**** BEGIN LOGGING AT Mon Apr 11 23:59:56 2005 Apr 12 00:26:29 morning Apr 12 00:27:24 morning Apr 12 07:21:37 ta-da! Apr 12 07:22:00 * NAiL soldered in a brand spanking new 512mb pendrive instead :) Apr 12 07:24:49 <[g2]-away> NAiL, on the nslu2 ? Apr 12 07:24:54 yeah Apr 12 07:25:12 <[g2]-away> flash chips ? Apr 12 07:25:20 Removed port2, and soldered a flash drive inside the case. Kinda looks like a daughterboard ;) Apr 12 07:25:23 no Apr 12 07:25:39 Just don't want a pendrive permanently sticking out Apr 12 07:25:46 So I soldered it on inside instead :) Apr 12 07:25:46 <[g2]-away> ah Apr 12 07:25:57 <[g2]-away> mechanical stability and visuals Apr 12 07:26:06 yeah Apr 12 07:26:17 Even took nice pictures during the process Apr 12 07:26:24 <[g2]-away> cool Apr 12 07:26:26 <[g2]-away> url ? Apr 12 07:26:33 (and managed to destroy two 128mb flashdrives as well) Apr 12 07:26:39 <[g2]-away> doh! Apr 12 07:26:42 I'll get the pics up later tonight Apr 12 07:26:47 <[g2]-away> is that MB or Mb Apr 12 07:26:51 Gotta go set up a server Apr 12 07:26:54 MB Apr 12 07:27:23 <[g2]-away> you're running Openslug right ? Apr 12 07:27:26 yeah Apr 12 07:27:35 <[g2]-away> I've been playing with native compiles Apr 12 07:27:53 Haven't done any of that since I compiled samba Apr 12 07:28:31 Now that I've got more room, I might start again Apr 12 07:28:52 sigh Apr 12 07:28:55 <[g2]-away> I know that you already do this, but I've got to say it anyway Apr 12 07:29:12 <[g2]-away> make sure you flash stick is mounted noatime Apr 12 07:29:17 yeah ;) Apr 12 07:29:22 <[g2]-away> that's not the default Apr 12 07:29:24 I know Apr 12 07:29:36 btw, what filesystem is best to use? Apr 12 07:29:48 <[g2]-away> I use ext3 Apr 12 07:29:55 I don't need journalling on the flash, to be honest Apr 12 07:30:07 <[g2]-away> probably not Apr 12 07:30:28 <[g2]-away> but there's less of a need for fsck Apr 12 07:30:41 yeah, but the logging adds extra wear ;) Apr 12 07:30:52 Hmm... would it be possible to use jffs2? Apr 12 07:31:01 <[g2]-away> I think the flash sticks spread out the writes anyway Apr 12 07:31:05 It has compression, and is "nice" to flashdrives? Apr 12 07:32:00 <[g2]-away> I'll play around with reiserfs on the HD when I get time, but that probably be a while Apr 12 07:32:03 Oh well.. Gotta go. Be back later tonight. I'll put up the pics then Apr 12 07:33:31 <[g2]-away> tc Apr 12 08:29:22 Is it possible to format that 512mb-stick of mine with jffs2? Apr 12 08:30:11 and... I don't really need to partition the disk, do I? I can just use the whole stick as a partition? Apr 12 08:44:50 if not jffs2, what fs would people recommend for use on a flash stick? Apr 12 09:05:04 NAiL: a flash stick is a block device, so JFFS2 won't work - use a normal file system like ext2/3 or vfat Apr 12 09:05:33 vfat is probably your best pick if you need windows support, otherwise ext2 Apr 12 09:11:30 Jacmet: Considering that the stick is soldered on inside the slug, I don't think I need vfat support for it ;) Apr 12 09:13:26 <[g2]-away> Jacmet, I think the guy from FlashLinux is running jffs2 on usb flash sticks Apr 12 09:14:02 <[g2]-away> I know he modified some drivers Apr 12 09:20:47 NAiL: ahh, you're talking about a normal NOR flash chip - I thought you meant a USB flash stick Apr 12 09:20:55 * Jacmet is tired Apr 12 09:21:24 NAiL: in that case - cramfs or squashfs for a readonly FS, otherwise jffs2 Apr 12 09:24:40 Jacmet: Uhmm.. It's a USB flash stick. Like the ones you can insert into a Wintendo PC without doing anything Apr 12 09:24:46 wb, [g2] :)= Apr 12 09:25:41 <[g2]> thx Apr 12 09:25:49 NAiL: ok, then you indeed cannot use jffs2 and you'll have to use ext2/3 or vfat or whatever Apr 12 09:25:52 Flash sticks like that ar NAND, right? Apr 12 09:26:47 NAiL: yes, but you don't have lowlevel access to them - the usb interface simply emulates a hard drive (or rather a block device) Apr 12 09:26:56 ah, yes Apr 12 09:26:58 that;s why you cannot use jffs2 Apr 12 09:27:14 and you need to use a "normal" fs Apr 12 09:36:38 Jacmet: There's actually support for jffs2 on block devices in 2.6.11 Apr 12 09:39:50 NAiL: perhaps, but it doesn't really make sense - the idea of jffs2 is basically to ensure data consistency and handle wear levelling, but because those lowlevel details of the flash chips are not available over the USB interface it doesn't work Apr 12 09:40:30 NAiL: the USB controller will try to do it's own wear levelling, so JFFS2 and the controller will probably end up working against each other Apr 12 09:40:57 that's why a normal FS like ext2 (preferably mounted noatime) is a much better choice Apr 12 09:42:03 Quite possible Apr 12 09:58:27 hmm ipkg problems Apr 12 10:07:22 yeah. my build fails too. Apr 12 10:08:19 mine was working.. until something got corrupted and I got the SMB fail error Apr 12 10:08:31 but I think I just installed to many packages and it ran out of room Apr 12 10:08:35 different problem I think. Apr 12 10:09:03 yeah I just formatted my drive again and I'll unsling it again Apr 12 10:19:19 hey Apr 12 10:21:51 hi Apr 12 10:22:27 Wrecked my usb-key, so I had to get another one. Got a 512mb one today, and it's working like a charm ;) Apr 12 10:22:45 Gonna put up some nice pics later tonight, when I get home Apr 12 10:23:09 cool Apr 12 10:23:34 hey i got a quick question...... Apr 12 10:27:10 has anyone put nslu2 on batteries? Apr 12 10:27:48 its certianly posible....but such a waste Apr 12 10:28:39 heh... Would be a fun thing to do ;) Apr 12 10:29:09 jipi: have a look at http://webuser.fh-furtwangen.de/~dersch/gphoto/remote.html Apr 12 10:30:17 if i buy a Maxtor One Touch backup thing...will it worg with the slug or will i have to install sofware on my machines and the slug.... Apr 12 10:30:50 jipi: In fact, my slug runs on my robot using NiMh AA battery packs. Apr 12 10:31:06 gosh!!!!!!!!!!!!!! Apr 12 10:31:26 Do you mean whether or not the "One Touch" function will work? AFAIK, it won't. The enclosure will work nicely IIRC Apr 12 10:31:28 i was just thinking about putting my nslu2 on my remote control car. Apr 12 10:31:38 I use an 8 cell pack and regulate it using a DC/DC switch. Apr 12 10:31:49 how long does say four good batteries last Apr 12 10:31:53 I'm using a web cam for vision. Apr 12 10:32:17 <[g2]> I've run the nslu2 on 4 niMh AA batteries for a little while Apr 12 10:32:41 I got 2100mAh and it would run for hours if it wasn't for my motors and power-sucking sensors. Apr 12 10:33:23 The power consumption of the slug isn't that big, is it? 5V 2A, of which I guess 1A is "extra" for the USBPWR for external laptop drives etc? Apr 12 10:33:24 toobad the ixp420 processor isn't optimised for suspend mode? Apr 12 10:33:43 It was measured to be around 400mA Apr 12 10:33:52 with no HD of course! Apr 12 10:33:53 has any1 tried turing there slug into a wireless "g" access point (not bluetooth) Apr 12 10:34:07 400mah... So a 2300mah battery back should last a while ;) Apr 12 10:34:40 I would think so. Unless it's working extra hard the whole time. Apr 12 10:35:02 yeah Apr 12 10:36:23 caplink: how do u control your camera? are there any API/commands openly available? Apr 12 10:37:42 jipi: isn't my page Apr 12 10:51:53 man room is idle.. Apr 12 10:52:19 hey Apr 12 10:55:41 Yeah, we try to out-idle each other ;) Apr 12 10:56:13 I'm just busy trying to figure out why linux doesn't see my PDC20265... Apr 12 10:56:35 k Apr 12 12:49:57 03bzhou * 10unslung/make/postgresql.mk: upstream upgrade from 8.0.1 to 8.0.2 Apr 12 12:51:20 hey, anyone have any idea why ipkg update errors with a "couldn't write" error? Apr 12 12:51:42 ~nslu2-general Apr 12 12:51:43 #nslu2-general is the IRC channel specifically set up for end-user questions about the NSLU2 device. Questions that are not related to the nslu2-linux development projects should not be asked in #nslu2-linux, but should be asked in #nslu2-general instead. Apr 12 12:54:06 In the Slugbug tracker, if a bug is assigned to NSLU2-linux who is responsible for working on it? Apr 12 12:54:35 I need to find out how to fix something for appWeb Apr 12 12:56:12 Particularly, I would like to find out what the status is of bugs 81 and 101 Apr 12 12:56:43 copperbeech, I was just going to ask you that question! Apr 12 12:56:47 :-) Apr 12 13:04:01 03jeanfabrice * 10unslung/make/ (fontconfig.mk thttpdphp.mk): Corrected dependencies for thttpdphp Apr 12 13:30:55 <[g2]> kergoth, HEY! Apr 12 13:34:16 03jp30 * 10unslung/make/ (libcurl.mk libbt.mk): fix libbt build Apr 12 15:08:31 03jp30 * 10unslung/make/ (xchat.mk vte.mk gtk.mk): xchat, vte, gtk: clean up staging mechanism Apr 12 15:11:16 03jeanfabrice * 10unslung/make/thttpdphp.mk: Bumped IPK version Apr 12 15:27:48 hmm... Anyone know where I can find the mount options for ext3? Apr 12 15:33:52 nm, found what I needed Apr 12 15:34:10 Another question.. Where does openslug put packages by default? Apr 12 15:35:52 deploy/ipk Apr 12 15:53:05 hi Apr 12 17:09:46 anyone know what's going to happend to OE after the BK torubles? Apr 12 17:36:16 nettie: no word yet. Apr 12 17:56:22 yeah I'm rather curious about that myself Apr 12 17:57:20 nettie: OE will just move to another CM system, and we will move with them. Apr 12 17:57:55 kergoth is evaluating different alternatives now - we've got until July to move. Apr 12 18:02:36 In reality the stuff we use BK for in nslu2-linux could be done with just about any CM system which is at least as good as CVS or SVN. The only thing we would miss is the two-step nature of pushing stuff to nslu2-linux and then kergoth pulling it into oe proper. Apr 12 18:07:52 03jp30 * 10unslung/make/libbt.mk: fix libbt build Apr 12 18:19:31 how's that? Apr 12 18:20:00 you getting changes from us is far, far, far, far less painful with a peer to peer system like bk or tla than with anything else, since your repo can easily be just a branch of ours Apr 12 18:20:11 cvs or svn would be a huge pain in the ass, and that's understating it Apr 12 18:20:29 hi kergoth Apr 12 18:20:42 kergoth: I don't think any of the nslu2-linux users keep multiple clones. we all just sync to the nslu2-linux clone. Apr 12 18:20:48 ? Apr 12 18:20:50 you're not listening. Apr 12 18:20:57 the nslu2-linux oe repo is a clone of ours Apr 12 18:21:01 i dont give a shit what users do Apr 12 18:21:10 i'm talking about how changes flow between nslu2-linux and upstream oe Apr 12 18:21:11 and back Apr 12 18:21:26 with a client/server based system like cvs or svn, your tree wouldnt be a branch of ours Apr 12 18:21:30 so you'd have to pull the changes over manually. Apr 12 18:21:33 likely via diff/patch. Apr 12 18:22:10 kergoth: I'm not stating what you should decide (which would obviously need to be much more sophisticated than CVS or SVN). I was stating that nslu2-linux people have much less sophisticated needs,. and so anything you decide will be more than good enough for us. Apr 12 18:22:35 your _users_ have limited needs of your repo Apr 12 18:22:42 your repo has more advanced needs, to stay in sync with us. Apr 12 18:22:53 yes, that is exactly what I was saying. Apr 12 18:23:14 I think my previous observation got lost... I'm workign on nslu2-linux with two working repos and four active (non-backup) ones... Apr 12 18:23:15 thanks for putting it more concisely :-) Apr 12 18:23:46 My needs certainly go beyond CVS - if I use remote CVS I loose local source managment! Apr 12 18:24:04 jbowler: agreed Apr 12 18:25:18 I really like bk, but monotone seems equivalent (I don't care about the GUI) - it seems to me that the OS development community will have bk-equivalent functionality by July. Apr 12 18:26:51 anyway, I'm happy to leave the decision in kergoth's capable hands, and just follow whatever OE does. Apr 12 18:27:30 (as long as he doesn't decide to use ClearCase :-) Apr 12 18:28:20 Or require everyone to buy a bk license ;-) Apr 12 18:28:51 how much is an individual bk license? Apr 12 18:30:22 What is it anyway? Per seat or per server (to use the MS licensing terminology) - i.e. per user or per-repostitory? Apr 12 18:31:22 Still, this is moot - McVoy seemed to be happy to give $0 licenses to kernel dev, so long as they didn't work at OSDL. Apr 12 18:35:52 what's the definition of "kernel dev" ? Apr 12 18:36:18 per user, $2600/seat (negotiable?) http://www.linux-mag.com/content/view/457/2244/ Apr 12 18:36:39 kernel dev = (I believe) write access to kernel (sp?).bkbits.net Apr 12 18:38:04 $2600 per year. Apr 12 18:39:00 The article isn't clear - I think that's $2600/year/seat, which means $13000/seat outright purchase Apr 12 18:39:30 The McVoy comment is: Larry suggested, "if Linus and Andrew and the others moved elsewhere, we'd glady comp them licenses", referring to their current employment with OSDL''' from http://kerneltrap.org/node/4966 Apr 12 18:39:57 maybe now linux will start using a true open source solution Apr 12 18:40:50 That's a certainty - the April 6 Torvalds posting said that he was spending a week off line testing out what to use... Apr 12 18:41:58 cool Apr 12 18:43:59 There's a non-SCM aspect too - bkbits.net provides a good, high perf, high reliability central server. Whatever people use in place of it won't be as good (sourceforge?) Apr 12 18:45:59 * jacques shudders at the mention of sourcescourge. Apr 12 18:50:36 [g2]: my depmod(etc) fix means that RDEPENDS += kernel-module-vfat automatically pulls in kernel-module-fat, I'd like to simplify the openslug-image packages and move the file system stuff to .conf to simplify local configuration - thoughts? Apr 12 18:55:18 jbowler: the idea is that openslug-image is meant to have everything that you need for the defined openslug baseline functionality (not sure whether this includes VFAT or not). The stuff in the conf is then the optional stuff. Apr 12 18:56:44 so someone can get a bare-bones openslug image by defining the EXTRAS in their local.conf to be nothing. Apr 12 18:57:15 but the default distributed binary includes whatever is in the openslug.conf by default. Apr 12 18:57:52 Right; I would add kernel-module-ntfs (requested twice on the list) plus ufs and iso9660 (CDROM). Apr 12 18:58:14 so I think I agree that vfat support should be in openslug.conf, not openslug-image, so that those who don't want vfat support (noting that nothing else in the base image relies on vfat support) can easily remove it by simply not including it in their local.conf Apr 12 18:58:28 yes, same for them Apr 12 18:58:50 What about kernel-module-msdos? 'cause in fact it is pretty useless - it just hides the DOS long file names! Apr 12 18:59:09 I'm attempting to setup a 1.0GB USB flash drive to hold programs for unslung. Apr 12 18:59:20 we just have to be careful to ensure that nothing depends on something in EXTRAS in openslug.conf, knowing that some people will not have those extras installed. Apr 12 18:59:38 So far, the nslu2 crashes hard during formatting. Apr 12 18:59:51 (unslung version 3.18) Apr 12 19:00:07 Are 1GB USB drive known to be problematic? Apr 12 19:00:10 rwhitby: I wasn't suggesting that, I was suggesting installing them in the flash still (except maybe msdos) Apr 12 19:01:24 jbowler: things in openslug.conf still end up in the flash image. but only if you don't override that variable in your local.conf Apr 12 19:01:57 I'm saying that nothing in openslug-image should depend on something which is in the EXTRAS variable in openslug.conf Apr 12 19:01:58 Ok, my misunderstanding - yes, that's what I intended. Apr 12 19:02:30 Right - I'll do a sanity check build (zap the EXTRA_ stuff to empty). Apr 12 19:03:19 Cause, for instance, VoodooZ will have a much smaller EXTRAs than the default, so he has room for his AI software in the flash. Apr 12 19:03:43 Anyone know the steps the web interface does during a format? Maybe I can try the process manually, and see where it is bomging out. Apr 12 19:04:04 Yes: I did that with 'jonslug' - just kept stripping stuff out until it didn't boot ;-) Apr 12 19:04:43 jbowler: cool. BTW, I am very glad that you have become an active OpenSlug developer :-) Apr 12 19:05:00 <[g2]> hey jbowler ! I'm just reading the comments you made, I'll reply in a minute Apr 12 19:05:31 There's probably some ~ thingy in here for expressing embarassment isn't there, ~redface? Apr 12 19:06:26 stat64 fix: current no-go. It fixes linux-uclibc and breaks glibc (linux) in the exact (opposite) way. Apr 12 19:06:47 jbowler: this is a meritocracy - no need for embarrassment due to identified merit. Apr 12 19:07:21 <[g2]> rwhitby-web, I happy too that jbowler is such an active OpenSlugger too, especially when he said he "like to work alone" :) Apr 12 19:08:24 yes, we have drawn him into the web :-) Apr 12 19:12:11 <[g2]> rwhitby-web, jbowler I agree on the .conf and EXTRAs Apr 12 19:12:55 <[g2]> I may not always be a good about following such a rule, but it certainly makes sense Apr 12 19:12:59 [g2]: I hope so, as we'd discussed that previously :-) Apr 12 19:13:18 Ok, I'll get that one pushed next - it adds OPENSLUG_FILESYSTEMS to openslug.conf with instructions (comments) Apr 12 19:13:27 <[g2]> rwhitby-web, we've discussed lots of stuff :) Apr 12 19:13:39 <[g2]> jbowler, great Apr 12 19:13:50 [g2]: so my AccessSlug and MediaSlug will both be running OpenSlug soon .... Apr 12 19:14:01 <[g2]> Cool Apr 12 19:14:08 <[g2]> I natively built the kernel today Apr 12 19:14:36 <[g2]> not with the IXP4xx stuff, but the current default OpenSlug kernel Apr 12 19:16:16 self-hosting. nice :-) Apr 12 19:16:42 <[g2]> soon Apr 12 19:17:21 <[g2]> I also got a JTAG setup for the gateworks board today Apr 12 19:17:37 <[g2]> I zeroed out the Redboot partition Apr 12 19:17:45 <[g2]> then reflashed it Apr 12 19:18:06 <[g2]> so I'll be doing a bunch of testing with APEX on the avila (gateworks) board Apr 12 19:22:01 back later Apr 12 20:26:21 Pinback: check the custom partition sizes wiki page Apr 12 20:26:59 we know of no issues with large USB drives. but you may well be the first to try a 1GB drive. Apr 12 20:33:12 I've tried setting up the partition entries using busybox, and also with cfdisk on my suse box. Apr 12 20:34:04 I think the one time I saw an error message it was failing at swapon time. Apr 12 20:35:56 But do I understand right that it is necessary to trick the NSLU2/unslung to get the flash to be formatted/used like a HDD? Apr 12 20:37:13 Almost feels like a (flash) geometry problem. Apr 12 20:43:16 your understanding is corret Apr 12 20:43:24 oops, he's gone. Apr 12 20:53:48 <[g2]> nite guys Apr 12 21:17:12 I'm looking for an opinion on how to package something that has no released tarball -- only available via svn. I've got a tarball snapshot I've made, but where to host it? Apr 12 21:17:56 This is libvorbisidec, "Tremor" - the integer decoder for ogg/vorbis Apr 12 21:27:59 i've seen a couple of packages getting source from CVS repo Apr 12 21:28:19 i guess svn should take a similar approach Apr 12 21:30:02 but it probably makes sense to use a snapshot tag instead of head (or the equivalent thing in svn) Apr 12 21:33:03 not to mention that if it's straight from a repo, it typically has to be autogened, and then strange stuff -- like autotool dependancies and stuff. Easier to have a make dist'ed tarball. Apr 12 21:33:36 Which I have, but I guess the real question is whether or not that's kosher, and where that tarball would be hosted Apr 12 21:39:01 rpedde: I am VERY interested in that package. Apr 12 21:39:55 is the tarball all the necessary stuff to build it? Apr 12 21:41:29 rpedde, ping Apr 12 21:41:38 Yeah. I have a package that i've been running for a month and a half, along with a patched vorbis-tools that makes oggdec use libvorbisidec. Apr 12 21:41:38 transcodes in better-than-realtime. Apr 12 21:41:38 ideal for slimserver, etc. Apr 12 21:41:41 I've just been, well... slow in committing it. :/ Apr 12 21:41:58 I've been reading your Blog and website. Apr 12 21:42:53 for things that are hard to get (or missing) we have been hosting them so that they can be gotten by the group to get. Apr 12 21:43:11 when HH.org was down we used a tarball stash for builds. Apr 12 21:43:45 I can throw it up on my site and get it committed now, but I don't have a feel for how much bandwidth that would be. Apr 12 21:43:57 I don't want to slam myself. Apr 12 21:44:00 how big is the tarball? Apr 12 21:44:30 Errr... Apr 12 21:44:39 Pretty small -- 380k. Apr 12 21:44:54 probably not a problem, for me to host, but I don't know how "official" that is Apr 12 21:45:16 we actually can host it at nslu2-linux.org. Apr 12 21:45:20 if you want. Apr 12 21:47:02 I also have a sqlite 2.8.x package that coexists with the sqlite3.x package. The 3.x package that's in there is called "sqlite". Should I call my 2.8 package sqlite2? Apr 12 21:47:47 it would be good to differentiate it. Apr 12 21:47:51 :) Apr 12 21:48:56 rpedde: we can put in in nslu.sf.net/downloads Apr 12 21:49:14 k. Apr 12 21:49:18 you can have write access to that area to update it Apr 12 21:50:06 Okay... I'll put it there. Apr 12 21:51:28 ok, you should be able to scp it up there now. Make sure you version the tarball so we can transition to new releases seamlessly. Apr 12 21:52:17 it's 'libvorbisidec-cvs-20050221.tar.gz', and the package matches that Apr 12 21:52:19 :) Apr 12 21:52:43 look in the Unslung Makefile "upload" target to see the full pathname - replace ".../feeds/...." with ".../downloads/libvorbisidec-cvs-20050221.tar.gz Apr 12 21:52:57 rwhitby-web: this was the last piece of the puzzle to make the wl-hdd do what I need. :-) Apr 12 21:53:27 cool Apr 12 21:53:36 rpedde, does this also have vorbis-tools? or is that separate? Apr 12 21:53:59 I have a vorbis-tools.mk that patches oggdec to use libvorbisidec Apr 12 21:54:09 But it's pretty intrusive Apr 12 21:54:32 can you call it vorbis-itools? Apr 12 21:54:33 I'm not sure how to package that. Apr 12 21:55:01 I want to use ogg123 to decode a stream. Apr 12 21:55:03 Or maybe just make oggdec, and call it oggdect or something Apr 12 21:55:34 I don't have ogg123 patched, but it should be pretty easy. It's *almost* a drop-in replacement for libvorbisfile. Not entirely, but pretty close. Apr 12 21:56:06 03rpedde * 10unslung/ (3 files in 2 dirs): Initial commit of sqlite 2.x, to coexist with sqlite3. Apr 12 21:56:51 so we would need to tell ogg123 to use libvorbisidec. Apr 12 21:57:04 instead of libvorbisdec Apr 12 21:57:12 Instead of vorbisfile Apr 12 21:57:30 okay. Apr 12 21:57:42 I was going to do that, but I got all I needed from oggdec -- I'm just ogg->wav'ing, so that was good enough for me. Apr 12 21:58:19 I can look at the mpg123, though... shouldn't be too bad. Lemme get the lib uploaded and commited first. Apr 12 21:58:38 s/mpg/ogg/ Apr 12 21:58:55 I"m ogg->usbaudio. Apr 12 22:00:34 You got libao to work on usbaudio off the slug? Cool... Apr 12 22:07:12 ka6sox-office: what are you using for the audio output plugged into the USB slot ? Apr 12 22:08:39 Griffin iMic Apr 12 22:09:07 nic Apr 12 22:09:10 nice Apr 12 22:09:34 I think that this will port nice to the wl-hdd Apr 12 22:09:47 ooh.. and I have one of those laying around... I didn't know those worked... Apr 12 22:10:32 can you say wl-pod? Apr 12 22:12:12 woah! get out! I hadn't seen the wl-hdd. That's veeeery cool. That runs the same hardware as the slug? Apr 12 22:13:03 mipsel Apr 12 22:13:13 so similar to openwrt. Apr 12 22:13:28 wl500g.dyndns.org Apr 12 22:13:37 but oe is running on it? Apr 12 22:13:41 geez... where have I been? Apr 12 22:14:18 it uses unlung packages. Apr 12 22:14:34 so its the std crosstool builds. Apr 12 22:15:06 gack... I'm ordering one. must have wireless portable daap server. Apr 12 22:15:29 I ordered one from Newegg tonight...$83 Apr 12 22:16:31 mine will run off of CF (512mb or 1GB) Apr 12 22:17:15 I've got a 40gb laptop drive hanging around here... that would be perfect. Apr 12 22:21:05 in my case I will be streaming to it...and it only needs a small amount of space for packages and *some* local audio..but not much Apr 12 22:22:04 03rpedde * 10unslung/ (make/libvorbisidec.mk sources/libvorbisidec/control): Initial checkin of libvorbisidec -- the ogg vorbis integer decoder Apr 12 22:23:11 Now, as far as vorbis-tools goes... does it make any sense to have non-tremor decoders in ogg123 and oggdec? Apr 12 22:24:07 not probably as we need the integer versions anyways. Apr 12 22:24:09 I mean, should there be both? Because native oggdec is something like 10x realtime. It's essentially unuseable Apr 12 22:24:31 we need it to be *small* and work. Apr 12 22:26:39 hrm.. I didn't strip the libs... I guess I should have, eh? Apr 12 22:27:27 the wl-hdd is only 16MB. Apr 12 22:27:45 Stip libs, compile -Os ;-) Apr 12 22:27:53 s/stip/strip/ Apr 12 22:27:55 oof. yeah... probably should strip them. :) Apr 12 22:32:02 its not like we are going to run gdb on these libraries. Apr 12 22:33:59 gdb isn't much help anyway unless the compilation is done with -g (which adds enormous amounts of symbol data...) Apr 12 22:34:26 So what's in a binary that's unstripped, but not compiled with -g? Apr 12 22:34:42 There is stuff there, but it's not debugging symbols... what is it, exactly? Apr 12 22:34:56 just curious... Apr 12 22:36:21 function and data symbols - extern or static values. Some of that may be required by ld.so, but I think the dynamic link stuff is in a separate section (not sure about that.) Apr 12 22:38:58 Any idea why if it wasn't compiled -g it wouldn't automatically jettison that baggage? Or is there value in it I don't see? Apr 12 22:41:19 ld might do so when building a DLL - I'm not sure. The function names have value for the simple back trace (though I don't know whether that's possible on MIPS with the standard compilation options) Apr 12 22:42:26 Is google broken, or is it just me? Apr 12 22:42:40 The main reason is historical - before ELF (and dwarf and so on) that was the only symbol information. Apr 12 22:43:06 sorry... when I go to google, I get redirected to google.com/xhtml -- the pda version. strange. Apr 12 22:43:30 google: works for wl-hdd, for me... Apr 12 22:43:51 but it is indeed the pda version Apr 12 22:44:57 firefox and epiphany both redirect to /xhtml for me. Apr 12 22:45:29 I'm getting it with galeon too Apr 12 22:45:32 Konqueror (on gentoo): no - I can select xhtml by typing it in, but no redirection Apr 12 22:45:46 heh.. I was just apt-getting konq to test it. Apr 12 22:46:07 and also with mozilla Apr 12 22:46:12 firefox, epiphany and galeon are all gecko based Apr 12 22:46:18 rpedde, I"m getting xhtml too! Apr 12 22:46:27 so probably no surprise there. bad browser sniffing? Apr 12 22:46:37 might be b0rk3n Apr 12 22:46:39 I thought I was nuts for a second. :) Apr 12 22:46:44 on google's side Apr 12 22:46:54 lemme try konqui Apr 12 22:47:11 konqui works right Apr 12 22:47:37 score one for google. Apr 12 22:47:46 lol Apr 12 22:48:01 they don't often blow it. Apr 12 22:48:29 at least not this long... I've never caught them in an obvious thing like this before. Apr 12 22:49:29 * ka6sox-office sees egg all over the face of g00gle. Apr 12 22:51:38 bah, only the 1 in 20 that use non-IE will ever notice Apr 12 22:52:17 there are a number of us who *don't* do windows. Apr 12 22:52:21 i just want /. to get the story so google fixes it Apr 12 22:52:22 exactly. and probably it works on ff on win32, because the *have* to have tested that after rolling in a code change, so it's probably just mozilla users on linux, a vanishingly small fraction of users Apr 12 22:52:50 ka6sox-office, I'm in that number :-) Apr 12 22:52:52 if I actually had a windows machine, I'd verify that... Apr 12 22:52:59 I tried all the Mozilla identifications Konqueror supports - no redirection Apr 12 22:53:06 if I have to use windows, I install firefox Apr 12 22:53:21 I *do* have a windows machine but its offline Apr 12 22:53:55 aha! Apr 12 22:54:00 it's only 216.239.39.99 Apr 12 22:54:00 strange... konq spoofing a moz user-agent works? Apr 12 22:54:13 216.239.37.99 and 216.239.57.99 are fine Apr 12 22:54:38 maybe first in a rolling upgrade? Apr 12 22:54:40 used for VHDL programming and programming obscure things that require DOS and a parallel port. Apr 12 22:55:06 sounds like "world of warcraft" to me. Apr 12 22:55:25 WoW is online... Apr 12 22:55:33 I play NWN on my Macs Apr 12 22:55:49 konq spoofing - works in what sense? It just changes the browser id string, google may be sniffing more stuff but there isn't that much to look at... Apr 12 22:56:15 jbowler, go to 216.239.39.99 Apr 12 22:57:24 216.239.39.99 in konq works... no redirection. Apr 12 22:57:43 hmm Apr 12 22:57:47 Same - no redirection. I did get a cookie from it (which I rejected) - could google be using a cache of my IP (which is masqueraded) Apr 12 22:58:09 ah, they just fixed it Apr 12 22:58:24 it's working for me too now on 216.239.39.99 Apr 12 22:58:32 oops Apr 12 22:58:34 now it's not Apr 12 22:58:40 maybe it's a load balancer Apr 12 22:58:43 its fixed... Apr 12 22:58:50 (or I got the *good* ones.) Apr 12 22:58:56 and some of the servers on the other side are not fixed yet Apr 12 22:59:27 I can sit there and reload 216.239.39.99 and about 50% of the time I get redirected now Apr 12 22:59:44 I guess eventually they'll fix all the servers Apr 12 23:02:03 guess the fun is over, then :) Apr 12 23:02:11 ah.. Apr 12 23:02:24 well..I've found the bug in my VHDL so I can go home. Apr 12 23:03:43 ka6sox-office, congrats :-) Apr 12 23:04:25 thanks...it was a PITA to find...buried in a nested state machine. Apr 12 23:10:18 ka6sox: you have a slug with audio right now? Apr 12 23:10:40 yes..but I need to fix libao Apr 12 23:11:06 it works on OpenSlug Apr 12 23:11:09 because I can patch ogg123, but I can't test it with no audio out Apr 12 23:11:09 with Coreaudio Apr 12 23:11:19 nite Eno Apr 12 23:11:47 if you patch it I'll sync with the repo and build it. Apr 12 23:11:54 and then test it. Apr 12 23:12:07 I think I'm close here Apr 12 23:12:13 cool Apr 12 23:15:58 catch you on the flipside Apr 12 23:16:32 aight.. I'll get this to where I think it's working and commit it. It'll be called ivorbis-tools Apr 12 23:30:54 Well, it compiles but I can't actually *play* anything, as I have no audio and libao gets all mad Apr 12 23:36:38 libao is b0rk3n Apr 12 23:36:59 I"m a hardware guy but am learning this *software* stuff ;) Apr 12 23:37:01 mkay Apr 12 23:37:08 I can't even get the wav target to work. Apr 12 23:40:24 :( Apr 12 23:42:17 03rpedde * 10unslung/ (3 files in 2 dirs): Initial checkin of vorbis-tools compiled for Tremor - the ogg vorbis integer decoder Apr 12 23:47:33 03rpedde * 10unslung/make/libvorbisidec.mk: configure for arm assembly -- significantly faster Apr 12 23:48:58 mkay... if libao worked, everything else would, I think. At least, I've got the ivorbis-tools installed and it does oggdec okay. I'm guessing that ogg123 would work if I could make it output to file. Apr 12 23:50:23 sounds good. Apr 12 23:56:01 if you compile libvorbisidec for mips, you'll want to comment out the -D_ARM_ASSEMBLY_ define Apr 12 23:56:38 I do intend to do that if I can. Apr 12 23:57:20 I am going to order that thing from newegg, so I might be back here bugging you. Apr 12 23:57:41 well...its not bugging me! Apr 12 23:58:06 I appreciate your help here. Apr 12 23:58:47 and checking in those codecs Apr 12 23:59:28 I needed to be pushed into doing it anyway.. my packages were diverging from unslung and people on mt-daapd nightlies were complaining. :) **** ENDING LOGGING AT Tue Apr 12 23:59:56 2005