**** BEGIN LOGGING AT Fri Jun 24 23:59:57 2005 Jun 25 00:00:25 maybe a glitch in its version sort Jun 25 00:01:50 hmm... should probably have been openvpn_2.0-rc1 Jun 25 00:02:10 quite possibly Jun 25 00:02:44 duh.. no difference :( Jun 25 00:03:49 :-\ Jun 25 00:06:21 hi ilg_ Jun 25 00:06:27 hi Jun 25 00:06:52 I'm working on the loop mount Jun 25 00:07:08 the /boot script work, but Jun 25 00:07:36 I'm trying to write the script to generate the loop file... Jun 25 00:08:07 for ext2 it was no problem, Jun 25 00:08:37 but for jffs2, if I try to mount it, the kernel issues Attempt to mount non-MTD device "/dev/loop1" as J Jun 25 00:08:37 FFS2 Jun 25 00:09:14 any other ideas as filesystem for flash? Jun 25 00:11:14 why not ext2? Jun 25 00:11:57 ilg_: is this external flash? Jun 25 00:12:06 usb flash Jun 25 00:12:38 vfat, and the future root fs is a file there Jun 25 00:14:50 jffs2 was designed for MTD, external flash is not MTD Jun 25 00:15:17 it's an ATA interface Jun 25 00:15:44 ok, then ext2? Jun 25 00:15:52 yes, that's a good choice Jun 25 00:18:21 the usb stick itself will do wear levelling for the vfat fs Jun 25 00:19:01 I would like to initialize this ext2 with with the /build/tmp/rootfs. do you see any problem? Jun 25 00:19:51 NAiL: try syncing now. I had two heads (one yours and one jbowlers) for org.nslu2-linux.openembedded, so I merged them and pushed back to our repo Jun 25 00:20:14 ilg_: unslung or openslug? Jun 25 00:20:20 openslug Jun 25 00:20:40 then you should be able to mount loop, mke2fs, then turnup Jun 25 00:21:12 rwhitby: Well... almost. I got the patch to nslu2-io.c, but still not the last samba stuff ;) Jun 25 00:21:14 allow turnup to copy the rootfs to it. Jun 25 00:21:31 NAiL: what is the rev id of the samba stuff? Jun 25 00:21:49 uh.. how do I find that? Jun 25 00:22:06 good question. Jun 25 00:22:13 we need viewmtn running on nudi ... Jun 25 00:22:56 BTW, are you all running mtn update after the sync? a sync doesn't touch the working dir. Jun 25 00:23:22 (that's a big difference from bk) Jun 25 00:23:26 5cba07aac0196e37bd9c62615603a532f135c2c4, 3783ac4c44f799b108df17b5007f68dffcf17d43, 07ec89b6d456e2e9e0bd70da845bdef77454b4b6 <-- those three hasn't come through Jun 25 00:24:05 The two last ones are just test stuff Jun 25 00:24:19 rwhitby@linux:/home/slug/openembedded$ monotone cat revision 5cba07aac0196e37bd9c62615603a532f135c2c4 Jun 25 00:24:19 new_manifest [ba491e43a208dc1e315808afe21387ec69f61d04] Jun 25 00:24:19 old_revision [3783ac4c44f799b108df17b5007f68dffcf17d43] Jun 25 00:24:19 old_manifest [2dc115fcbdba400df63486477716df4a5ac66c69] Jun 25 00:24:19 patch "packages/samba/samba_3.0.14a.bb" Jun 25 00:24:21 from [ef4befa1dc2da3dbe1d5d39c24579a15e27a8766] Jun 25 00:24:22 to [fd736ca2de34a77d7df470d9a0c0e3c7173bcdc1] Jun 25 00:24:55 lemmesee... the two *first* ones (they're pasted in reverse order), are just test-stuff Jun 25 00:25:43 Looks like I have all three here. Jun 25 00:25:55 rwhitby@linux:/home/slug/openembedded$ monotone cat revision 3783ac4c44f799b108df17b5007f68dffcf17d43 Jun 25 00:25:55 new_manifest [2dc115fcbdba400df63486477716df4a5ac66c69] Jun 25 00:25:55 old_revision [07ec89b6d456e2e9e0bd70da845bdef77454b4b6] Jun 25 00:25:55 old_manifest [ba491e43a208dc1e315808afe21387ec69f61d04] Jun 25 00:25:55 patch "packages/samba/samba_3.0.14a.bb" Jun 25 00:25:57 from [fd736ca2de34a77d7df470d9a0c0e3c7173bcdc1] Jun 25 00:25:58 to [ef4befa1dc2da3dbe1d5d39c24579a15e27a8766] Jun 25 00:26:00 rwhitby@linux:/home/slug/openembedded$ monotone cat revision 07ec89b6d456e2e9e0bd70da845bdef77454b4b6 Jun 25 00:26:02 new_manifest [ba491e43a208dc1e315808afe21387ec69f61d04] Jun 25 00:26:04 when I mt update my joe blow repo, I don't get the last three Jun 25 00:26:04 old_revision [b3045979914c294a5890ddf1499ce9aa65e093e3] Jun 25 00:26:06 old_manifest [c61d71f48dde0d57bb6dcec9fee93403493101cf] Jun 25 00:26:08 patch "packages/samba/samba_3.0.14a.bb" Jun 25 00:26:10 from [b8bf603e2d69edf5092bd0ebdac0944817d5cc73] Jun 25 00:26:12 to [fd736ca2de34a77d7df470d9a0c0e3c7173bcdc1] Jun 25 00:27:07 let me try my joe blow repo Jun 25 00:27:37 you did a pull from mtn.nslu2-linux.org/org.nslu2-linux, and then an mtn update, right? Jun 25 00:27:59 yeah... Jun 25 00:28:01 this is odd Jun 25 00:28:07 I've got the revisions now Jun 25 00:28:23 So my joe blow repo is up to date Jun 25 00:28:34 as well as my working repo Jun 25 00:28:51 did you check heads when the revs were missing? Jun 25 00:29:18 rwhitby@linux:/home/slug/test/openembedded$ monotone update Jun 25 00:29:18 monotone: selected update target 34f21f15d9abcc06ca233c9c087ef03ee76186c0 Jun 25 00:29:18 monotone: moving packages/samba/files -> MT/tmp/7 Jun 25 00:29:18 monotone: moving packages/samba -> MT/tmp/6 Jun 25 00:29:19 monotone: moving packages -> MT/tmp/5 Jun 25 00:29:19 monotone: moving dir MT/tmp/5 -> packages Jun 25 00:29:21 monotone: moving dir MT/tmp/6 -> packages/samba Jun 25 00:29:23 monotone: moving dir MT/tmp/7 -> packages/samba/files Jun 25 00:29:25 monotone: updating packages/linux/openslug-kernel-2.6.11.2/nslu2-io.c to 2da1d3f5c0aa0804c5769588337077ddcb35c5e9 Jun 25 00:29:28 monotone: updating packages/samba/files/init to 4c1c08ceed1b75cbfdd24d9c3b187a74efa798f0 Jun 25 00:29:30 monotone: updating packages/samba/samba.inc to 4fe5749411f7c835d4187d1e302eaddf8606fe80 Jun 25 00:29:32 monotone: updating packages/samba/samba_3.0.14a.bb to fd736ca2de34a77d7df470d9a0c0e3c7173bcdc1 Jun 25 00:29:34 monotone: updated to base revision 34f21f15d9abcc06ca233c9c087ef03ee76186c0 Jun 25 00:29:35 I was at 5cba07aac0196e37bd9c62615603a532f135c2c4, iirc Jun 25 00:29:40 that's an update on my joe blow repo which was idle for 18 hours Jun 25 00:31:43 You didn't get the nslu2-io.c update? Jun 25 00:31:57 ah, yeah, it's there Jun 25 00:33:25 btw, how do I add a kernel module? (CONFIG_TUN in this case, need to test openvpn) Jun 25 00:35:29 defconfig Jun 25 00:35:57 Nothing more is needed? It's built as an ipk automagically? Jun 25 00:36:01 rebuild the kernel and OE will emit the ipk Jun 25 00:36:18 and next time we build for the feed it will appear there Jun 25 00:36:36 OK, here's my theory on this whole situation: Jun 25 00:37:02 1/ #oe's server is serving org.openembedded.* Jun 25 00:37:13 2/ our server is serving org.nslu2-linux.* Jun 25 00:37:37 3/ when you commit a revision it ends up tied to a particular branch Jun 25 00:37:59 (that's a theory) Jun 25 00:38:25 4/ (theory) when you propagate from one branch to another, it doesn't copy the data, it just puts in a reference Jun 25 00:39:36 cool, whatever updates have gotten through is making my build get farther Jun 25 00:39:48 5/ (theory) when we push back to #oe's server, it ignores any revisions which are not tied to org.openembedded.*, and therefore ends up with references with missing targets, and viewmtn doesn't display those revisions which have references with missing targets. Jun 25 00:39:56 jaques: Where did it stop before? Jun 25 00:40:03 libtiff Jun 25 00:40:40 dyoung said to expect problems with fontconfig but I think my autoconf stuff is new enough and it got past that Jun 25 00:40:46 bbiab - time to kick a ball in the back yard with a 4 yr old :-) Jun 25 00:40:58 :-) have fun Jun 25 00:41:00 rwhitby: Cool. Be nice ;) Jun 25 00:41:55 so my theory is that if both #oe and our servers served org.* then all would be sweet. Jun 25 00:42:05 (cause all revisions would end up in both) Jun 25 00:42:28 the theory is backed up by the fact that when I sync with #oe it keeps sending the same stuff again and again - it doesn't "stick" Jun 25 00:42:48 will check the actual revisions sent later, to confirm they are nslu2-linux branch revisions. Jun 25 00:43:07 rwhitby-away: Yeah, do so. That bit is a tad above my head Jun 25 00:46:09 * NAiL likes monotone Jun 25 00:47:07 (When it works the way I expect it to) Jun 25 00:53:52 oh my build died again Jun 25 00:54:04 on optware mysql Jun 25 00:54:17 checking for correct ltmain.sh version... no Jun 25 00:54:17 *** Gentoo sanity check failed! *** Jun 25 00:54:17 *** libtool.m4 and ltmain.sh have a version mismatch! *** Jun 25 00:54:17 *** (libtool.m4 = 1.5.16, ltmain.sh = 1.5.6) *** Jun 25 01:23:14 The way to check my theory will be for us to create an org.openembedded.nslu2-linux branch, and do some work on that. When we propagate that back to org.openembedded.dev, it should show up. Jun 25 01:23:29 then we would just keep the org.nslu2-linux.* branches for non-oe stuff Jun 25 01:23:48 and never propagate stuff between org.openembedded and org,nslu2-linux Jun 25 01:31:07 Who was complaining about cron not working on openslug earlier? Jun 25 01:35:14 rwhitby-away: hmmm.. That means I need to change stuff in my setup? Jun 25 01:43:01 NAiL: I will update the Master Makefile with the new branches. Jun 25 01:43:23 Just make sure everything is committed to the current branch, and no new changes until I'm done (probably in three or four hours) Jun 25 01:44:15 rwhitby-away: Ok, so I can commit one more thing? :) Jun 25 01:44:21 yep Jun 25 01:44:55 let me know here when you're done, and what the final head of org.nslu2-linux.openembedded is after you commit and sync that back to the server. Jun 25 01:45:10 done. I'll try to get some sleep now. Head is ef2de85880c4b7ec8c89465705bc3acb5ce02378 Jun 25 01:45:23 and make sure that your joe blow repo has all our latest changes in it Jun 25 01:46:10 joe blow up to date Jun 25 01:47:41 monotone: updating packages/cron/cron_3.0pl1.bb to 9c0ed90c6c5d57086ff22cbd6bfa0846316796bb Jun 25 01:47:41 monotone: updating packages/linux/openslug-kernel-2.6.11.2/defconfig to 9518cd19bea8afd803d60f8b412c04fb5452cd65 Jun 25 01:47:41 monotone: updating packages/meta/openslug-packages.bb to 7134faff44011b1b05d333bfa444131483aa468c Jun 25 01:47:41 monotone: updating packages/openvpn/openvpn_2.0.bb to ae85854f0ba5009ce57f8ffec27b697e75217455 Jun 25 01:47:48 was that all of them? Jun 25 01:48:25 uh, forgot one teeny little update to openslug-packages Jun 25 01:49:42 ok, now I'm done. And that's all the updates. head is now ~e0a64f00ee41ae48055371e3ab48070e5aff8020 Jun 25 01:49:49 except the tilde ;) Jun 25 01:52:09 Did that last commit just make a new branch (or whatever the term is)? Jun 25 01:52:33 Joe blow insists that "up to date" is ef2de85880c4b7ec8c89465705bc3acb5ce02378, not e0a64f00ee41ae48055371e3ab48070e5aff8020 Jun 25 01:52:43 monotone: branch 'org.nslu2-linux.openembedded' is currently unmerged: Jun 25 01:52:43 0e46f5262b548693d0460cf2c1b103cc455e4aae rwhitby@nslu2-linux.org 2005-06-25T06:31:11 Jun 25 01:52:43 e0a64f00ee41ae48055371e3ab48070e5aff8020 oyvind@repvik.org 2005-06-25T08:28:01 Jun 25 01:52:58 monotone: common ancestor ef2de85880c4b7ec8c89465705bc3acb5ce02378 found Jun 25 01:52:58 monotone: trying 3-way merge Jun 25 01:52:58 monotone: [merged] e939e179c1552200dcb9eb26b181ac051f11e4be Jun 25 01:53:13 monotone: successful exchange with mtn.nslu2-linux.org Jun 25 01:53:13 monotone: branch 'org.nslu2-linux.openembedded' is currently merged: Jun 25 01:53:13 e939e179c1552200dcb9eb26b181ac051f11e4be rwhitby@nslu2-linux.org 2005-06-25T06:35:15 Jun 25 01:54:32 hmm.. now both my repos are @ e0a... Jun 25 01:54:45 I'd expect them to be @ e939? Jun 25 01:55:47 try sync and heads Jun 25 01:56:19 monotone: branch 'org.nslu2-linux.openembedded' is currently merged: Jun 25 01:56:19 e0a64f00ee41ae48055371e3ab48070e5aff8020 oyvind@repvik.org 2005-06-25T08:28:01 Jun 25 01:56:41 sync, update, heads... Jun 25 01:57:14 is that synced with #oe too? Jun 25 01:57:32 what do you mean? Jun 25 01:58:13 nm Jun 25 02:09:23 NAiL: http://ewi546.ewi.utwente.nl/tmp/viewmtn/revision.psp?id=051fc6f9e28e16d274e2cd95db273eb514d8fb33 Jun 25 02:09:44 saw it Jun 25 02:12:04 I see that many of my changes hasn't made it Jun 25 02:12:42 really? Jun 25 02:13:37 no changes to cron or openslug-packages Jun 25 02:14:15 can you do a new joe blow from utwente only, and see if they are there in the checked out files? Jun 25 02:14:17 not openvpn either Jun 25 02:14:28 (don't sync that one with mtn.nslu2-linux.org) Jun 25 02:15:07 yeah, as soon as I figure out how to change Jun 25 02:16:49 syncing what I believe is right ;) Jun 25 02:20:38 hmm Jun 25 02:20:39 monotone: verifying new revisions (this may take a while) Jun 25 02:20:39 monotone: read from fd 4 (peer ewi546.ewi.utwente.nl) failed, disconnecting Jun 25 02:25:38 Ok, now it feels like I might get some sleep. I'm gonna go for it. Jun 25 02:25:55 I'll probably be back in a coupla hours Jun 25 02:30:40 ok Jun 25 02:48:17 sleepless in trondheim Jun 25 02:48:37 synced with ewi, my changes aren't there Jun 25 02:50:55 rwhitby-away: a18267d52d3d7431e7643dafe74393852f1be261 koen@openembedded.org 2005-06-24T12:43:40 Jun 25 03:07:24 NAiL: back now, with no distractions Jun 25 03:11:20 Well... ewi's repo looks like it has even fewer of my patches than nslu2's Jun 25 03:14:24 I'm no scm guru, to put it mildly, so I can't diagnose much on my own. But I'll provide any info you wish :) Jun 25 03:14:26 ok, I'm pulling a joe blow repo from ewi, and the comparing with ours Jun 25 03:15:42 uuuuh Jun 25 03:16:13 my ewi joeblow repo differs from the nslu2 repo in a lot of ways. This being one: Jun 25 03:16:16 DISTRO_NAME = "OpenSlug" Jun 25 03:16:18 -DISTRO_VERSION = "2.1-beta" Jun 25 03:16:21 +DISTRO_VERSION = "2.0-beta" Jun 25 03:17:51 Well, I think NAiL is beating everyone else on monotone/openslug commmits. Jun 25 03:18:11 I guess I should see about making that cvs thingo work. Jun 25 03:18:21 cia? Jun 25 03:18:31 s/cvs/cia/ Jun 25 03:18:49 nslu2-io.c, cron, openslug-packages, openvpn, samba... Everything looks like it did before I started committing anything Jun 25 03:18:58 its annoying because it wants to have a private repo to look at. Jun 25 03:19:06 and its called by cron. Jun 25 03:19:16 cron :-( Jun 25 03:19:26 Yeah. Jun 25 03:19:37 cron works in openslug if you pull from the right repo :P Jun 25 03:20:08 however, if I can figure out what the "post-commit" support they supposedly added is; it can be restructured to work the way the svn/bk one work Jun 25 03:20:41 dyoung: It doesn't use the "note_commit" hook in monotone? Jun 25 03:21:21 http://venge.net/monotone/docs/Hook-Reference.html#Hook-Reference Jun 25 03:25:20 http://lists.gnu.org/archive/html/monotone-commits-nodiffs/2005-05/msg00272.html Jun 25 03:25:23 the python script is in the monotone source tarball actually, in contrib Jun 25 03:25:45 that one is cron based. Jun 25 03:25:56 Still poking around for one that works better. Jun 25 03:27:19 looks like the hooks for a better one have only just been added recently Jun 25 03:27:50 Yeah. Jun 25 03:28:09 still digesting Lua too Jun 25 03:33:27 "they" decided to go with a cron method because the note_commit only works locally, and theres no easy way to propogate that setting for all developers. At least according to one mailing list post. Jun 25 03:34:21 rwhitby: I don't get this. in viewmtn, I can see the latest commit I did in org.openembedded.dev/org.nslu2-linux.openembedded. But when I sync I don't get any of them... Jun 25 03:35:03 NAiL: I have a theory ... Jun 25 03:35:10 verifying it now Jun 25 03:35:47 rwhitby: heh... looking at the timestamp at ewi joeblow, it's from yesterday Jun 25 03:36:41 NAiL: if you sync joeblow with mtn.nslu2-linux.org/org,nslu2-linux.openembedded, then my theory is that those others will magically appear. Jun 25 03:37:20 rwhitby, monotone has a irc channel Jun 25 03:37:26 irc.oftc.net #monotone Jun 25 03:40:56 I'll idle there for a while. Jun 25 03:41:22 rwhitby: mtn just disconnects me, dunno why Jun 25 03:43:12 monotone: warning: not currently serving requested collection 'org.nslu2-linux.org' Jun 25 03:43:14 monotone: fd 5 (peer 80.202.25.169:56230) processing finished, disconnecting Jun 25 03:43:33 then two after that succeeded Jun 25 03:45:35 um, anyone know how to undelete in linux ... ? Jun 25 03:46:13 I just deleted my private key *again* :-/ Jun 25 03:46:53 I never had much luck with any ext2/3 undelete tool Jun 25 03:47:43 well that sucks :-( Jun 25 03:47:51 /comfort rwhitby Jun 25 03:49:40 so now I have the public keys of anyone who commited anything in my db. But not the keys of anyone who hasnt commited anything. Jun 25 03:49:52 I dont get that behaviour... but... Jun 25 03:50:27 you get the keys for any certs you get Jun 25 03:51:05 threw my laptop, so have to use treo now ... Jun 25 03:52:30 ok, this time around I'm gonna put my mt private key on a CD or something ... Jun 25 03:52:43 hehe Jun 25 03:52:54 That's not such a bad idea Jun 25 03:53:48 mtn still hasn't finished... It disconnects me right after rebilding merkle trees or something like that Jun 25 03:54:07 my theory was not confirmed Jun 25 03:55:09 NAiL: someone else will need to look in the server log for you ... Jun 25 03:56:06 what are merkle trees anyway? Jun 25 03:56:23 lord knows Jun 25 03:56:38 merkle is norwegian dialect for "strange" Jun 25 03:56:40 or "odd" Jun 25 03:58:02 that's apt for what we are seeing ... Jun 25 03:58:10 yes ;) Jun 25 03:59:02 Ok, but sooner or later, org.nslu2-linux.openembedded is going to be org.openembedded.nslu2-linux, right? Jun 25 03:59:56 I expect so Jun 25 04:00:12 Lappy abuser. Jun 25 04:00:28 we will have to reapply all changes Jun 25 04:00:30 I backed up my keys in multiple places. Jun 25 04:00:33 I think I'm going to sit back, and not meddle in the affairs of wizards Jun 25 04:00:50 I've got all my changes, so it's no problem. Jun 25 04:01:04 dyoung: we've got him fooled too ... Jun 25 04:01:24 hee hee Jun 25 04:01:50 wizards don't rm -rf monotone one level too high .. Jun 25 04:01:55 that "merging-whole-repos" stuff is kinda beyond me Jun 25 04:01:57 hahaha :P Jun 25 04:02:28 Theres nothing interesting happening in the #monotone channel. Jun 25 04:02:46 just some talk of certifing 0.20. Jun 25 04:02:54 rwhitby-treo: Even after seven years of linux, I managed to rm -rf /etc/* Jun 25 04:03:01 the current status is http://venge.net:9000/ Jun 25 04:03:16 Fortunately, I was able to reconstruct most of it while the box was running ;) Jun 25 04:03:44 are they responsive to a question? Jun 25 04:04:18 Will it help if I beat them into submission? Jun 25 04:04:27 I didnt ask anything yet; since I'm not sure what to ask. Jun 25 04:04:35 I seriously need to sleep. Jun 25 04:04:36 ah no. Jun 25 04:04:44 and noone else has asked anything either. Jun 25 04:05:47 sleep? isnt it like noon your time? Jun 25 04:06:01 rwhitby-treo: How many changes have been done to "our" repo anyways? I've done a bunch, and jbowler's done two? Is there anything else? Jun 25 04:06:14 I dunno how to join two networks with my current IRC clients Jun 25 04:06:27 dyoung: Yes. I'm sick, so I can't fall asleep... I'm considering dringing a pint to fall asleep Jun 25 04:06:38 rwhitby-treo: which client? Jun 25 04:06:40 11 changes or something? Jun 25 04:06:48 no, that's pretty much all of them Jun 25 04:07:09 Miranda on win Jun 25 04:07:22 Never seen ;) Jun 25 04:08:05 I'll do an xchat or something in vmware ... Jun 25 04:09:08 BlueSlug reception is holding at the far corner of the house ... Jun 25 04:10:00 I'm probably not awake enough to ask intellegent questions there right now. Jun 25 04:11:12 in fact.... Jun 25 04:11:19 hehe Jun 25 04:11:23 g'nite Jun 25 04:11:32 nite. Jun 25 04:13:10 hmm.. samba is still seriously b0rked Jun 25 05:04:13 OK, the monotone database is going to have to be rebuilt from scratch. No more changes until that is complete please. Jun 25 05:37:33 Hmm - I just took a sync of #oe's repo, then dropped all @nslu2-linux.org keys from it, then did a db check. No change in the result (14 unreferenced manifests both before and after dropping the keys). This means that nothing in #oe's database depends on any @nslu2-linux.org keys. Jun 25 05:40:38 At least that means that us redoing our monotone database won't affect them if they choose to keep the existing one instead of redoing it from scratch at their official transition point. Jun 25 08:27:59 Master Makefile has been removed for repair and monotone repo restructuring. Jun 25 10:12:48 03bzhou * 10unslung/ (4 files in 2 dirs): bluez-libs bluez-utils bluez-hcidump (bluetooth userland packages) ready for cross testing Jun 25 14:30:30 hello Jun 25 14:30:47 I seem to be having some compiler problems Jun 25 14:31:08 but definitely don't know enough about gcc to do much about it Jun 25 14:31:17 is anybody here? Jun 25 14:33:06 I'm here Jun 25 14:33:16 But I don't know much about gcc Jun 25 14:34:08 btw, as soon as monotone gets back up, I'm submitting a patch for openvpn so it'll pass the selftest Jun 25 14:49:46 coolness Jun 25 14:50:04 that inline thing might be causing more problems later, though Jun 25 14:54:58 Yeah, I know :( Jun 25 14:55:26 But I agree with Yolnan.. It does sound like a gcc bug Jun 25 15:10:21 Lovely Jun 25 15:10:27 e2fsck gives me this: Jun 25 15:10:27 Pass 2: Checking directory structure Jun 25 15:10:27 Killed Jun 25 15:11:33 heh... not strange Jun 25 15:11:41 My swap wasn't activated Jun 25 15:14:00 That was a short nap ;) Jun 25 15:17:22 Is there a way to make e2fsck use less memory? I don't like the fact that it might just silently fail when I boot and a fsck is forced. Jun 25 15:21:28 And... isn't it an idea to check the root fs before mounting&sviewel? Jun 25 15:41:42 Am I the only one talking here? ;) Jun 25 15:47:51 currently....yes Jun 25 15:50:45 * NAiL feels alone Jun 25 15:51:00 Why can't I be sleepy now, when there's nobody here? ;) Jun 25 15:52:15 I've slept less than 8 hrs of the last 50 :( Jun 25 15:55:33 have some kids... you'll be sleepy all the time (except when you have coffee) Jun 25 15:56:18 heh Jun 25 15:56:21 but true. Jun 25 15:56:37 * ka6sox heads for the la-Z-Boy Jun 25 15:56:54 I should have said "have a kid" - at least for me, one was all it took Jun 25 15:57:14 ByronT: Yeah, have kinda had one ;) Jun 25 15:57:25 ByronT: If you get two, they manage to entertain themselvs after a while :P Jun 25 15:58:11 you're just trying to trick me into having another! have you been talking to my parents???? Jun 25 15:58:25 hahahaha, no :P Jun 25 15:58:58 I'm gonna have to watch you more closely now... you've got me wondering Jun 25 15:59:17 But I've found that two of them little buggers eases things a bit Jun 25 15:59:26 You should try it :P Jun 25 15:59:53 How old's your kid? Jun 25 16:02:33 almost two years old... Jun 25 16:02:41 I have some old pictures of ths bugger Jun 25 16:02:46 Ah, perfect timing for another one then ;) Jun 25 16:02:50 www.toddcomp.com/Christian.html Jun 25 16:03:47 "some"? Jun 25 16:04:13 look at the date... there's another year of photos that I haven't posted Jun 25 16:04:51 hehe Jun 25 16:06:59 thats nothing...this is what takes up 85MB of space on my server : http://www.websb.net/~vog/justin.html Jun 25 16:09:11 Volcanic Fog? Jun 25 16:09:29 hm. Jun 25 16:09:49 n6vog Jun 25 16:10:04 /whois n6vig Jun 25 16:10:07 /whois n6vog Jun 25 16:10:23 MIL Jun 25 16:10:41 slugarmy.mil ? Jun 25 16:11:32 dyoung-zzzz, I went to see all 5 of the famous volcanoes of Costa Rica and couldnt' see a one of the calderas...until on the airplane on the way out I saw all 5 from the air. Jun 25 16:11:43 Mother in Law. Jun 25 16:13:56 they were all covered in vog. Jun 25 20:02:44 morning, glc Jun 25 20:11:54 NAiL: almost ready with the new Master Makefile. Are you able to test it? Jun 25 20:12:13 rwhitby-away: Always prepared ;) Jun 25 20:12:19 It's at the usual place? :) Jun 25 20:14:02 not yet Jun 25 20:15:18 I guess I can safely remove my old joe-blow repos? Jun 25 20:15:38 yep Jun 25 20:20:24 rwhitby-away: btw, I wanna push the udev changes, but it being a "large" change I feel I should have the core devs blessing? Jun 25 20:21:00 morning NAiL Jun 25 20:21:13 "large" as in possibly affecting lots of stuff :P Jun 25 20:25:24 NAiL: does it work? Jun 25 20:26:21 rwhitby-away: Yup. mr_claus and I have been running it for a little while now Jun 25 20:26:46 NAiL, more importantly, if it breaks something; are you prepared to fix it? Jun 25 20:27:39 heh, that was my next question :-) Jun 25 20:27:48 dyoung-zzzz: To the best of my ability, yes :) Jun 25 20:28:36 It hasn't broken anything in the standard build as far as I can see. I've played around a bit Jun 25 20:29:06 The necessary devices are created as they should, rules work like they should, the daemon starts on cue etc.. Jun 25 20:29:56 I'd say go for it then. I wouldnt mind having a workable udev in there. Jun 25 20:30:08 dyoung-zzzz: here's the scoop on the new Master Makefile. Jun 25 20:30:18 It's just disabling the daemon if there's a problem anyways Jun 25 20:30:33 there are now only two branches we are concerned with: org.openembedded.nslu2-linux and org.nslu2-linux.dev Jun 25 20:30:55 I saw the addition of org.openembedded.nslu2-linux on ewi Jun 25 20:30:57 the former has our version of openembedded, and would be pushed back to org.openembedded.dev by core devs only. Jun 25 20:31:15 the latter is everything else for us - no longer separate branches for top-level, unslung and openslug. Jun 25 20:31:41 Aha, okay. Jun 25 20:34:11 the first thing the new Makefile does is replace itself with a co in-place of org.nslu2-linux.dev (which has that same Makefile at the top-level) Jun 25 20:35:41 dyoung-zzzz: new Makefile is up, please test. this is completely untested. Jun 25 20:37:08 rwhitby-away: Can I replace the old Makefile in my work repo with the new one? Since it's got my keys and stuff in that db? Jun 25 20:37:29 NAiL: no, the world has changed. Jun 25 20:37:34 heh, ok Jun 25 20:37:48 You have exported your keys and made a hard-copy of them, right? Jun 25 20:38:15 After July 1, if you loose your private key, you will not be able to reuse that cert id ever again. Jun 25 20:39:20 I haven't touched that repo, but I dunno how to export my keys Jun 25 20:39:52 monotone pubkey and monotone privkey Jun 25 20:40:07 burn them on a CD, or print them out on paper. Jun 25 20:42:25 or both :) Jun 25 20:46:00 so is the new Makefile doing something? Jun 25 20:46:21 monotone: verifying new revisions (this may take a while) Jun 25 20:47:08 I'm thinking of making the initial wget Makefile just be a stub, and only including enough stuff to check out the real one from monotone. Jun 25 20:48:04 so basically, "wget http://www.nslu2-linux.org/Makefile && make && make" ? Jun 25 20:48:13 yeah Jun 25 20:48:20 neat Jun 25 20:48:59 monotone: misuse: could not guess default signing key Jun 25 20:48:59 make: *** [monotone/nslu2-linux.db] Error 1 Jun 25 20:49:26 just doing "wget;make" Jun 25 20:49:50 which command? Jun 25 20:50:20 monotone: setting default collection to org Jun 25 20:50:30 monotone: setting default server to mtn.nslu2-linux.org Jun 25 20:50:30 monotone: setting default collection to org Jun 25 20:50:55 re-pasted, with two lines to give some context ;) Jun 25 20:50:59 ok Jun 25 20:56:05 Makefile updated Jun 25 20:57:12 ok, retrying Jun 25 21:00:23 also made "make setup" safe for multiple runs Jun 25 21:03:24 also, after the first successful run, you should be able to get the latest Makefile by "make update" Jun 25 21:03:42 (since the top-level dir is itself a monotone checkout) Jun 25 21:03:53 nice Jun 25 21:04:12 which is what allows us to just have a stub for the first wget Jun 25 21:08:16 ok, I'll be out for up to six hours. Feel free to fix problems and push. Jun 25 21:08:53 Ok, I'll try ;) Jun 25 21:11:27 you should be able to just add your private key to the database that the Makefile creates, and be able to mtn sync Jun 25 21:12:47 cool Jun 25 21:14:55 okay launching mine. Jun 25 21:18:33 I don't get how I import that key of mine in mt Jun 25 21:18:35 should be where we were before the snags started Jun 25 21:18:48 NAiL: all your changes should be in there Jun 25 21:19:25 back later ... Jun 25 21:20:26 rwhitby-away: some of them are, I see.. Jun 25 21:20:28 but not all :P Jun 25 21:22:20 There's "genkey", and "dropkey". But I can't find a way to import. Jun 25 21:22:51 read Jun 25 21:23:12 monotone -d .db read < nameofkeyfile Jun 25 21:24:07 the export should have been monotone -d .db privkey > privkeyfilename Jun 25 21:24:56 ... privkey > privkeyfilename.... Jun 25 21:27:06 yeah, works great now :) Jun 25 21:27:46 Mine is going without error so far. Jun 25 21:34:15 yeah, seems to work fine here too Jun 25 21:34:27 morning, jbowler Jun 25 21:34:40 morning NAiL Jun 25 21:34:42 I should cheat and link the downloads directory. Jun 25 21:35:24 Yes, everyone should cheat! (It's a waste of time, and net bandwidth, downloading the stuff again!) Jun 25 21:36:08 I'm just gonna move it from the old repo. I don't need that repo anymore Jun 25 21:36:29 yeah, good plan Jun 25 21:37:45 saved me ~314mb downloads and took me ~2 secs. That's about 157MB/sec. Good bandwidth :P Jun 25 21:38:22 So did someone manage to fix the corruption in the mtn.nslu2-linux.org monotone database? Jun 25 21:38:32 jbowler: new repo Jun 25 21:38:43 and new makefile Jun 25 21:38:57 Yes, I saw, but it's exporting 'org' which suggests desparation. Jun 25 21:39:35 And the openembedded db seems to be growing unreferenced manifests at an alarming rate. Jun 25 21:40:05 (It was 5, now it's at 13). Jun 25 21:47:21 yeah, I see that Jun 25 21:47:57 I suspect that much of it was caused by my commits that didn't make it through Jun 25 21:48:23 (ie, successfull push but never showed up anywhere) Jun 25 21:49:40 Maybe, but my 5 count was from a pull of the openembedded db after you had tried the commits. Jun 25 21:49:58 Still, monotone regards it as a 'minor' error (and it is easy to fix.) Jun 25 21:50:08 I got to 8 before the repo disappeared Jun 25 21:50:25 Unless it's a symptome of something else... Jun 25 21:51:18 jbowler: btw, you wouldn't happen to be running samba on openslug? Jun 25 21:52:06 After the changes we committed, I'm hitting a bug Jun 25 21:53:28 Atleast this time I'm getting the same result when I pull as an anon user as I have in my work repo Jun 25 21:54:40 I'm running a very old samba on a very old openslug, but I haven't ported over to the new one yet - I want to change to openldap at the same time. Jun 25 21:55:31 Well... you'll probably hit the same bug as I do. All files/folders show up as nameless folders. Jun 25 21:56:00 You can add files and they show up on the server, but on the client they'll show up as a nameless folder Jun 25 21:57:10 Codepage problem? Jun 25 21:58:11 well.. it's the same between slug->windows and slug->linux. And it's not complaining about *anything* in the logs Jun 25 22:00:53 I'm using the same config as before, only new binaries. The old crosscompiled version worked, the new natively compiled version worked. The new crosscompiled version does not. Jun 25 22:05:39 Where "old" was 10 or 14a before the path changes? Jun 25 22:06:27 14a worked before those changes Jun 25 22:08:17 That's what I don't get. The changes shouldn't affect that.. Jun 25 22:11:29 Gonna revert to the one in the default feed and see if that works Jun 25 22:12:18 It might be the feed version was built with gcc 3.4.3 Jun 25 22:13:13 hmm Jun 25 22:14:50 We're using gcc 3.4.4 now? Do you know when that changed? Jun 25 22:15:41 Two weeks ago? Jun 25 22:17:36 2005-06-03, or a bit after that Jun 25 22:18:59 hmm. the one in the feed acts the same way Jun 25 22:20:41 woohoo... --debuglevel=9 looks like it will give me a couple of days reading Jun 25 22:21:04 doubleclicking on a folder output what looks like 50-60k log Jun 25 22:22:30 samba in the feed was built 5 days ago, same problem Jun 25 22:29:38 mr_claus___: :) Jun 25 22:31:34 jbowler: samba seems totally unaware of the fact that it's outputting garbage. I wonder if there's a bug introduced in 3.0.14a which strikes when you crosscompile. I know fer sure it works natively compiled. Jun 25 22:32:20 So the cross version which worked was earlier than 14a? Jun 25 22:32:57 uh, yeah Jun 25 22:32:57 I added 14a quite recently. Jun 25 22:33:34 hmm.. no... I got the same problem with 3.0.10 cross a while back Jun 25 22:34:13 Ok, I guess that's good... I didn't actually test 14a - I just copied the patches from 10... Jun 25 22:34:37 on the 9th of April I had that problem Jun 25 22:34:50 And I don't think it went away until I compiled it natively Jun 25 22:35:27 The file site/armeb-linux may contain the answer - it has a lot of settings for cross compiling samba. Jun 25 22:36:15 aha Jun 25 22:37:03 Those site files scare me, I don't think anyone bothers to check them when a new revision is added. Jun 25 22:37:35 I haven't looked at them at all Jun 25 22:37:40 ;-) Jun 25 22:37:54 omg Jun 25 22:38:11 Quite. Jun 25 22:38:32 Ok, I really don't think I'm gonna fiddle with that :P Jun 25 22:38:52 What is that anyways? Jun 25 22:38:57 Well, I'd be tempted just to delete every samba_cv and see what happens. Jun 25 22:39:10 jbowler: My thoughts exactly ;) Jun 25 22:39:12 ? what - the samba_cv lines? Jun 25 22:39:35 those, amongst others... Everything in the site file, actually Jun 25 22:39:58 looks kinda like "workarounds" for configure or something Jun 25 22:39:59 All of those lines cause configure to skip the relevant test and use the contained result. Jun 25 22:40:16 wohoo... /me did a good guess :) Jun 25 22:40:40 yes - good pattern recognition! Jun 25 22:40:59 you get that when you sit and stare at logs scrolling by :P Jun 25 22:41:10 everything working re Master Makefile? Jun 25 22:41:27 Mine is still going. Seems to work; havnt tried pushing anything back yet though. Jun 25 22:41:33 but... If I compile samba natively, and that works... I can take the output from there and check with the site file what is wrong, correct? Jun 25 22:41:48 Yes! Jun 25 22:42:01 rwhitby-away: mine works, changes pushed so I'm up to date. joeblow repo is at same rev that my work repo is Jun 25 22:42:16 Ok, I'll start a native compile then :D Jun 25 22:42:48 * NAiL just figured out something about site files. Jun 25 22:42:53 <-- proud :D Jun 25 22:43:31 jbowler: did you see the new repo layout? Jun 25 22:43:33 Yes, that was quick of the mark. I stared at one problem for about a week until I worked out what was going on. Jun 25 22:44:09 rwhitby-treo: I've seen the new db (and branch), but I lost power so I don't know what happened in the channels. Jun 25 22:44:57 I haven't downloaded the new makefile yet - I've just been db check'ing everything I can (I've only found serious errors in the original nslu2-linux db). Jun 25 22:45:02 jbowler: I've got backlogs, but I'm not gonna paste 'em here :P Jun 25 22:45:27 I think I can work it out - and the Makefile will tell me the correct procedure for building a db. Jun 25 22:45:34 yep Jun 25 22:45:41 My only problem is ensuring I don't lose my local changed files. Jun 25 22:46:11 jbowler: it's got everything that you pushed previously Jun 25 22:47:03 Right, I've got local edits though. Jun 25 22:47:25 cp -r without MT should fix that Jun 25 22:48:06 jbowler: I don't need to do a compile now, do I. I only need to do a ./configure (with all the prefixes we're using in the .bb/.inc I assume) Jun 25 22:48:21 s/prefixes/"--'s"/ Jun 25 22:48:37 Yes, then the output is in one of those config files. Jun 25 22:49:39 Oh, config.log (cross) shows where the site file comes in. Jun 25 22:52:09 hmm? Jun 25 22:52:14 It's configuring now Jun 25 22:54:17 that reminds me I was going to ask kergoth something... Jun 25 22:54:30 Can't remember what :P Jun 25 23:05:28 jbowler: From the logs I read I guess samba is stuffing up on a BE/LE-thing Jun 25 23:12:03 Very likely - I think perl has the same problem (but not as a result of a site setting - something else). Jun 25 23:12:31 ok, now configure is done. Which file is most useful to compare with the sitefile? Jun 25 23:13:56 aha... Jun 25 23:14:08 grep samba_cv_ config.log > somefile Jun 25 23:15:01 yeah the perl issue is tricky Jun 25 23:15:30 it errors on loading dynamic modules Jun 25 23:18:32 I might've hit (one of) the jackpot(s) Jun 25 23:18:33 samba_cv_HAVE_NATIVE_ICONV=no Jun 25 23:18:39 samba_cv_HAVE_NATIVE_ICONV=${samba_cv_HAVE_NATIVE_ICONV=yes} Jun 25 23:20:10 That would explain it. Jun 25 23:20:55 There's many differences Jun 25 23:21:24 It looks like someone just dumped that stuff in without thinking about it. Jun 25 23:21:40 Well, we still have bitkeeper... Jun 25 23:22:28 back later Jun 25 23:22:33 for 4 more days. Jun 25 23:23:54 Ah, true, but I have the converted bk history as a montone db. This could come in useful... Jun 25 23:24:18 hey Nail, i just started a flashing "FrankenSlug" from my wife's iMac Jun 25 23:24:36 Pretty much all of the armeb site file (including the samba stuff) was there when kergoth checked it in in Sep 2004. Jun 25 23:25:20 ByronT: Update that page about how to get tftp or http up, will ya? ;) Jun 25 23:25:30 several gotchas Jun 25 23:25:36 ooh... Jun 25 23:25:39 I did it http Jun 25 23:25:45 cool Jun 25 23:26:12 the Redboot load does NOT like long filenames or any symbols other than 1 period Jun 25 23:26:27 so you have to rename the bin file to something like Unslung.bin Jun 25 23:26:35 jbowler: Now... If I remove all of samba from the site file, build samba cross, look at the config file, then we'll know what actually *needs* to be there, right? Jun 25 23:27:11 ByronT: Very nice to know. Should update that other page I can't remember the name of too with that one Jun 25 23:27:43 but the biggest gotcha is catching the Redboot window Jun 25 23:27:57 I estimate it to be somewhere around 1 in a Jun 25 23:28:04 billion that you catch it Jun 25 23:28:07 NAiL: yes, in that the differences might need fixing in the site file. Jun 25 23:28:17 only might? Jun 25 23:28:22 with the NIC in default settings.... Jun 25 23:28:27 In fact it is possible to assign to those cv variables on the configure command line. Jun 25 23:28:41 I think that's a better approach. Jun 25 23:29:06 jbowler: Ok, I'll yank out the stuff, build samba, see what differs, then we'll figure out how to fix it :D Jun 25 23:29:21 See the openldap .bb file (openldap is where I discovered about this.) Jun 25 23:29:57 Oh, and that setting might be relevant to perl too - openldap opens it's own DLLs. Jun 25 23:30:07 s/it's/its/ Jun 25 23:30:42 aha Jun 25 23:30:52 ok, I'll fix samba first, then go on with perl! :P Jun 25 23:30:59 Indeed, it would seem that I'm the only person who does it this way, apart, amusingly, from the other of nail_11.21.bb Jun 25 23:31:06 s/other/author/ Jun 25 23:31:45 if you guys are talking about samba configure differences between native and cross, you can also look at unslung/make/samba.mk Jun 25 23:31:47 I'm my own author :P Jun 25 23:32:27 eno-away: do you know about perl differences native/cross? Jun 25 23:33:31 sorry no. In the case of samba, i ended up comparing the result config.h, and try to patch configure.in or add cv_ variables Jun 25 23:34:13 Ah, ok, that will certainly be useful. Jun 25 23:35:04 took me 3 IPK_VERSIONs to get samba right Jun 25 23:35:04 I spent three days comparing perl config.caches and came up empty Jun 25 23:36:26 Yeah, I started to look at it but couldn't work out the endianness stuff, it looked like cross was compiling little endian, but couldn't be sure. Jun 25 23:36:33 Ok, there's 9 differences between cross and native in samba Jun 25 23:37:16 optware/make/samba.mk has 20 overrides Jun 25 23:38:06 bitbake might be better at get the cross-autoconf right though. Jun 25 23:39:13 I feel like ripping samba out entirely, and fixing configure on a per .bb basis. Jun 25 23:39:36 I see there's differences between 3.0.10 and 3.0.14 as well Jun 25 23:40:23 Would anyone kill me if I did? Jun 25 23:41:10 ByronT: With NIC in default setting? You've got an MDI-X nic, I assume? Jun 25 23:41:14 Ask mickeyl in #oe. It's armeb only (linux and linux-uclibc), I'm not sure wha that changes. Jun 25 23:41:58 Come to think of it, it's Mac. All of their nics are probably MDI-X Jun 25 23:48:22 I'll try a build without all the stuff in site, with a few things added in the .bb. See if I can get it working again. Jun 25 23:51:55 Ok, everything is in monotone now, right? How do I add a symlink in oe-packages? openvpn isn't there... Jun 25 23:52:09 Yes, that's the right thing to do. I can't see any reason why OE would object to the approach. Jun 25 23:52:38 me neither. But I'll build and see if it works or not first Jun 25 23:53:22 I'd like to have samba running... All my music is currently unavailable :P Jun 25 23:55:34 The primary problem is that the site files differ for each arch/os combination, and the settings may too. Jun 25 23:55:53 So it may be easier to leave things which are clearly arch/os dependent in the site file (that's what it should be used for). Jun 25 23:55:55 Yeah, I can't fix site files for other archs Jun 25 23:56:36 http://pastebin.ca/15875 <-- which ones do you think is arch/os dependent? Jun 25 23:56:41 The things I did in ntp and openldap were not arch dependent (I think). Jun 25 23:57:48 jbowler: on the topic ntp.. I'm running openntpd. 1/4th the mem usage and ~1/9th the size on disk Jun 25 23:59:02 DIRENT64 yes, MAKEDEV no, I'm not sure about the others. Jun 25 23:59:42 Any easy way to figure it out? Jun 25 23:59:51 I don't see why any of them would matter though - I suspect cross would work with just the settings it works out. **** ENDING LOGGING AT Sat Jun 25 23:59:56 2005