| 00:21.18 | *** join/#nslu2-linux k_velikov-away (~Evil_Geni@24-205-124-58.bb-cres.charterpipeline.net) |
| 00:31.20 | Christopher-bbl | anyone around? |
| 00:32.47 | jacques | yah |
| 00:34.03 | Christopher | my /dev/sdb1 always comes up read only when I reboot |
| 00:34.10 | Christopher | where would I go to fix this? |
| 00:35.04 | jacques | hmm, anything in dmesg that tells why? |
| 00:35.16 | jacques | I'm assuming it's not specified ro in /etc/fstab |
| 00:36.26 | Christopher | no |
| 00:36.37 | Christopher | it seems to be performing some sort of check on the drive |
| 00:36.50 | Christopher | it deletes several unreference inodes |
| 00:36.58 | Christopher | unreferenced inodes |
| 00:37.21 | jacques | it dies this every time it boots? |
| 00:37.27 | Christopher | it eventually remounts itself rw |
| 00:37.28 | jacques | can you umount and fsck ? |
| 00:37.30 | jacques | ah |
| 00:37.41 | Christopher | but then I have to start all the processes with logs manually |
| 00:37.51 | Christopher | because they can't write to their logs so they don't start |
| 00:38.22 | jacques | is this a recent unslung ? |
| 00:38.28 | Christopher | 3.18 |
| 00:38.51 | jacques | is sdb a hard drive or flash ? |
| 00:38.55 | Christopher | hd |
| 00:39.19 | jacques | hmm. I would really try fscking /dev/sdb1 (after umounting it) |
| 00:39.32 | Christopher | ok |
| 00:39.32 | jacques | or boot without it attached and then attach and fsck |
| 00:39.39 | jacques | since it may not umount |
| 00:40.06 | Christopher | yeah, it didn't |
| 00:40.15 | Christopher | i'll try that after dinner and see what happens |
| 00:40.23 | Christopher | bbl, thanks |
| 00:40.29 | jacques | good luck |
| 00:47.00 | *** join/#nslu2-linux k_velikov-away (~Evil_Geni@24-205-124-58.bb-cres.charterpipeline.net) |
| 00:48.07 | *** join/#nslu2-linux k_velikov-away (~Evil_Geni@24-205-124-58.bb-cres.charterpipeline.net) |
| 00:48.09 | dyoung | Christopher, its probably running the quotas when its ro. |
| 00:48.41 | dyoung | and may be getting stuffed up because it needs fscking. |
| 01:14.15 | Christopher | fscking... |
| 02:49.19 | *** join/#nslu2-linux k_velikov-away (~Evil_Geni@24-205-124-58.bb-cres.charterpipeline.net) |
| 03:27.01 | glc | rwhitby-away: Looks good! |
| 03:36.06 | *** join/#nslu2-linux ka6sox (~tking@adsl-66-159-198-116.dslextreme.com) |
| 04:10.33 | CIA-8 | 03bzhou * 10unslung/ (make/libtiff.mk sources/libtiff/control): upstream upgrade from 3.7.1 to 3.7.2; generates control file |
| 04:50.01 | *** join/#nslu2-linux k_velikov-away (~Evil_Geni@24-205-124-58.bb-cres.charterpipeline.net) |
| 06:51.20 | *** join/#nslu2-linux k_velikov-away (~Evil_Geni@24-205-124-58.bb-cres.charterpipeline.net) |
| 07:18.43 | *** join/#nslu2-linux jf-work (~none@home.bobo-rousselin.com) |
| 07:19.48 | *** join/#nslu2-linux perlguru (www-data@netblock-66-159-209-60.dslextreme.com) |
| 07:19.56 | jf-work | morning |
| 07:19.58 | perlguru | goodmorning |
| 07:25.00 | *** join/#nslu2-linux ruDelph (~delph@194.154.83.193) |
| 07:29.00 | *** join/#nslu2-linux eno-away (~eno@adsl-64-164-10-56.dsl.snfc21.pacbell.net) |
| 08:06.43 | jf-work | anyone interested in an eggdrop package ? |
| 08:14.57 | dyoung | what is eggdrop again? |
| 08:15.22 | ka6sox | irc bot |
| 08:15.30 | jf-work | "Eggdropt is the world most popular irc bot" |
| 08:15.32 | ka6sox | (just looked it up) |
| 08:15.46 | ka6sox | dyoung: you've got mail! |
| 08:21.04 | jacques | this is cool |
| 08:21.25 | dyoung | got a working gcc 4.0 crosstool? |
| 08:21.53 | jacques | nope |
| 08:21.57 | jacques | qemu-armeb -L /home/packages/oe/nslu2/bb_openslug_default/tmp/rootfs bin/busybox |
| 08:21.57 | jacques | BusyBox v1.00 (2005.04.29-21:04+0000) multi-call binary |
| 08:21.58 | jacques | ... |
| 08:23.25 | jacques | I can run the openslug busybox on my x86 machine |
| 08:23.41 | jacques | since qemu now supports armeb (in the latest version) |
| 08:25.24 | dyoung | nice. |
| 08:25.36 | dyoung | so you figure its more ready for mainstream work now? |
| 08:26.00 | perlguru | QEmuSlug ;) |
| 08:26.05 | jacques | I think we could definitely find some uses for it now |
| 08:26.18 | jacques | a couple of caveats |
| 08:26.43 | jacques | it still only supports user emulation on arm - running apps, not booting an OS |
| 08:28.03 | jacques | and it does not work for unslung - I get segfaults - I think this is related to the glibc version and the suspicion that some hard-float calls are still being generated even with a soft-float toolchain |
| 08:28.33 | jacques | since I am not running a hard-float kernel on my x86 box, that would explain those segfaults |
| 08:29.20 | jacques | things seem to work fine for openslug (more recent glibc), tho things that use /proc don't work right - not positive what the deal is there |
| 08:30.42 | jacques | heh I mispoke above - of course I am running a hard-float kernel on my x86, but I am not running a kernel fp emulation kernel |
| 08:30.57 | jacques | I am only guessing about the unslung segfaults tho |
| 08:31.49 | *** join/#nslu2-linux ruDelph (~delph@194.154.83.193) |
| 08:32.09 | jacques | I think it's same reason why we get segfaults if we try to use the libc libs that we compile (instead of the libc binaries from the distributed firmware) |
| 08:32.21 | jacques | (in unslung) |
| 08:34.45 | jacques | I was impressed by how easy qemu is to configure/compile/use |
| 08:35.40 | jacques | ./configure --target-list=armeb-user && make && cp armeb-user/qemu-armeb /usr/local/bin |
| 08:47.10 | jacques | s/mispoke/misspoke/ |
| 08:47.51 | dyoung | cool. |
| 08:48.00 | dyoung | I ought to give it a try after this other thing is finished. |
| 08:50.54 | jacques | hi ka6sox |
| 08:51.20 | ka6sox | hiya jacques |
| 08:52.09 | *** join/#nslu2-linux k_velikov-away (~Evil_Geni@24-205-124-58.bb-cres.charterpipeline.net) |
| 08:52.44 | ka6sox | that matrix of things that work with Crosstool is getting more RED as it goes along. |
| 08:52.53 | jacques | my thoughts exactly |
| 08:53.18 | jacques | ppor mipsel |
| 08:53.19 | jacques | poor |
| 08:53.28 | ka6sox | and in our case building native would be a real PITA. |
| 08:53.43 | jacques | I don't think crosstool has a single working combination for mipsel |
| 08:53.52 | ka6sox | ugh. |
| 08:54.01 | ka6sox | Oleg has something that he uses for the WL stuff. |
| 08:54.24 | jacques | but then again, if you want a uclibc toolchain, I think andersee's toolchain builder has something that works |
| 08:54.41 | ka6sox | they are using uclibc for Wiley |
| 08:55.19 | jacques | oh, and I read on the crossgcc list (the main list for crosstool support) that kegel has some personal rule about not including patches in crosstool if they "add new features" as opposed to "fixing bugs" |
| 08:55.43 | jacques | and he has some strange ideas of what constitutes "adding new features" |
| 08:56.06 | jacques | Wiley ? |
| 08:56.27 | jacques | BTW, we have Lennert to thank for the armeb support in qemu |
| 08:56.45 | ka6sox | I made an off hand remark about what I call my ASUS WL-HDD and it stuck. |
| 08:57.01 | jacques | heh cool :-) |
| 08:57.31 | ka6sox | so we have Hard Wileys, Wiley Std, and Wiley Delux |
| 08:58.26 | ka6sox | lennert got a slug to do dev on? |
| 08:58.37 | jacques | I don't recall |
| 08:59.09 | ka6sox | brb...I need to ship off this design to the Board House. |
| 08:59.29 | jacques | ok |
| 08:59.32 | ka6sox | have to be out the door in 4.0hrs. |
| 08:59.49 | ka6sox | (that Is I have to be out the door in 4hrs) |
| 09:00.57 | dyoung | we sent lennert a slug. |
| 09:01.18 | jacques | I think that was wise :-) |
| 09:01.41 | dyoung | I've been meaning to write him and see if it ever arrived. theres a bug I've been meaning to look at and wanted his input. |
| 09:06.02 | rwhitby-away | last I heard was when he sent me an email saying that a package had arrived for him, and that I should pay the guy who sent it |
| 09:11.29 | jacques | hi rwhitby |
| 09:12.15 | rwhitby | hey |
| 09:38.30 | jf-work | ~seen caplink |
| 09:38.32 | jbot | caplink <~joerg@pD951747B.dip.t-dialin.net> was last seen on IRC in channel #nslu2-linux, 90d 11h 22m 8s ago, saying: 'night all'. |
| 09:38.48 | ka6sox | ~seen caplink811 |
| 09:38.49 | jbot | caplink811 <~Joerg@dsl-213-023-147-175.arcor-ip.net> was last seen on IRC in channel #nslu2-linux, 15d 16h 18m 3s ago, saying: 'k, fine idea'. |
| 09:38.55 | jf-work | thx :) |
| 09:39.05 | ka6sox | ~seen caplink-811 |
| 09:39.06 | jbot | i haven't seen 'caplink-811', ka6sox |
| 09:39.07 | jf-work | ~seen caplink-away |
| 09:39.08 | jbot | jf-work: i haven't seen 'caplink-away' |
| 09:39.21 | ka6sox | I saw him earler today. |
| 09:39.26 | jf-work | good ! |
| 09:39.54 | ka6sox | he's still around. |
| 09:43.07 | jf-work | got eggdrop compiled. The question is : where should I put it in the filesystem as it normally stays in a user home directory and doesn't obey the --prefix rules ? |
| 09:43.28 | rwhitby | what type of files does it install? |
| 09:44.06 | jf-work | binaries, conf files, dl extensions |
| 09:45.25 | rwhitby | does it have an install target, and what does that target do? |
| 09:47.09 | jf-work | defaultly, the target install is $HOME/user. This can be changed with whatever we want, but there is no bin, sbin, etc, etc... |
| 09:48.37 | rwhitby | might have to just install things manually |
| 09:51.36 | *** join/#nslu2-linux ka6sox (~ka6sox@adsl-66-159-198-116.dslextreme.com) |
| 09:58.22 | NAiL | That's odd... In gentoo eggdrop installs in /usr/something. I wonder if there's a patch |
| 09:59.51 | NAiL | hrrm... /opt/eggdrop actually |
| 10:01.15 | NAiL | <PROTECTED> |
| 10:01.38 | jf-work | NAiL : I can have eggdropt install in /opt/eggdrop |
| 10:02.01 | jf-work | I was wondering if some would prefer to have /opt/sbin, /opt/bin and /opt/etc |
| 10:03.08 | NAiL | That'll be messy with the config-file... You'd have to change the path to motd, help-files and everything ;) |
| 10:03.43 | jf-work | 'coz there's a lot of bunch under eggdrop/ : doc, filesys, logs, modules, scripts, filesys, language... |
| 10:04.40 | jf-work | anyway, eggdrop package will require post install work for the user as eggdrop cannot be run as root. End user will have to create a user |
| 10:05.05 | NAiL | Yeah |
| 10:05.22 | NAiL | There's an install-script that comes with gentoo that does that... Might be interesting? |
| 10:06.54 | jf-work | NAiL: any idea where I can find it ? |
| 10:07.40 | NAiL | http://www.repvik.info/eggdrop-installer |
| 10:08.26 | jf-work | NAiL: ty |
| 10:09.29 | NAiL | np |
| 10:11.01 | jf-work | NAiL : this script seems to install eggdrop like the original 'make install' does it ("Install eggdrop for a specific user, creating the directories and files |
| 10:11.01 | jf-work | needed for eggdrop to run securely and safely.") |
| 10:11.34 | NAiL | hmm |
| 10:12.04 | NAiL | Doesn't install the binaries in the homedir tho |
| 12:20.19 | *** join/#nslu2-linux jf-work_ (~none@home.bobo-rousselin.com) |
| 13:19.57 | *** join/#nslu2-linux ByronT-Away (byrontodd@adsl-220-208-160.bhm.bellsouth.net) |
| 13:47.51 | *** join/#nslu2-linux caplink811 (~Joerg@dsl-082-082-080-125.arcor-ip.net) |
| 14:04.41 | jf-work | caplink811-away: ping |
| 14:23.54 | jf-work | caplink811-away: you've got mail too.. |
| 14:29.47 | *** join/#nslu2-linux pTweety (~pTweety@dialin-212-144-058-151.arcor-ip.net) |
| 14:39.11 | *** join/#nslu2-linux _chris_ (~chris@pcp04638208pcs.gambrl01.md.comcast.net) |
| 14:52.50 | *** join/#nslu2-linux dyoung (~dyoung@dyoung.nslu2-linux) |
| 14:52.57 | VoodooZ_Work | wb |
| 15:12.15 | CIA-8 | 03ptweety * 10unslung/make/ (11 files): some older perl-*.mk are now generating control files. |
| 15:46.49 | *** join/#nslu2-linux perlguru (~perlguru@cc444495-a.hnglo1.ov.home.nl) |
| 15:46.59 | perlguru | hi |
| 16:04.43 | CIA-8 | 03caplink811 * 10unslung/make/tcl.mk: tcl compiles now cross and native, changes by jeanfabrice |
| 16:08.22 | *** join/#nslu2-linux zoobab (zoobab@genba.ffii.org) |
| 16:27.18 | *** join/#nslu2-linux jf-away (~jeanfabri@home.bobo-rousselin.com) |
| 16:44.53 | jf-away | ² |
| 16:45.58 | perlguru | <PROTECTED> |
| 16:47.01 | schula | ³ |
| 16:52.11 | *** join/#nslu2-linux Ksilebo (russ@d14-69-12-117.try.wideopenwest.com) |
| 17:03.53 | *** join/#nslu2-linux [g2] (~g2@g2.nslu2-linux) |
| 17:06.36 | VoodooZ_Work | The Man is in the house! |
| 17:07.16 | [g2] | VoodooZ_Work, hey --- Who'd I miss ? |
| 17:07.50 | VoodooZ_Work | you of course! |
| 17:07.51 | VoodooZ_Work | <PROTECTED> |
| 17:08.00 | VoodooZ_Work | You're the MAN! |
| 17:08.13 | VoodooZ_Work | what's new? |
| 17:09.05 | [g2] | VoodooZ_Work, you don't do any fpga stuff do you ? |
| 17:09.23 | VoodooZ_Work | no, not yet |
| 17:09.42 | VoodooZ_Work | but It would be useful for some highly specialised vision stuff. |
| 17:10.55 | perlguru | article about PWC: http://kerneltrap.org/node/5066 |
| 17:11.13 | VoodooZ_Work | cool |
| 17:15.34 | VoodooZ_Work | So it "might" end up back in the kernel after all. we'll see. |
| 17:15.57 | VoodooZ_Work | perlguru: Do you have a webcam? if so, which one? |
| 17:17.21 | perlguru | yes, a logitech one (from my legomindstorm vision set) |
| 17:19.49 | VoodooZ_Work | cool. Which chipset does it uses? And does it work unde rlinux? |
| 17:25.17 | perlguru | dont know. i dont use it that often. and if i do, i use it with M$N. |
| 17:29.01 | VoodooZ_Work | ic |
| 17:37.15 | *** join/#nslu2-linux [cc]smart (~smart@gw.ptr-62-65-149-158.customer.ch.netstream.com) |
| 17:41.32 | *** join/#nslu2-linux jacques (~jacques@jacques.nslu2-linux) |
| 17:49.04 | CIA-8 | 03caplink811 * 10unslung/sources/tcl/control: makefile now produce the control |
| 17:58.57 | *** join/#nslu2-linux beewoolie-akf (~beewoolie@florence.buici.com) |
| 18:03.40 | *** join/#nslu2-linux copperbeech (~copperbee@82-40-186-72.cable.ubr04.newm.blueyonder.co.uk) |
| 18:06.49 | *** join/#nslu2-linux jacques (~jacques@jacques.nslu2-linux) |
| 18:14.11 | jf-away | caplink811-away: ping |
| 18:14.48 | jf-away | caplink811-away: can we discuss about tcl in one or two hours from now ? |
| 18:16.01 | jf-away | caplink811-away: it is now 08:15pm my time, I will be bk around 1 hour |
| 18:35.14 | [g2] | VoodooZ_Work, can you peek at something for a minute ? |
| 18:44.28 | VoodooZ_Work | yep. what? |
| 18:47.23 | VoodooZ_Work | [g2]: Are you there Master? :) |
| 19:02.13 | copperbeech | anyone ever see a build error which looks like "undefined reference to `operator new(unsigned)'" ? |
| 19:02.55 | copperbeech | ... or at least have any ideas what it might be about? |
| 19:03.07 | jp30-work | missing libstdc++? |
| 19:03.49 | copperbeech | ahh - that's an ipkg isn't it? |
| 19:04.25 | jp30-work | yes; what context are you seeing the error in? |
| 19:04.59 | copperbeech | already got that installed... |
| 19:05.30 | copperbeech | I'm trying to install the new version of AppWeb (natively - for my sins) |
| 19:05.45 | copperbeech | hmmm install = build and install... |
| 19:06.23 | jp30-work | in the unslung build env? (or natively natively? :) |
| 19:06.41 | copperbeech | in the unslung build env |
| 19:08.13 | jp30-work | have you looked at the gcc commandline that's producing the error and looked to see whether there's a -lstdc++ (also whether it is gcc or g++ that is being invoked) |
| 19:09.54 | jp30-work | ...also check that libstdc++ is staged |
| 19:11.44 | copperbeech | gcc both times... I have to go off for a few mins - be right back |
| 19:13.21 | *** join/#nslu2-linux ka6sox (~ka6sox@adsl-66-159-198-116.dslextreme.com) |
| 19:13.24 | jp30-work | see if that command line works with an explicit -lstdc++. i take it that appWeb /does/ actually contain some c++ code? |
| 19:16.07 | caplink811-away | jf-away: jf-work: ping, are you around? |
| 19:21.46 | jf-away | caplink811-away: yep, have to leave 30mn while kid go to bed |
| 19:22.11 | jf-away | will you be there in 30mn ? |
| 19:22.39 | caplink811-away | jf-away: yeah, i think so, np |
| 19:22.45 | jf-away | ok, see u |
| 19:23.02 | jp30-work | caplink811-away, i'm having some trouble building tcl |
| 19:23.19 | caplink811-away | jp30-work: native, or cross? |
| 19:24.09 | jp30-work | well, actually on a cross machine, but the trouble is with the tcl-source rule |
| 19:24.38 | caplink811 | jp30-work: describe it pls |
| 19:25.57 | jp30-work | well, for some reason make tcl-source always produces make: Nothing to be done for `tcl-source', even though there is no file match downloads/tcl* |
| 19:26.22 | jp30-work | seems like it should be an obvious typo somewhere, but i can't see one |
| 19:27.07 | jp30-work | (and of course make tcl-unpack does not succeed, because of the missing sources) |
| 19:29.27 | jf-away | jp30-work: the lack of dash (-) between the filename and the version fooled me |
| 19:29.45 | jf-away | did you check this at it appears more than one in the mk |
| 19:30.07 | caplink811 | just one moment, will have a look |
| 19:30.22 | jp30-work | jf-away: ls: downloads/tcl*: No such file or directory |
| 19:30.59 | jf-away | will check this is 20m, have it compiled cross fine |
| 19:31.04 | jf-away | s/is/in |
| 19:32.04 | caplink811 | compiles native also fine, cause the source was already in downloads :-( |
| 19:33.38 | caplink811 | k, there are blanks, insteed of a tab in the makefile |
| 19:33.45 | jp30-work | ah! |
| 19:33.54 | caplink811 | dpwnloads now |
| 19:34.14 | jp30-work | can you checkin the fixed version of tcl.mk? |
| 19:34.18 | caplink811 | yep |
| 19:35.32 | caplink811 | cia have a lag this evening, commit is done |
| 19:36.24 | jp30-work | looks like it's working. thanks! i'll let you know if there is any build trouble |
| 19:37.14 | jp30-work | there is... |
| 19:37.17 | jp30-work | cat /share/hdd/data/home/unslung/packages/sources/tcl/configure.in.patch /share/hdd/data/home/unslung/packages/sources/tcl/strstr.c.patch | patch -d /share/hdd/data/home/unslung/packages/builds/tcl8.4.9 -p1 |
| 19:37.17 | jp30-work | patching file unix/configure.in |
| 19:37.17 | jp30-work | patch: **** malformed patch at line 12: @@ -32,7 +32,9 @@ |
| 19:39.28 | caplink811 | yeah, i know, this is why i tried to reached jf-away , something went wrong at transfer |
| 19:39.44 | jean-fabrice | 'back |
| 19:39.47 | caplink811 | hi jean-fabrice , have you read your email? |
| 19:39.52 | jean-fabrice | yeah |
| 19:40.14 | jean-fabrice | what did you commit ? all the patches ? |
| 19:40.22 | caplink811 | so, it would nice, if you can commit your configure.in.patch, and yes |
| 19:41.01 | caplink811 | strstr.c.patch is ok, but i think, configure.in,patch is malformatted |
| 19:41.28 | caplink811 | maybe something like blank versus tab, or cr/lf vs. cr |
| 19:45.33 | jean-fabrice | have commit configure.in.patch |
| 19:45.40 | caplink811 | thx |
| 19:46.33 | Christopher | jacques: the fsck seemed to help, thanks\ |
| 19:46.43 | jacques | Christopher, glad to hear it! |
| 19:46.58 | jean-fabrice | even strstr.c.patch have diff |
| 19:48.00 | jp30-work | jean-fabrice, we should try updating php to 5.0.4 sometime |
| 19:48.35 | jean-fabrice | jp30-work: okay. everyday's a challenge with you ;) |
| 19:48.47 | *** join/#nslu2-linux Power_Lapi (~Frank@dsl-084-060-097-208.arcor-ip.net) |
| 19:49.14 | CIA-8 | 03jeanfabrice * 10unslung/sources/tcl/strstr.c.patch: recommit strstr.c.patch |
| 19:49.34 | jp30-work | maybe not today, i have stuff to do this afternoon, but something to think about |
| 19:50.26 | caplink811 | jean-fabrice: fine, patching works fine now |
| 19:50.40 | jp30-work | ...probably also a good idea to wait for user feedback on the threadsafe build first too. |
| 19:58.17 | jean-fabrice | does CIA-8 have troubles ? |
| 19:58.59 | caplink811 | yeah, seems to have a lag this evening |
| 19:59.40 | jean-fabrice | have fix tcl.mk to have staging produce the libtcl.so |
| 20:05.55 | *** part/#nslu2-linux jeanfabrice (~jeanfabri@jeanfabrice.nslu2-linux) |
| 20:06.01 | *** join/#nslu2-linux jeanfabrice (~jeanfabri@jeanfabrice.nslu2-linux) |
| 20:10.58 | jeanfabrice | It would be nice to have a common named file on both wiley & slug so we can identify the box during postinst for conditional, no ? or maybe we can use 'uname -m'... |
| 20:13.23 | jp30-work | jeanfabrice, i think we should uname if it can be guaranteed to be installed |
| 20:13.44 | CIA-8 | 03jeanfabrice * 10unslung/Makefile: Promote tcl for Wiley |
| 20:14.26 | jeanfabrice | humm |
| 20:15.00 | jp30-work | inspecting /proc/cpuinfo also a possibility |
| 20:15.25 | jeanfabrice | not guarantee for uname on wiley |
| 20:15.46 | jp30-work | i was just looking on a slug, it's not part of the firmware there either |
| 20:16.23 | jeanfabrice | wiley have a .version at the top but there's only the fw release in it |
| 20:16.33 | jeanfabrice | don't you have such ? |
| 20:17.43 | jp30-work | maybe that's the solution: unslung slugs have a .unslung in the root directory of the root fs |
| 20:17.59 | jp30-work | it contains a version like this: V2.3R25-uNSLUng-standard-3.17-beta |
| 20:19.55 | ByronT | I thought that was deprecated... |
| 20:20.04 | ByronT | I could be wrong |
| 20:20.15 | jp30-work | hmm |
| 20:21.02 | jp30-work | jeanfabrice, you'd better make sure to talk to the unslung firmware guys to check on this. |
| 20:21.43 | jp30-work | when i said that uname would be best, i meant that it would be best if we didn't have to change existing firmware to implement this |
| 20:22.08 | jeanfabrice | yep. I was just wondering how usefull it could be for conditionnal action on package postinst and see if you could find it usefull |
| 20:22.21 | jp30-work | well, i haven't needed it yet |
| 20:22.59 | jeanfabrice | as an example, creating user account on slug and wiley implies differents action |
| 20:23.32 | jp30-work | come to think of it, it's always possible to generate that information at compile time and write it into the script |
| 20:23.47 | jeanfabrice | for wiley, as /etc/passwd is stored in ramfs, we need to make some trick to make the change permanent |
| 20:24.11 | jp30-work | its very arcane and tricky to change /etc/passwd on the slug too |
| 20:24.26 | jeanfabrice | wouldn't it denature the .mk to have it create an account ? |
| 20:24.50 | jeanfabrice | denature = change its initial and original goal |
| 20:24.57 | *** join/#nslu2-linux pTweety (~pTweety@dialin-212-144-058-151.arcor-ip.net) |
| 20:25.00 | jp30-work | two things... |
| 20:26.08 | jp30-work | 1) i mean if a postinst script needed to know whether it was on a slug or wiley, the .mk could modify it with sed so it knew (e.g. sed -e s/@MACHINE@/$(UNSLUNG_TARGET)/) |
| 20:26.31 | jp30-work | 2) my view is that unslung packages should just refrain from modifying /etc/passwd |
| 20:26.38 | jeanfabrice | okay |
| 20:26.45 | jp30-work | this came up with apache and mysql |
| 20:26.57 | jp30-work | ...which is why apache runs as nobody, and mysql as root |
| 20:27.14 | jeanfabrice | does nobody has a shell on slug ? |
| 20:27.20 | jeanfabrice | s/has/have |
| 20:27.37 | eno-away | currently postgresql also modifies /etc/passwd |
| 20:27.39 | jp30-work | i hope not, let's check |
| 20:27.54 | jeanfabrice | and what id/groupid is nobody ? |
| 20:28.04 | senneth | nobody:x:99:99:Nobody:/: |
| 20:28.11 | eno-away | uid=99(nobody) gid=99(nobody) |
| 20:28.18 | jeanfabrice | ok, got the same |
| 20:28.48 | *** join/#nslu2-linux rwhitby-away (~rwhitby@rwhitby.nslu2-linux) |
| 20:29.35 | jeanfabrice | without shell ? wiley has /bin/false |
| 20:29.40 | eno-away | wish we had useradd/groupadd on unslung |
| 20:30.40 | jp30-work | eno, it looks like you are just modifying /etc/passwd on the slug - i don't know that that will survive a reboot (even though it is in jffs, the linksys boot scripts like to overwrite it unexpectedly) |
| 20:30.58 | jeanfabrice | still as an example, I working on eggdrop which refuse to run as root. And nobody has no shell.... need an eggdrop account, no ? Should I let the end user create it himself ? |
| 20:32.00 | eno-away | there's a "cp /etc/passwd /share/hdd/conf/" step in postinst |
| 20:32.30 | jp30-work | eno, ok, i see it now |
| 20:32.34 | eno-away | don't know if that's sufficient or not |
| 20:32.45 | jeanfabrice | but that do not work on wile ;) |
| 20:32.50 | jeanfabrice | wiley |
| 20:33.05 | jeanfabrice | sorry :( |
| 20:33.07 | eno-away | i understand, but that predates wiley |
| 20:33.08 | jp30-work | that's another good reason just to run postgresql as root |
| 20:33.59 | jeanfabrice | eno-away : j/k |
| 20:34.49 | eno-away | i understand the reason here, but OTOH, i'd like to see a cross platform useradd/groupadd working |
| 20:35.05 | jp30-work | jf, why does eggdrop need to have a shell in /etc/passwd? does the user have to become that user manually to start it? |
| 20:35.48 | eno-away | i gotta go, will read irc log later |
| 20:35.52 | jp30-work | (if it owns files, that's a reason not to use nobody - nobody should never be able to write to anything) |
| 20:36.43 | jeanfabrice | jp30-work: it appears that eggdrop has to be launch with a unprivileged account |
| 20:36.51 | pTweety | jp30-work: I've upgraded to the latest apache version in the feed an now have 30 threads running. What is the right httpd.conf option to reduce it to about 10? |
| 20:37.00 | *** join/#nslu2-linux Ksilebo (~russ@d14-69-185-70.try.wideopenwest.com) |
| 20:37.23 | pTweety | There are such a lot of different numbers in this file :-( |
| 20:38.09 | jp30-work | pTweety, you need to tune ThreadsPerChild and MaxClients. if MaxClients > ThreadsPerChild, apache will start multiple processes, which you probably don't want |
| 20:39.06 | jp30-work | jeanfabrice, can you patch it to setuid(nobody) if it detects that it's running as root |
| 20:41.26 | jeanfabrice | jp30-work: that could be a solution yes |
| 20:43.40 | jp30-work | pTweety, pls let me know if you think the number of threads needs to be tuned for apache. I figured that threads are pretty low cost. also, apache can't deal with more clients than threads (especially an issue when keepalive is being used). |
| 20:44.48 | jp30-work | it may yet be a faq to disable keepalive if you expect to serve many clients concurrently |
| 20:45.35 | jp30-work | jf, i have to go, keep me posted re: eggdrop |
| 20:46.27 | pTweety | jp30-work, I don't much about apache. But for my purpose 10 threads are enough. That was why I asked. |
| 20:49.08 | pTweety | Anyway, I'm going to do some test on apache and php the next days and will report any strang observations back to you |
| 20:50.27 | pTweety | night all |
| 20:57.53 | *** join/#nslu2-linux Ksilebo^ (~russ@d14-69-185-70.try.wideopenwest.com) |
| 21:20.17 | jeanfabrice | should log and pidfile go under /opt/var ? |
| 21:24.02 | jeanfabrice | (wiley have /var mount on ramfs) |
| 21:25.46 | *** join/#nslu2-linux dyoung (~dyoung@dyoung.nslu2-linux) |
| 21:44.38 | *** join/#nslu2-linux dr_nick (~markus@p5090D4FD.dip.t-dialin.net) |
| 22:11.31 | CIA-8 | 03jeanfabrice * 10unslung/make/php-thttpd.mk: Fixed libxml2 search path |
| 22:19.35 | *** join/#nslu2-linux jaro (~ja4aro@195.47.42.18) |
| 22:21.44 | jaro | hi jeanfabrice, can you please tell me from what sources did you manage to build thttpd-php package? |
| 22:25.14 | jeanfabrice | jaro: I'm currently using source form thttpd package and php package |
| 22:25.23 | jeanfabrice | s/form/from |
| 22:27.51 | jaro | Thanks a lot. Anyway - did you try to build that combination of packages using uclibc instead of full libc? I'd like to put it into the flash, but libc won't fit there. |
| 22:29.10 | jeanfabrice | I currently build them using a uclibc toolchain |
| 22:29.37 | jeanfabrice | so then build ;) |
| 22:29.43 | jeanfabrice | s/then/they |
| 22:30.35 | jeanfabrice | (asus wl500g only support uclibc for now, and I'm building package for such box) |
| 22:35.06 | CIA-8 | 03jeanfabrice * 10unslung/Makefile: Demote quagga for wiley as it needs fixes |