Jun 19 15:38:40 luke-jr_: That's only if you're going to build and provide the compiler for oe to use Jun 19 15:38:58 PREFERRED_PROVIDER_virtual/arm-linux-gcc = "gcc-cross" Jun 19 15:40:15 That's saying the preferred arm compiler is provided by gcc-cross... Jun 19 15:40:31 You then tell it which version of gcc-cross you'd prefer Jun 19 15:41:04 proferred providers and proferred versions are two different things Jun 19 15:42:06 right, but is not the preferred provider needed for it to even mess w/ version? Jun 19 15:45:00 seems to work, thx Jun 19 15:45:07 * luke-jr_ ponders if GCC 4.0.0 is stable enough Jun 19 15:48:42 koen: Just came across another VC system that sounds interesting. Just wanted to make sure you are aware of it: http://www.selenic.com/mercurial/ Jun 19 15:51:17 koen isn't here... Jun 19 15:52:14 luke-jr_: Thanks for point it out. pinging kergoth. Jun 19 17:18:47 http://pastebin.ca/15056 -- after cc1 eats RAM continuously. Any ideas? Jun 19 17:28:11 <[g2]-in-and-out> don't use cc1 :) Jun 19 19:40:08 i am having trouble with building opie at the 'make' command, could someone please assist me? **** ENDING LOGGING AT Sun Jun 19 23:59:56 2005