**** BEGIN LOGGING AT Wed Jan 13 02:59:57 2010 Jan 13 03:09:27 sigh, so many Xorg packages Jan 13 03:17:45 swalker: why the sigh? Hate Xorg? :) Jan 13 03:30:08 not particularly Jan 13 03:31:23 swalker: you working on those packages? Jan 13 04:19:14 cshore: I'm in the process of finally adding automated upstream checking for them for my upstream page Jan 13 04:19:54 cool...that's the chart of upstream vs in OpenWRT packages? Jan 13 04:20:46 right Jan 13 05:21:52 lars * r19113 /trunk/tools/ (libuuid/Makefile mtd-utils/Makefile): Fix libuuid header location and reenable building of mkfs.ubifs Jan 13 06:52:13 juhosg * r19114 /trunk/target/linux/ar71xx/ (6 files in 3 dirs): ar71xx: fix 320S33B flash chip support Jan 13 07:23:35 updated openwrt/upstream, http://pastehtml.com/view/5sxrat8.html Jan 13 08:50:53 <[florian]> cshore: pong Jan 13 08:51:24 [florian]: hi! Jan 13 08:51:55 [florian]: erm, what's my username on the svn, and does it use my SSH public key? Jan 13 08:52:10 <[florian]> cshore: check your mail Jan 13 08:53:53 [florian]: which one (vertical, gmail, or csolve?) Jan 13 08:53:58 <[florian]> cshore: csolve Jan 13 08:54:33 ok, great, thanks! Jan 13 08:54:56 <[florian]> cshore: please have a try :) Jan 13 08:56:42 well the checkout is working Jan 13 09:00:27 [florian]: if a log message is longer that 80 chars, should I use newline, or should I leave it one long line, and let it wrap? Jan 13 09:01:49 <[florian]> cshore: let it wrap, that's what we almost all do Jan 13 09:02:02 [florian]: also, did you get a chance to add the -A option to dnsmasq or should I do that? Jan 13 09:02:25 <[florian]> cshore: I invite you to read this page: https://dev.openwrt.org/wiki/CommitPolicy Jan 13 09:02:36 I found that one Jan 13 09:02:37 <[florian]> cshore: you should do that, with xMff agreement Jan 13 09:02:50 I got Access denied Jan 13 09:04:26 <[florian]> cshore: please make sure that your ssh key is in your ssh-agent Jan 13 09:05:18 <[florian]> cshore: and that you use the correct username and protocol to connect to the svn server, I will check with Kaloz Jan 13 09:06:12 ok, I think it's probably the username (protocol is ssh+svn, right? Jan 13 09:06:22 <[florian]> yep Jan 13 09:06:29 <[florian]> username should be the one I mailed you Jan 13 09:08:39 okay, and it worked for pulling, but not for pushing Jan 13 09:09:31 <[florian]> are you using svn or git-svn? Jan 13 09:09:46 svn Jan 13 09:10:05 <[florian]> ok, please make sure that your svn copy is fully transitionned to svn+ssh by using svn switch --relocate Jan 13 09:10:26 <[florian]> if it worked for pulling and did not ask for your ssh passphrase it probably did not use the ssh version Jan 13 09:10:45 I did a fresh checkout, svn co svn+ssh://svn.openwrt.org/openwrt/trunk openwrt Jan 13 09:11:14 <[florian]> still wrong as you are missing the username here Jan 13 09:11:29 <[florian]> you really need the username to be set Jan 13 09:11:53 sorry...I did the username...without the username prompted for a password, so I realized I did something wrong Jan 13 09:12:35 <[florian]> no problem Jan 13 09:13:23 <[florian]> is it better now? Jan 13 09:14:10 okay, I'm trying again: svn co svn+ssh://username@svn.openwrt.org/openwrt/trunk trunk, worked Jan 13 09:14:42 <[florian]> good, I guess you can now commit something :) Jan 13 09:15:43 okay, I applied a patch and used the existing commit message (same patch as I tried before, but got failed), and got Jan 13 09:15:43 svn: Access denied Jan 13 09:15:53 <[florian]> ok, let me check with Kaloz then Jan 13 09:16:30 AIUI svn+ssh should already be using my SSH key, right? Jan 13 09:16:35 that is checkout Jan 13 09:16:48 <[florian]> yes it should be Jan 13 09:17:16 so the problem shouldn't be the key Jan 13 09:17:25 <[florian]> I guess not Jan 13 09:17:38 <[florian]> can you still check that the corresponding private key is available? Jan 13 09:17:49 <[florian]> like killing your ssh-agent and starting a new checkout should be ok Jan 13 09:17:57 <[florian]> it should prompt you for the private key passphrease Jan 13 09:21:29 hmm....it's not prompting me Jan 13 09:22:43 ok, I killed gnome-keyring and it prompted me Jan 13 09:25:31 ok, when I tried to commit it prompted the the SSH passphrase, which I entered successfully, but I still got Access denied Jan 13 09:25:49 <[florian]> ok, then it's definitively the account setup which is missing something Jan 13 09:25:54 <[florian]> I have asked Kaloz about this Jan 13 09:27:57 what's xMff's OpenWRT email address (in case he's not online before I go to bed) Jan 13 09:29:38 <[florian]> xm @ subsignal.org Jan 13 09:58:18 Hi Jo, could you look at #5017...I'm thinking of adding that patches, except not removing option local (would break too many people I think), and updating for options that have been added since. What do you think? Jan 13 10:01:35 cshore: yes makes sense, option local should be preserved, the rest is fine Jan 13 10:02:21 cshore: there is also this: http://wiki.openwrt.org/doc/uci/dhcp#custom.domain Jan 13 10:02:42 ... which translates to -A and --ptr-record Jan 13 10:06:03 xMff: also, can you access uci lists in LuCi (i.e. to add some of the dnsmasq options to the dnsmasq LuCI) Jan 13 10:11:05 yes Jan 13 10:17:36 xMff: IIRC that wasn't possible when I first was using LuCI...is it now? Jan 13 10:18:26 juhosg * r19115 /trunk/target/linux/generic-2.6/files/ (2 files in 2 dirs): Jan 13 10:18:26 kernel: change debounce logic in the gpio-buttons driver Jan 13 10:18:26 * thanks to Nuno Gonçalves Jan 13 10:18:31 juhosg * r19116 /trunk/target/linux/ar71xx/files/arch/mips/ar71xx/ (22 files): ar71xx: reduce button threshold values Jan 13 10:47:16 hee, i've got my retailer chasing Lex for the source code to my infineon adsl card Jan 13 10:49:22 lupine_85: wat dsl card? Jan 13 10:49:52 http://blog.lupine.me.uk/2010/01/07/adsl2-pci-modem Jan 13 10:50:02 looks very amenable to OpenWRTification Jan 13 10:50:37 but if I can get the toolchain from the makers, I'm happy to just compile an ipv6 module for the ancient kernel and leave it at that Jan 13 10:56:10 lupine_85: nice it is a danube Jan 13 10:56:19 openwrt has a dsl driver for those alresdy Jan 13 10:56:23 where did you buy the card ? Jan 13 10:56:39 it must be running as a pci device not master Jan 13 10:56:44 veryinteresting indeed Jan 13 11:01:57 blogic: www.linitx.com Jan 13 11:02:01 £35! cheap as chips Jan 13 11:02:54 (it's a full SBC that can talk to my network through a total of three different kinds of ethernet interface - including the PCI one) Jan 13 11:03:40 so this is a pci card that has an embedded computer on it which runs linux? Jan 13 11:04:16 yes Jan 13 11:04:21 crazy Jan 13 11:04:25 so says everyone else Jan 13 11:04:40 sounds like a must-have for the obscure hardware collector Jan 13 11:04:43 but if you come up with a PSU and some electrical tape, you can get rid of the PCI component :) Jan 13 11:04:55 it has pins for 5V DC Jan 13 11:05:20 it's like: "yo dawg, i heard you like computers, so we put a computer in your computer" Jan 13 11:05:56 hehe Jan 13 11:07:06 http://linitx.com/viewproduct.php?prodid=12181 Jan 13 11:07:07 this one ? Jan 13 11:07:13 can you make a photo of it ? Jan 13 11:07:28 openwrt boots on several danube boards already and dsl is known to be working Jan 13 11:07:34 there's a photo on the lex.com.tw website Jan 13 11:08:07 url? Jan 13 11:08:50 http://www.lex.com.tw/product/images/ADSL2+modem-big.jpg :) Jan 13 11:09:19 that links times out for me Jan 13 11:09:29 i get it, dcc? Jan 13 11:09:30 tw dont like me Jan 13 11:09:30 workshere Jan 13 11:09:33 blogic: by "known to be working", do you mean there's an open-source / reverse-engineered driver for the infineon adsl chip included in the openwrt distribution Jan 13 11:09:34 ? Jan 13 11:09:39 * lupine_85 puts the image elsewhere Jan 13 11:09:44 FOSS driver Jan 13 11:09:48 has a firmware blob Jan 13 11:09:55 cool :) Jan 13 11:09:59 ifx contributed it to owrt and i merged it last month Jan 13 11:10:01 well Jan 13 11:10:04 2-3 months ago Jan 13 11:10:05 blogic: http://imagebin.ca/view/STwOebCs.html Jan 13 11:10:16 jow_laptop: thx Jan 13 11:10:49 blogic: that is pretty awesome Jan 13 11:10:55 lupine_85: it is indeed Jan 13 11:11:05 so in theory, I should just be able to flash the ifxmips package onto the board and it'll "just work"? Jan 13 11:11:13 I love that "in theory" Jan 13 11:12:02 http://www.lupine.me.uk/img/random/ADSL2+modem-big.jpg Jan 13 11:13:31 as you said, in theory Jan 13 11:13:37 you need to make several things Jan 13 11:13:51 first try to find out if there is a way to enable serial uplaod on the board Jan 13 11:15:25 do devices like these have a bootloader too? Jan 13 11:17:02 yep Jan 13 11:17:03 uBoot Jan 13 11:17:27 we think there's a serial header on the board but I've not had chance to try it Jan 13 11:17:37 if you can get access to that, that should allow for an upload Jan 13 11:20:41 danage: yes Jan 13 11:20:47 lupine_85: you need uboot access really Jan 13 11:21:43 aye, bit of a pain Jan 13 11:22:01 is there a way to flash the device from having root access to it in linux? Jan 13 11:22:24 (the retailer is willing to give it a go and sell them with openwrt on instead of the lex firmware if it works) Jan 13 11:22:44 or is the serial console really needed? Jan 13 11:24:07 can I jut use the web-interface's firmware upload, if I don't care whether it bricks the device or not? Jan 13 11:24:28 well if it bricks it, you'll need the serial console :) Jan 13 11:24:33 heh Jan 13 12:07:51 lupine_85: well telnet or ssh access might give you pretty good functionality too Jan 13 12:48:07 danage: basically, I've got someone who's willing to give a blind flash a try if I give him a walkthrough Jan 13 13:14:36 lars * r19117 /trunk/tools/Makefile: mtd-utils depends on libuuid Jan 13 16:35:31 ping xMff Jan 13 16:36:09 pong cshore (I'm still at work) Jan 13 16:37:43 I'm working on preinit, mount_root, and firstboot. I've compined preinit/mount_root into /etc/preinit.d/ and hooks, and I'm working on doing something similar under /lib/firstboot for firstboot. but there is a catch Jan 13 16:38:13 mount_root and firstboot want to share some code...where should the shared code be kept? Jan 13 16:39:36 maybe /etc/functions.sh ` Jan 13 16:40:25 that seems like bloating functions.sh too much to me. Isn't functions.sh loaded for a lot of scripts in OpenWRT? Jan 13 16:42:57 how about making /etc/functions.d/boot (with the idea that /etc/functions.d could become a place where functions that were not used often enough for /etc/functions.sh could be stored) Jan 13 16:44:01 yeah, I did something similar years ago when working on a whiterussian based firmware Jan 13 16:44:27 but actually it does not matter where you put it, if you're going to split it up into small chunks then put it in /lib Jan 13 16:44:34 okay Jan 13 16:44:42 /lib/functions? Jan 13 16:44:46 /etc is not really appropriate for that imo, it clutters backups and stuff Jan 13 16:45:13 hmm...should my preinit stuff go under /lib too then? Jan 13 16:45:25 it's not configurable by the user, only at build time Jan 13 16:45:36 probably Jan 13 16:45:56 the thing is, if I backup /etc I expect it to contain more or less only configuration files Jan 13 16:46:09 right...makes sense to me Jan 13 16:46:13 now imagine I backup /etc, flash a newer firmware and restore the backup Jan 13 16:46:22 then I'll overwrite integral system parts with older versions Jan 13 16:46:33 hmmm...yes that would be bad Jan 13 16:51:13 really /etc/function.sh should be /lib/functions/function.sh or something, but it's too widely used now to change, eh? Jan 13 16:52:38 hm well I'd like to see it in /lib too, but we should discuss that with the others Jan 13 16:53:17 the code in the repo can be changed, problem is the 3rd party stuff Jan 13 16:53:48 right Jan 13 16:54:15 would a symlink work? Jan 13 16:54:28 sure Jan 13 16:59:28 allright, heading home. bbl Jan 13 18:39:47 jow * r19118 /trunk/package/ (2 files in 2 dirs): Jan 13 18:39:47 [package] base-files, mac80211 (#6491) Jan 13 18:39:47 - fix wep key handling in iw with mac80211 based drivers Jan 13 18:39:47 - sanitize keys where necessary Jan 13 18:39:47 - put a procedure prepare_key_wep() into /sbin/wifi for use by other driver backends Jan 13 19:25:46 xMff: thanks :-) Jan 13 19:31:49 jow * r19119 /trunk/package/wpa_supplicant/files/wpa_supplicant.sh: [package] wpa_supplicant: add WEP key handling to conform with documented behaviour in mac80211.sh (#6493) Jan 13 19:37:09 xMff: ahh, I was just checking if that prepare_key_wep procedure could also be used in the changes I made to wpa_supplicant :-) Jan 13 19:38:37 xMff: nvm, seems you already took that into account, great. Jan 13 19:38:54 jow * r19120 /trunk/include/download.mk: [buildroot] allow download urls using the file:// schema (#6480) Jan 13 19:39:20 stintel: yep Jan 13 20:21:41 jow * r19121 /trunk/package/opkg/patches/003-fs_overlay_support.patch: [package] opkg: fallback to dest->root_dir if specified overlay_root does not exist on filesystem (#6498) Jan 13 20:31:45 jow * r19122 /packages/libs/openldap/ (Makefile files/ldap.init): [packages] openldap: remove dependency on db74 (#6476) Jan 13 20:34:32 jow * r19123 /packages/libs/openldap/patches/020-autofs-schema.patch: [packages] openldap: add autofs/automount schema suitable for Mac OS X, Linux and others (#6454) Jan 13 20:36:56 jow * r19124 /packages/net/krb5/ (Makefile files/krb5kdc patches/002-krb5kdc-dir-to-etc.patch): [packages] krb5: move persistent database to /etc (#6475) Jan 13 20:47:56 anyone know why sox just hangs when I try to use it? I do get the help screen, but nothing else seems to function... Jan 13 20:48:10 strace it Jan 13 20:48:19 (no) Jan 13 20:49:33 ok, thats a good idea Jan 13 20:50:36 :) Jan 13 21:07:46 <[florian]> cshore: ping? Jan 13 21:07:57 [florian]: pong Jan 13 21:08:09 <[florian]> cshore: does your svn account works now? Jan 13 21:08:34 <[florian]> cshore: check your query Jan 13 21:08:37 * Bartman007 joins the convo Jan 13 21:09:51 my query? Jan 13 21:09:57 cshore: this morning you pasted you were using svn+ssh://username@, when it should be svn+ssh://openwrt@ Jan 13 21:10:30 I just wasn't used the actual username...I wasn't sure if it was a secret Jan 13 21:10:49 <[florian]> cshore: a query is a personnal message, which should open a new irc window in your client Jan 13 21:11:03 I used svn co svn+ssh://openwrt@svn.openwrt.org/openwrt/trunk Jan 13 21:11:08 <[florian]> cshore: I did not want to paste the complete URL here, sorry if that confused you Jan 13 21:11:24 can you do the query again Jan 13 21:11:44 <[florian]> sure Jan 13 21:13:48 hmmm...I don't seem to be getting it Jan 13 21:14:06 <[florian]> weird, let me try again Jan 13 21:14:41 cshore: that would explain why you never responded to mine this morning, or the one I just sent :) Jan 13 21:14:57 Bartman007: yep Jan 13 21:15:10 could my firewall be a problem? Jan 13 21:15:32 <[florian]> cshore: can you connect to other machines using ssh? Jan 13 21:15:58 <[florian]> cshore: can you check that svn.openwrt.org <-> 78.24.191.177 is ok for you too? Jan 13 21:16:31 [florian]: I can svn local machines. let me try the address you just listed Jan 13 21:16:43 cshore: I would more likely suspect an issue w/ pidgin Jan 13 21:17:30 PTY allocation request failed on channel 0 Jan 13 21:17:30 ( success ( 2 2 ( ) ( edit-pipeline svndiff1 absent-entries commit-revprops depth log-revprops partial-replay ) ) ) Connection to 78.24.191.177 closed. Jan 13 21:17:31 (I assume that's what your using since your client responds to a version query with "Purple IRC") Jan 13 21:17:35 I'd say it worked Jan 13 21:17:39 yep Jan 13 21:17:46 Barman007: smart man Jan 13 21:17:50 or woman Jan 13 21:18:17 Jan 13 21:18:28 I'm not worried Jan 13 21:18:48 and I'm reasonably certain I'm a man Jan 13 21:18:52 :) Jan 13 21:19:53 [Florian]: I can ssh to local machines (i.e. other hosts) with my private key Jan 13 21:20:19 <[florian]> cshore: ok, so I guess the issue might very well be on openwrt'side Jan 13 21:20:22 Bartman007: do you have Jabber Jan 13 21:20:46 yes. agb@jabber.ccc.de Jan 13 21:20:47 <[florian]> I go to bed now guys Jan 13 21:20:55 night [florian] Jan 13 21:20:58 [florian]: gnight Jan 13 21:21:03 <[florian]> thanks Jan 13 21:22:29 ok, I added you (crazycshore@gmail.com), so you can message me if you want...since IRC isn't working in Pidgin. At some point I want to use znc or something like that on my router, which should help Jan 13 21:23:01 queries at any rate Jan 13 21:24:10 xMff: I'm seeing tons and tons of _llseek messages: _llseek(3, 0, [0], SEEK_SET) = 0 Jan 13 21:25:00 netprince_: hm what is fd 3? A file Jan 13 21:25:03 ? Jan 13 21:25:46 Barman007: could you try query again? Jan 13 21:28:57 I would have to guess OggS, I'm new to strace... http://openwrt.pastebin.ca/1750179 Jan 13 21:30:11 guess I could have snipped that file a bit... sorry Jan 13 21:30:43 netprince_: look for something like open("/foo/...", ...) = 3 Jan 13 21:31:21 open("/lib/libc.so.0", O_RDONLY) = 3 Jan 13 21:31:32 that cant be good Jan 13 21:32:12 no Jan 13 21:32:16 wait, here is a closer one to the end: open("/tmp/test/sample.ogg", O_RDONLY|O_LARGEFILE) = 3 Jan 13 21:32:18 it is opened and closed several times Jan 13 21:32:28 open("/tmp/test/sample.ogg", O_RDONLY|O_LARGEFILE) = 3 @ L306 Jan 13 21:32:42 looks okay so far Jan 13 21:32:54 it also reads an initial 64k chunk from the file successfully Jan 13 21:33:17 the subsequent seek attempts look like it tries to read more and fails with that Jan 13 21:33:37 maybe a large-file issue or endianess (bad offsets) Jan 13 21:35:09 I fear without seeding printf()s into the code or using gdb(server) it'll be difficult to tell what's going on Jan 13 21:35:42 how large is the test.ogg file? Jan 13 21:35:48 35M Jan 13 21:35:57 hmm Jan 13 21:37:13 I'm trying to read from stdin instead... Jan 13 21:37:58 hah, seems to work from stdin Jan 13 21:39:08 works for me for now, xMff thanks for that! very helpful. Jan 13 21:39:48 well then Jan 13 21:40:19 did you want me to try something else? Jan 13 21:40:44 no :) Jan 13 21:40:50 lol Jan 13 21:41:25 it works now, don't touch Jan 13 23:23:14 how does one get a kernel tree patched without buliding the kernel? Jan 13 23:24:18 will make prepare do it? Jan 13 23:25:21 make target/linux/{clean,prepare} V=99 QUILT=1 Jan 13 23:25:46 or without QUILT=1 if you want the buildroot to apply the patches as well Jan 13 23:26:09 automatically I mean Jan 13 23:26:20 else you have to do quilt push etc. Jan 13 23:26:24 if the patches are applied automatically can I unapply to where I want via quilt? Jan 13 23:26:29 via quilt pop? Jan 13 23:26:35 only if you use QUILT=1 Jan 13 23:26:49 then you can add them one by one Jan 13 23:26:57 or move back and forth with push / pop Jan 13 23:27:18 what order do they get applied in (that is what dirs first?) Jan 13 23:27:31 generic, target Jan 13 23:27:51 just check patches/series Jan 13 23:27:56 after preparing the tree Jan 13 23:29:47 hmmm...ok how about just building the toolchain (make tools; make toolchain?) Jan 13 23:30:37 make toolchain/compile I think Jan 13 23:38:27 make toolchain/install is probably what you want. Jan 13 23:39:29 actually realized what I was looking for what make tools/compile (maybe install too) Jan 13 23:40:11 or do you need to the toolchain to do make target/linux/prepare? Jan 13 23:41:19 I guess I'll find out...I issued the command Jan 13 23:41:24 when you run just plain `make' in the build system it hits: tools/install toolchain/install target/compile package/cleanup package/compile package/install package/rootfs-prepare target/install (in that order) Jan 13 23:41:53 right...I don't want to do a full make though...just enough to edit the kernel tree Jan 13 23:42:37 you might be able to get away without a toolchain build if you are just editing the tree, but you need the toolchain to build the kernel Jan 13 23:42:47 right Jan 14 01:13:46 ping [florian] **** ENDING LOGGING AT Thu Jan 14 02:59:57 2010