[02:53] <[mbm]> hmm
[03:06] <[mbm]> nbd: we could do the tracing by overriding the default make rules
[03:06] <nbd> how?
[03:06] <[mbm]> well, I mean we could write a .o.c: rule that does the "CC foo.c" thing
[03:07] <nbd> you mean overriding the rules of the individual packages?
[03:08] <[mbm]> well, most packages don't define those rules, they just depend on make defaults
[03:08] <nbd> i think we should get the build system tracing right first
[03:09] <nbd> with different verbosity levels
[03:09] <nbd> then we can see about what we do to the package builds
[03:09] <tziOm> ipkg.c:(.text+0xc): undefined reference to `ipkg_op'
[03:09] <nbd> tziOm: did you change any busybox settings?
[03:09] <tziOm> alot
[03:10] <[mbm]> and you obviously broke it
[03:10] <tziOm> didnt change the cod
[03:10] <tziOm> e
[03:10] <nbd> but you changed the settings
[03:10] <nbd> that can be enough to screw things up
[03:10] <tziOm> seems so
[03:12] <tziOm> I set "compile all sources at once"
[03:12] Action: [mbm] wonders why we're expected to fix something tziOm broke
[03:12] <tziOm> and MD5 => 0 .. 1k bigger and 3 times faster..
[03:13] <tziOm> but I dont think im doing anything _wrong_ .. maby stuff you havent tried?
[03:13] <[mbm]> < nbd> but you changed the settings
[03:13] <tziOm> mbm - whats wrong with You?
[03:13] <[mbm]> no, what's wrong with you
[03:14] <[mbm]> you complain that all the settings are wrong, then you change them all, and then you complain that it doesn't work
[03:14] <tziOm> is it wrong to go through configuration with a careful eye? ..reading and optimizing?
[03:15] <tziOm> i dont just change to change, not born yesterday
[03:15] Action: nbd wonders if tziOm is using gentoo linux
[03:15] <tziOm> slack
[03:15] <nbd> ah
[03:15] <tziOm> since -94
[03:16] <nbd> anyway, changing lots of options and pasting single lines from the logs doesn't give us _anything_ that we can use
[03:16] <tziOm> the log did say something
[03:17] <tziOm> I cant know what you understand and not
[03:17] <[mbm]> we have no idea what you changed, and you're not giving us much information beyond "it broke."
[03:17] <nbd> what you pasted doesn't tell us anything other than 'it's broken'
[03:18] <nbd> if you can't filter the information for us, then you'll have to give us the information in a way that allows us to look for what's relevant
[03:18] <[mbm]> start with the defaults, make sure you can compile that way, then change the settings individually, so that when something breaks you know why it broke
[03:18] <nbd> right
[03:18] <tziOm> think it started when i disabled ipkg support.
[03:18] <tziOm> dont need it
[03:20] <tziOm> now it works
[03:20] <tziOm> and ipkg is disabled.. dunno what happened
[03:20] <[mbm]> :P
[03:21] <tziOm> why elinks and not w3m, which is better? size?
[03:22] <[mbm]> try compiling w3m and see if it's smaller
[03:23] <tziOm> elinks wont compile.. should add some menuconfig configurable options to it.. like --without-openssl (if it works)
[03:25] <tziOm> hmm.. it says --with-openssl=(bah)/staging_dir_mipsel/usr
[03:31] <CIA-16> nbd * r3901 / (2 files in 2 dirs): don't compile libipkg in busybox if ipkg is disabled
[03:31] <nbd> :)
[03:32] <[mbm]> hmm .. could export a MAKEFILES from the top level
[03:32] <tziOm> so.. debugging is bit helpful afterall?
[03:32] <[mbm]> and basically override the default rules globally
[03:32] <nbd> tziOm: sure it is, if it's done right
[03:33] <nbd> [mbm]: i wouldn't do it by default for every package. can introduce lots of potentially weird cases
[03:33] <nbd> [mbm]: i don't trust make to handle things like that properly
[03:33] <[mbm]> nbd: true .. just thinking about things
[03:37] <tziOm> openssl is compiled after elinks and elinks needs openssl...
[03:37] <nbd> i think we should have output similar to what the linux kernel prints for the build system trace
[03:37] <nbd> i.e. when it enters the package/foobar directory it should print CHK package/foobar
[03:37] <[mbm]> the packages/Makefile is used to setup up package dependancies
[03:38] <nbd> then some short strings for download, prepare, configure, compile and install
[03:38] <[mbm]> nbd: sounds good
[03:39] <[mbm]> could do that from the packages/rules.mk
[03:39] <nbd> yeah
[03:39] <tziOm> why the +s on /usr btw?
[03:39] <[mbm]> shouldn't be
[03:39] <tziOm> it is
[03:39] <[mbm]> isn't here
[03:39] <nbd> not on my router either
[03:40] <tziOm> is here.. current cvs, and not changed a thing in perms
[03:40] <nbd> you pulled it from cvs?
[03:40] <tziOm> svn even
[03:40] <[mbm]> we use svn
[03:40] <nbd> ah
[03:40] <nbd> [mbm]: is the cvs still active?
[03:40] <[mbm]> and it's probably a umask thing
[03:41] <[mbm]> nbd: hmm stragenly yes
[03:41] <tziOm> why is it installing shit i dont even want to hear about in this kind of system.. mtsql
[03:41] <tziOm> mysql...
[03:42] <[mbm]> because you told it to?
[03:42] <tziOm> if it is a dependancy.. I would very much like the log to tell me what required it.. because i dont
[03:43] <nbd> [mbm]: let's talk about the V= levels
[03:43] <nbd> [mbm]: how should we define them?
[03:44] <[mbm]> nbd: think they should define depth
[03:45] <tziOm> really!?!!
[03:45] <[mbm]> ?
[03:46] <tziOm> ...now its compiling postgresql also...
[03:46] <[mbm]> ...
[03:51] <tziOm> Your config system is so screwed!
[03:52] <[mbm]> there's a reason why we're writing buildroot-ng :P
[03:52] <tziOm> it is compiling everything I havent enabled I think..
[03:52] <[mbm]> no, you probably enabled it without knowing
[03:52] <tziOm> NO!
[03:52] <nbd> [mbm]: http://pastebin.com/756879
[03:52] <nbd> tziOm: did you enable the sdk?
[03:53] <tziOm> and not a single thing that could have use for libpng either
[03:53] <[mbm]> either that or he did DEVELOPER=1 build
[03:53] <[mbm]> nbd: yeah, something like that
[03:54] <tziOm> yeah, the sdk
[03:54] <[mbm]> the sdk builds all possible libraries
[03:54] <tziOm> yeah?
[03:54] <tziOm> shitty sdk
[03:55] <[mbm]> look, you mind not being so much of an asshole?
[03:55] <nbd> [mbm]: or maybe better: http://pastebin.com/756886
[03:55] <tziOm> sorry
[03:55] <tziOm> but im glad other sdks dont do that
[03:55] <[mbm]> the sdk is used to provide a full environment for building anything, so to build the sdk it has to build all the libraries that people might use
[03:56] <[mbm]> you don't need the sdk because you're building from buildroot
[03:56] <tziOm> hmm.. couldnt ppl choose that, maby?
[03:56] <[mbm]> they are given the choice, you just made the wrong choice
[03:56] <nbd> [mbm]: so how do we implement that kind of output in a clean way?
[03:57] <tziOm> choose what libs to compile
[03:57] <[mbm]> tziOm: turn off the sdk and choose
[03:57] <tziOm> thats ok, but hey
[03:57] <[mbm]> nbd: liked the other one better
[03:57] <nbd> ok
[03:58] Action: nbd can't think of a decent way to implement this without making it look too similar to the old mess
[03:58] <nbd> [mbm]: btw. remember that stupid LED spammer that put is crap in our trac?
[03:58] <nbd> s/is/his/
[03:59] <[mbm]> yeah, the ticket spam
[03:59] <nbd> he's now bugging madwifi
[03:59] Action: [mbm] hates spammers
[04:02] Action: nbd just sent otaku42 a link about that spammer
[04:02] <nbd> (the blog post with the right ip to firewall)
[04:03] <tziOm> tcpdump is huge
[04:03] <[mbm]> yep.
[04:03] <tziOm> It shouldnt install the /usr/lib/ipkg if ipkg is disabled, as is here
[04:04] <nbd> our system expects ipkg to be ther
[04:04] <[mbm]> we didn't expect anyone to be stupid enough to disable ipkg?
[04:04] <tziOm> what the fuck are you saying?
[04:04] Action: [mbm] repeats
[04:04] <[mbm]> we didn't expect anyone to be stupid enough to disable ipkg?
[04:05] <tziOm> Stupid asslicker
[04:05] <[mbm]> do you not understand english?
[04:09] <tziOm> is there no fancy way of compiling test c-code for arch, then running some fancy parser script through code to estimate quite well size of bin?
[04:10] <nbd> [mbm]: so... if $TOPDIR/Makefile calls $TOPDIR/package/Makefile, what do you think would be the right place to add the hook for the V= stuff?
[04:29] <[mbm]> :)
[04:30] <[mbm]> some fancy parser to estimate the size of a binary .. hmm a compiler?
[04:35] <nbd> :)
[05:00] <common> nbd?
[05:01] <common> thought about asking the openixp guys if they want to merge their code to openwrt?
[05:04] <nbd> iirc we were already in contact with the guy runing the project. ask kaloz, he's the one doing the ixp4xx port
[05:06] <common> i shot a wrv on ebay after i saw one can solder a 2nd minipci slot on it
[05:08] <common> i bet the soldering will be 'fun'
[05:11] <common> hrm
[05:11] <common> last changes on xscale are 2 month old
[05:11] <common> in trunk/
[05:14] <common> Kaloz SYN
[06:27] <[mbm]> .
[07:21] <[mbm]> nbd: do we really need BUILD_MAKEFLAGS passed on the make commandline?
[10:39] <nbd> [mbm]: i did this to keep it from passing on -j
[11:17] <tziOm> any particular reason nvram is located in /usr ? .. shouldnt it be in / ?
[12:18] <Kaloz> common: syn/ack
[15:23] <CIA-16> nico * r3902 /trunk/openwrt/package/ (14 files in 5 dirs): add rp-l2tp package (closes: #459).
[15:25] <nbd> {Nico}: ping
[15:26] <{Nico}> nbd: pong
[15:26] <nbd> {Nico}: i think it's about time we freeze the packages in trunk and port everything over to buildroot-ng
[15:27] <nbd> {Nico}: what do you think?
[15:27] <{Nico}> nbd: i have some updates pending, but yeah, i think its time
[15:28] <nbd> also: did you read the thread that i started on -devel?
[15:28] <nbd> (mail)
[15:29] <{Nico}> nbd: i read it and i agree
[15:30] <nbd> can you write a response?
[15:30] <{Nico}> nbd: will do
[15:31] <nbd> thanks
[16:23] <tziOm> in 2.4 kernel config .. the "Broadcom Home Network Division" .. is it the binary bcm43xx driver
[16:29] <nbd> no
[16:32] <tziOm> ndb - ok.. its the CONFIG_WL and /linux/drivers/net/wl
[16:32] <tziOm> nbd - whats the binary part of it?
[16:34] <tziOm> Are there alternatives to tcpdump that doesnt take up that much space?
[16:35] <nbd> wl is the driver for bcm43xx, right
[16:36] <tziOm> so it has to be a module? cant be put into kernel?
[16:40] <tziOm> nbd, how can I choose what gcc version to use?
[16:40] <CIA-16> nbd * r3903 /branches/buildroot-ng/openwrt/package/rules.mk: whitespace fixes
[16:41] <nbd> tziOm: you shouldn't mess with the gcc version
[16:41] <nbd> tziOm: some targets are quite sensitive to that
[16:41] <tziOm> seems you are using 3.4.6..
[16:43] <nbd> it depends on what target you choose
[16:43] <tziOm> nbd - after building .. does the openwrt-brcm-2.4-squashfs.trx (in my case) contain kernel and filesystem and everything?
[16:43] <nbd> broadcom 2.6 uses 3.4.6
[16:43] <nbd> broadcom 2.4 uses 3.4.4
[16:43] <nbd> some targets use 4.1.0
[16:45] <tziOm> nbd, I noticed BUSYBOX_CONFIG_BUILD_AT_ONCE .. gcc 4.1 and obove..
[17:01] <tziOm> could anyone help me understand how openwrt works when it comes to: building filesystem image from build_ARCH/root, building kernel/modules, and where are these placed? is everything combined in .trx file?
[17:02] <nbd> everything is in the trx file
[17:03] <nbd> it's a container for loader, kernel image and filesystem
[17:06] <tziOm> nbd, ok.. how do i generate that file after changeing kernel and root ?
[17:06] <tziOm> and where are the kernel modules?
[17:06] <nbd> how to generate it? just run make
[17:06] <nbd> the build system will take care of it
[17:07] <tziOm> wont it change my root (which I have modified) and change kenrnel config too (which is modified running make menuconfig ARCH=mips inside build_mipsel/linux)?
[17:08] <nbd> don't change the root manually
[17:08] <tziOm> Point is I want a bit more control/customizing than the "buildroot" menuconfig gives me..
[17:08] <nbd> openwrt is built on the assumption that all files are packaged
[17:08] <tziOm> is it no make root
[17:09] <tziOm> or make kernel and make final?
[17:09] <nbd> if you make sure that each of your modified files is in a package, then it'll do the right thing
[17:09] <tziOm> package?..
[17:09] <tziOm> nbd, lets say I delete some files and dirs from build_mipsel/root and change some config files..
[17:09] <nbd> don't manually edit stuff in build_mipsel/root
[17:10] <nbd> leave that directory alone
[17:10] <nbd> it's temporary
[17:11] <tziOm> ndb - its the easiest way.. cant you just tell me howto make a squashfs image from current root? is it no make option for going directly to that step?
[17:11] <nbd> no
[17:12] <tziOm> no? why so hostile?
[17:12] <nbd> there is no make option for that
[17:12] <nbd> that's what i meant
[17:12] <tziOm> can I cd to a dir and make to make it work then? or do I have to make a script to do it for me?
[17:13] <nbd> the build system was not designed for having its temporary directories messed with
[17:13] <tziOm> in the last case, could you please tell me commands to run to 1) make squashfs filesystem 2) make kernel and modules and 3) combine those into .trx image
[17:14] <nbd> i don't remember the exact command
[17:14] <nbd> commands
[17:14] <nbd> you have to look them up yourself
[17:14] <nbd> i never build images manually
[17:15] <tziOm> nbd - what is the common way of attacking it when you want to manually change kernel options and filesystem stuff then? I can probably copy the .config somewhere after my menuconfig ARCH=mips?
[17:16] <nbd> you can edit the kernel config in target/linux/brcm-2.6/config (for the broadcom 2.6 target)
[17:17] <tziOm> nbd - that sounds like the same as using menuconfig then copying .. better..
[17:18] <tziOm> nbd, im using 2.4 .. since 2.6 seems to be having problems with its bcm43xx support in openwrt
[17:21] <tziOm> nbd, are You familiar with the bootup process?
[17:21] <nbd> sure
[17:21] <tziOm> ok, what is run and in what order?
[17:22] <nbd> just look at the init scripts
[17:22] <tziOm> what is run first then
[17:24] <tziOm> You dont know?
[17:24] <milan_h> everything starts at /etc/preinit I think
[17:24] <tziOm> kernel calls preinit? .. no!
[17:24] <tziOm> or
[17:25] <milan_h> root@OpenWrt:~# cat /proc/cmdline
[17:25] <milan_h> [...] init=/etc/preinit [...]
[17:26] <nbd> yes, the kernel calls preinit
[17:26] <tziOm> yeah, ok. thanks milan_h
[17:26] <nbd> sorry, if i don't go into detail on questions like that, but i have better things to do
[17:26] <nbd> some things you can just look up yourself
[17:26] <tziOm> nbd, bah
[17:27] <tziOm> hmm.. but how is the squashfs stuff handeled?.. this happends after squashfs image in uncompressed to ram..
[17:28] <nbd> the squashfs image is not uncompressed
[17:28] <nbd> it's being mapped to a mtd partition and mounted directly
[17:29] <tziOm> nbd, so it will be used as is from flash?
[17:29] <nbd> right
[17:30] <tziOm> ok...
[17:33] <tziOm> the mtd is only binary? where can i find any docs on it?
[17:33] <nbd> it's a kernel subsystem
[17:33] <tziOm> not the util
[17:34] <nbd> do you even try to look in the build system to figure out things like that yourself?
[17:34] <tziOm> yep
[17:34] <tziOm> but found only mtd and mtd-static in build_mipsel/mtd
[17:34] <nbd> and obvious place to look would have been package/mtd
[17:34] <nbd> build_mipsel is a temporary directory
[17:35] <tziOm> there are other sources there, but thaks.
[17:36] <tziOm> is that mtd documented anywhere?
[17:36] <nbd> it has a command line help
[17:40] <db90h> i wish i had your patience nbd
[17:42] <tziOm> hmm.. so its running preinit to figure out if it should run failsafe and telnetd (without checking existance of telnetd..) .. then if not, its running mount_root (maby add /sbin/ here to avoid possible problems) .. mount root mounts tmpfs and does some mtd unlock linux stuff.. then it checks what fs it is, and if like my case squashfs it runs includes /bin/firstboot
[17:42] <tziOm> nbd, am I right this far
[17:47] <tziOm> you require hexdump actually... hmm..
[17:49] <nbd> it doesn't always mount tmpfs
[17:49] <nbd> if the jffs2 part is valid, it mounts this instead
[17:50] <tziOm> nbd, should the . /etc/preinit.arch in preinit be preinit.${arch} after setting arch ?
[17:51] <nbd> no
[17:55] <tziOm> nbd, I dont understand exactly what your is_dirty function does.. (and be warned, you are using a bin (hexdump) that ppl easily drop in config.. maby add comment)
[17:55] <nbd> you're the first one i know that actually wanted to drop it
[17:56] <tziOm> wont need it after the bootup...
[17:56] <nbd> maybe you don't need it
[17:57] <nbd> but it can be useful sometimes
[17:57] <tziOm> who needs it?
[17:57] <tziOm> it can be.. but not on the system im making,
[17:57] <nbd> feel free to write equivalent shell code if you like
[17:57] <tziOm> but anyway.. if there are other ways of getting the info you need, maby you should consider keeping the bootup requirements down
[17:58] <nbd> since there's lots of other stuff to do, i won't bother doing something about this right now
[17:59] <tziOm> nbd, can you please explain what it is you are doing in the is_dirty?
[17:59] <nbd> tziOm: checking a flag in a byte that we reserved inside the trx
[17:59] <tziOm> nbd, for?
[18:00] <nbd> read the code
[18:00] <nbd> don't make me explain every single line
[18:00] <nbd> because i won't do it
[18:00] <tziOm> if you should switch to jffs2..
[18:02] <tziOm> nbd, another thing maby ppl are more likely to do is dropping telnetd and including sshd (dropbear) .. this will cause boot failure in preinit, its not checking existance of telnetd before executing... nor checking for dropbear..
[18:03] <nbd> please take some more time to figure out how the system works
[18:03] <db90h> lol, leave him alone trzi0m.. these guys put enough uncompensated time into openwrt w/o being harassed
[18:03] <tziOm> its advice.. i could do it.. its 2 lines of shellscripting
[18:03] <tziOm> but i dont have access to code modifications
[18:03] <nbd> tziOm: you're making wrong assumptions ... lots of them
[18:04] <tziOm> nbd, for instance?
[18:04] <nbd> that dropbear is not included by default
[18:04] <nbd> that the dropbear init script is there even if you deselect it
[18:05] <tziOm> what?
[18:05] <nbd> or that telnet is always enabled
[18:05] <tziOm> i was talking about the bootup, and spesifically the FAILSAFE mode.
[18:05] <nbd> then say that
[18:05] <nbd> anyway
[18:05] <nbd> dropbear is unsuitable for failsafe mode
[18:06] <nbd> telnetd is much simpler
[18:06] <tziOm> so is telnetd if its not there, and why unsuitable?
[18:06] Action: nbd is getting more and more annoyed
[18:06] <db90h> trizi0m: even if u make valid arguments, its easy to be a critic
[18:06] <nbd> just be silent for a while
[18:06] <db90h> ;p
[18:07] <nbd> db90h: lots of 'maybe i want this', 'maybe i want that' ...
[18:07] <nbd> maybe there are lots of other places that are way more important...
[18:07] <db90h> yep ;)
[18:07] <nbd> or 'maybe' people that write code are more useful than people that complain about code :)
[18:07] <db90h> hehehe
[18:08] <nbd> anyway, next time i won't be so patient with this guy
[18:08] <db90h> i dunno how you are so patient to begin with.. i'd lose my mind quite quickly
[18:09] <db90h> you are a much better man than I ;)
[18:09] <nbd> ;p
[18:13] Action: nbd is happy with the progress that buildroot-ng is making
[18:13] <nbd> almost done rewriting most of the code in target/linux/image...
[18:53] <common> Kaloz: ACK
[18:53] <db90h> SYN
[18:53] <common> db90h: RST
[18:54] <db90h> ;p
[19:29] <Kaloz> common: so?
[19:32] <common> whats zup with wrv support?
[19:33] <common> nbd told me you were about to merge the 2.6 diffs from openixp, but i saw little to nothing in that direction
[19:49] <nbd> i didn't say that
[19:50] <nbd> i said that kaloz was working on xscale stuff
[19:53] <nbd> let's see if he can behave now
[20:03] <tziOm> no hardware watchdog in the wlhdd?
[20:03] <nbd> no
[20:09] <florian> is there a developper meeting tonight ?
[20:10] <nbd> no idea. haven't heard anything
[20:10] <florian> ok
[20:10] <nbd> florian: btw. did you read the email thread that i started on -devel?
[20:11] <florian> yes I did, did not take time to answer yet
[20:11] <nbd> ok. i'd like to have some feedback from all the devs, that's why i'm asking
[20:11] <florian> right
[20:12] <common> is it possible to get a ro account on the ml for non dev lurkers?
[20:13] <nbd> ro account? everybody can already send, so if someone were to get read access, he'd have an rw account already :)
[20:13] <nbd> anyway, this issue is still being discussed
[20:13] <common> everybody can send to the ml?
[20:14] <nbd> yeah
[20:15] <common> just subscribed to users and devel
[20:17] <common> devel is pending, users empty
[20:58] <[mbm]> wow, is it 20,000 questions day?
[20:59] <[mbm]> biggest problem with tzi0m is that he doesn't understand anything, yet it's all wrong and needs to be changed
[21:10] <[mbm]> the worst part is that all this work that was done explaining to tzi0m will just be lost in the irc archives
[21:40] <nbd> [mbm]: btw. i think we should have an official freeze for the work on kamikaze packages
[21:40] <nbd> until everything's ported over to buildroot-ng
[21:43] <[mbm]> hrm ..
[21:43] <[mbm]> you know what happened with the last freeze
[21:43] <[mbm]> but yeah, probably should
[21:46] <nbd> well, this is not a stability freeze
[21:46] <nbd> it's a 'force people to convert stuff to the new format' freeze
[21:46] <nbd> so the result will most certainly be different :)
[21:47] <[mbm]> btw about the MAKEFLAGS do we really need to pass those via the commandline? everything includes rules.mk so we should just be able to set it there
[21:48] <florian__> [mbm]: do we make a developper meeting tonight ?
[21:48] <[mbm]> florian__: I hadn't planned one
[21:48] <florian__> ok
[21:49] <nbd> [mbm]: you can't override it from rules.mk
[21:49] <nbd> [mbm]: make doesn't allow it for some reason - i tried it
[21:51] <[mbm]> weird
[21:51] <[mbm]> just looks nasty on the commandline
[22:00] <florian__> there is a lot of new people here, welcome !
[22:00] <nbd> yeah, ever since this channel was made public
[22:47] <crazy_imp> has someone running kamikaze build (2.6 kernel) on a wl-hdd2.5? my shows strange behavior like bringing the network interface up and down in about 30 sec distance and its still responding to the nvram ip (but there are still ~80% lost) and to the default ip there are 100% lost. telnet is imposible :(. any ideas?
[22:57] <crazy_imp> or can someone say, that it's an hopless idea and i should try another 2.4 or stay by rc5?
[23:29] <tziOm> nbd, there, duud?
[23:35] <nbd> no
[23:36] <tziOm> ok
[23:37] <tziOm> I know I am doing alot of strange stuff.. but maby you can gain on that too.. making future openwrt more failsafe/idiotproof...
[23:38] <tziOm> or if it is only ment to work with the default confuguration (busybox++) you should maby disable user-configurability of required stuff..
[23:43] <tziOm> What i am about to bug about now is I (hehe..) disabled ifconfig/route and went for ip only.. then looking at the scripts you use only ifconfig/route with no checking..
[23:45] <tziOm> ..even where you dont need to use either.. i.e for iface in $(/sbin/ifconfig -a | awk '{print $1}' | grep eth); => for iface in $(grep eth /proc/net/dev | cut -f1 -d:)
[23:47] <milan_h> how about providing a patch for that?
[23:48] <tziOm> I will rewrite most or all the init system and perhaps post it somewhere
[23:49] <tziOm> alot of stuff is done alot harder than needed, and slower
[23:49] <tziOm> ban me if you want!
[23:50] <blux> someone know if CFE is available in Srr?
[23:50] <blux> src sry
[23:52] <[mbm]> tziOm: so send patches
[23:53] Action: [mbm] sees a lot of questions and complaining but nothing useful
[23:53] <tziOm> When asking questions (have to understand to be able to help) you just harass me.
[23:54] <[mbm]> that's because you claim it's wrong without even knowing what it us
[23:54] <tziOm> alot of the init system is quite complicated (the squashfs => jffs2 mtd stuff.)
[23:54] <[mbm]> right
[23:54] <[mbm]> but it works
[23:54] <tziOm> [mbm], thats not true.. anyway. I will contribute -- lets just get a better tone here.
[23:55] <[mbm]> :P
[23:56] <[mbm]> there's tons of stuff we already know is broken
[23:56] <tziOm> its not nessesarly wrong, but it can be done simpler/better, with less requirements and less commands.. alot of things.. getting same functionality..
[23:56] <crazy_imp> btw, does someone have an idea whats the problem i described?
[23:56] <blux> for something there is no need to be done faster IMHO
[23:57] <CIA-16> nbd * r3904 /branches/buildroot-ng/openwrt/target/linux/au1000-2.6/Makefile: fix unnecessary kernel rebuild
[23:57] <[mbm]> if we already know it's broken then you repeating that it's broken without offering to fix it is just the same as you complaining without helping
[23:58] <CIA-16> nbd * r3905 /branches/buildroot-ng/openwrt/target/linux/ (17 files in 12 dirs): target/linux/image rewrite
[23:58] <[mbm]> and there's nothing we hate more than people who complain
[23:58] <nbd> [mbm]: now the stuff in target/linux/image looks much more like what we have in package/*/Makefile
[23:58] <[mbm]> nbd: cool.
[00:00] --- Mon Jun 5 2006
[03:06] <[mbm]> nbd: we could do the tracing by overriding the default make rules
[03:06] <nbd> how?
[03:06] <[mbm]> well, I mean we could write a .o.c: rule that does the "CC foo.c" thing
[03:07] <nbd> you mean overriding the rules of the individual packages?
[03:08] <[mbm]> well, most packages don't define those rules, they just depend on make defaults
[03:08] <nbd> i think we should get the build system tracing right first
[03:09] <nbd> with different verbosity levels
[03:09] <nbd> then we can see about what we do to the package builds
[03:09] <tziOm> ipkg.c:(.text+0xc): undefined reference to `ipkg_op'
[03:09] <nbd> tziOm: did you change any busybox settings?
[03:09] <tziOm> alot
[03:10] <[mbm]> and you obviously broke it
[03:10] <tziOm> didnt change the cod
[03:10] <tziOm> e
[03:10] <nbd> but you changed the settings
[03:10] <nbd> that can be enough to screw things up
[03:10] <tziOm> seems so
[03:12] <tziOm> I set "compile all sources at once"
[03:12] Action: [mbm] wonders why we're expected to fix something tziOm broke
[03:12] <tziOm> and MD5 => 0 .. 1k bigger and 3 times faster..
[03:13] <tziOm> but I dont think im doing anything _wrong_ .. maby stuff you havent tried?
[03:13] <[mbm]> < nbd> but you changed the settings
[03:13] <tziOm> mbm - whats wrong with You?
[03:13] <[mbm]> no, what's wrong with you
[03:14] <[mbm]> you complain that all the settings are wrong, then you change them all, and then you complain that it doesn't work
[03:14] <tziOm> is it wrong to go through configuration with a careful eye? ..reading and optimizing?
[03:15] <tziOm> i dont just change to change, not born yesterday
[03:15] Action: nbd wonders if tziOm is using gentoo linux
[03:15] <tziOm> slack
[03:15] <nbd> ah
[03:15] <tziOm> since -94
[03:16] <nbd> anyway, changing lots of options and pasting single lines from the logs doesn't give us _anything_ that we can use
[03:16] <tziOm> the log did say something
[03:17] <tziOm> I cant know what you understand and not
[03:17] <[mbm]> we have no idea what you changed, and you're not giving us much information beyond "it broke."
[03:17] <nbd> what you pasted doesn't tell us anything other than 'it's broken'
[03:18] <nbd> if you can't filter the information for us, then you'll have to give us the information in a way that allows us to look for what's relevant
[03:18] <[mbm]> start with the defaults, make sure you can compile that way, then change the settings individually, so that when something breaks you know why it broke
[03:18] <nbd> right
[03:18] <tziOm> think it started when i disabled ipkg support.
[03:18] <tziOm> dont need it
[03:20] <tziOm> now it works
[03:20] <tziOm> and ipkg is disabled.. dunno what happened
[03:20] <[mbm]> :P
[03:21] <tziOm> why elinks and not w3m, which is better? size?
[03:22] <[mbm]> try compiling w3m and see if it's smaller
[03:23] <tziOm> elinks wont compile.. should add some menuconfig configurable options to it.. like --without-openssl (if it works)
[03:25] <tziOm> hmm.. it says --with-openssl=(bah)/staging_dir_mipsel/usr
[03:31] <CIA-16> nbd * r3901 / (2 files in 2 dirs): don't compile libipkg in busybox if ipkg is disabled
[03:31] <nbd> :)
[03:32] <[mbm]> hmm .. could export a MAKEFILES from the top level
[03:32] <tziOm> so.. debugging is bit helpful afterall?
[03:32] <[mbm]> and basically override the default rules globally
[03:32] <nbd> tziOm: sure it is, if it's done right
[03:33] <nbd> [mbm]: i wouldn't do it by default for every package. can introduce lots of potentially weird cases
[03:33] <nbd> [mbm]: i don't trust make to handle things like that properly
[03:33] <[mbm]> nbd: true .. just thinking about things
[03:37] <tziOm> openssl is compiled after elinks and elinks needs openssl...
[03:37] <nbd> i think we should have output similar to what the linux kernel prints for the build system trace
[03:37] <nbd> i.e. when it enters the package/foobar directory it should print CHK package/foobar
[03:37] <[mbm]> the packages/Makefile is used to setup up package dependancies
[03:38] <nbd> then some short strings for download, prepare, configure, compile and install
[03:38] <[mbm]> nbd: sounds good
[03:39] <[mbm]> could do that from the packages/rules.mk
[03:39] <nbd> yeah
[03:39] <tziOm> why the +s on /usr btw?
[03:39] <[mbm]> shouldn't be
[03:39] <tziOm> it is
[03:39] <[mbm]> isn't here
[03:39] <nbd> not on my router either
[03:40] <tziOm> is here.. current cvs, and not changed a thing in perms
[03:40] <nbd> you pulled it from cvs?
[03:40] <tziOm> svn even
[03:40] <[mbm]> we use svn
[03:40] <nbd> ah
[03:40] <nbd> [mbm]: is the cvs still active?
[03:40] <[mbm]> and it's probably a umask thing
[03:41] <[mbm]> nbd: hmm stragenly yes
[03:41] <tziOm> why is it installing shit i dont even want to hear about in this kind of system.. mtsql
[03:41] <tziOm> mysql...
[03:42] <[mbm]> because you told it to?
[03:42] <tziOm> if it is a dependancy.. I would very much like the log to tell me what required it.. because i dont
[03:43] <nbd> [mbm]: let's talk about the V= levels
[03:43] <nbd> [mbm]: how should we define them?
[03:44] <[mbm]> nbd: think they should define depth
[03:45] <tziOm> really!?!!
[03:45] <[mbm]> ?
[03:46] <tziOm> ...now its compiling postgresql also...
[03:46] <[mbm]> ...
[03:51] <tziOm> Your config system is so screwed!
[03:52] <[mbm]> there's a reason why we're writing buildroot-ng :P
[03:52] <tziOm> it is compiling everything I havent enabled I think..
[03:52] <[mbm]> no, you probably enabled it without knowing
[03:52] <tziOm> NO!
[03:52] <nbd> [mbm]: http://pastebin.com/756879
[03:52] <nbd> tziOm: did you enable the sdk?
[03:53] <tziOm> and not a single thing that could have use for libpng either
[03:53] <[mbm]> either that or he did DEVELOPER=1 build
[03:53] <[mbm]> nbd: yeah, something like that
[03:54] <tziOm> yeah, the sdk
[03:54] <[mbm]> the sdk builds all possible libraries
[03:54] <tziOm> yeah?
[03:54] <tziOm> shitty sdk
[03:55] <[mbm]> look, you mind not being so much of an asshole?
[03:55] <nbd> [mbm]: or maybe better: http://pastebin.com/756886
[03:55] <tziOm> sorry
[03:55] <tziOm> but im glad other sdks dont do that
[03:55] <[mbm]> the sdk is used to provide a full environment for building anything, so to build the sdk it has to build all the libraries that people might use
[03:56] <[mbm]> you don't need the sdk because you're building from buildroot
[03:56] <tziOm> hmm.. couldnt ppl choose that, maby?
[03:56] <[mbm]> they are given the choice, you just made the wrong choice
[03:56] <nbd> [mbm]: so how do we implement that kind of output in a clean way?
[03:57] <tziOm> choose what libs to compile
[03:57] <[mbm]> tziOm: turn off the sdk and choose
[03:57] <tziOm> thats ok, but hey
[03:57] <[mbm]> nbd: liked the other one better
[03:57] <nbd> ok
[03:58] Action: nbd can't think of a decent way to implement this without making it look too similar to the old mess
[03:58] <nbd> [mbm]: btw. remember that stupid LED spammer that put is crap in our trac?
[03:58] <nbd> s/is/his/
[03:59] <[mbm]> yeah, the ticket spam
[03:59] <nbd> he's now bugging madwifi
[03:59] Action: [mbm] hates spammers
[04:02] Action: nbd just sent otaku42 a link about that spammer
[04:02] <nbd> (the blog post with the right ip to firewall)
[04:03] <tziOm> tcpdump is huge
[04:03] <[mbm]> yep.
[04:03] <tziOm> It shouldnt install the /usr/lib/ipkg if ipkg is disabled, as is here
[04:04] <nbd> our system expects ipkg to be ther
[04:04] <[mbm]> we didn't expect anyone to be stupid enough to disable ipkg?
[04:04] <tziOm> what the fuck are you saying?
[04:04] Action: [mbm] repeats
[04:04] <[mbm]> we didn't expect anyone to be stupid enough to disable ipkg?
[04:05] <tziOm> Stupid asslicker
[04:05] <[mbm]> do you not understand english?
[04:09] <tziOm> is there no fancy way of compiling test c-code for arch, then running some fancy parser script through code to estimate quite well size of bin?
[04:10] <nbd> [mbm]: so... if $TOPDIR/Makefile calls $TOPDIR/package/Makefile, what do you think would be the right place to add the hook for the V= stuff?
[04:29] <[mbm]> :)
[04:30] <[mbm]> some fancy parser to estimate the size of a binary .. hmm a compiler?
[04:35] <nbd> :)
[05:00] <common> nbd?
[05:01] <common> thought about asking the openixp guys if they want to merge their code to openwrt?
[05:04] <nbd> iirc we were already in contact with the guy runing the project. ask kaloz, he's the one doing the ixp4xx port
[05:06] <common> i shot a wrv on ebay after i saw one can solder a 2nd minipci slot on it
[05:08] <common> i bet the soldering will be 'fun'
[05:11] <common> hrm
[05:11] <common> last changes on xscale are 2 month old
[05:11] <common> in trunk/
[05:14] <common> Kaloz SYN
[06:27] <[mbm]> .
[07:21] <[mbm]> nbd: do we really need BUILD_MAKEFLAGS passed on the make commandline?
[10:39] <nbd> [mbm]: i did this to keep it from passing on -j
[11:17] <tziOm> any particular reason nvram is located in /usr ? .. shouldnt it be in / ?
[12:18] <Kaloz> common: syn/ack
[15:23] <CIA-16> nico * r3902 /trunk/openwrt/package/ (14 files in 5 dirs): add rp-l2tp package (closes: #459).
[15:25] <nbd> {Nico}: ping
[15:26] <{Nico}> nbd: pong
[15:26] <nbd> {Nico}: i think it's about time we freeze the packages in trunk and port everything over to buildroot-ng
[15:27] <nbd> {Nico}: what do you think?
[15:27] <{Nico}> nbd: i have some updates pending, but yeah, i think its time
[15:28] <nbd> also: did you read the thread that i started on -devel?
[15:28] <nbd> (mail)
[15:29] <{Nico}> nbd: i read it and i agree
[15:30] <nbd> can you write a response?
[15:30] <{Nico}> nbd: will do
[15:31] <nbd> thanks
[16:23] <tziOm> in 2.4 kernel config .. the "Broadcom Home Network Division" .. is it the binary bcm43xx driver
[16:29] <nbd> no
[16:32] <tziOm> ndb - ok.. its the CONFIG_WL and /linux/drivers/net/wl
[16:32] <tziOm> nbd - whats the binary part of it?
[16:34] <tziOm> Are there alternatives to tcpdump that doesnt take up that much space?
[16:35] <nbd> wl is the driver for bcm43xx, right
[16:36] <tziOm> so it has to be a module? cant be put into kernel?
[16:40] <tziOm> nbd, how can I choose what gcc version to use?
[16:40] <CIA-16> nbd * r3903 /branches/buildroot-ng/openwrt/package/rules.mk: whitespace fixes
[16:41] <nbd> tziOm: you shouldn't mess with the gcc version
[16:41] <nbd> tziOm: some targets are quite sensitive to that
[16:41] <tziOm> seems you are using 3.4.6..
[16:43] <nbd> it depends on what target you choose
[16:43] <tziOm> nbd - after building .. does the openwrt-brcm-2.4-squashfs.trx (in my case) contain kernel and filesystem and everything?
[16:43] <nbd> broadcom 2.6 uses 3.4.6
[16:43] <nbd> broadcom 2.4 uses 3.4.4
[16:43] <nbd> some targets use 4.1.0
[16:45] <tziOm> nbd, I noticed BUSYBOX_CONFIG_BUILD_AT_ONCE .. gcc 4.1 and obove..
[17:01] <tziOm> could anyone help me understand how openwrt works when it comes to: building filesystem image from build_ARCH/root, building kernel/modules, and where are these placed? is everything combined in .trx file?
[17:02] <nbd> everything is in the trx file
[17:03] <nbd> it's a container for loader, kernel image and filesystem
[17:06] <tziOm> nbd, ok.. how do i generate that file after changeing kernel and root ?
[17:06] <tziOm> and where are the kernel modules?
[17:06] <nbd> how to generate it? just run make
[17:06] <nbd> the build system will take care of it
[17:07] <tziOm> wont it change my root (which I have modified) and change kenrnel config too (which is modified running make menuconfig ARCH=mips inside build_mipsel/linux)?
[17:08] <nbd> don't change the root manually
[17:08] <tziOm> Point is I want a bit more control/customizing than the "buildroot" menuconfig gives me..
[17:08] <nbd> openwrt is built on the assumption that all files are packaged
[17:08] <tziOm> is it no make root
[17:09] <tziOm> or make kernel and make final?
[17:09] <nbd> if you make sure that each of your modified files is in a package, then it'll do the right thing
[17:09] <tziOm> package?..
[17:09] <tziOm> nbd, lets say I delete some files and dirs from build_mipsel/root and change some config files..
[17:09] <nbd> don't manually edit stuff in build_mipsel/root
[17:10] <nbd> leave that directory alone
[17:10] <nbd> it's temporary
[17:11] <tziOm> ndb - its the easiest way.. cant you just tell me howto make a squashfs image from current root? is it no make option for going directly to that step?
[17:11] <nbd> no
[17:12] <tziOm> no? why so hostile?
[17:12] <nbd> there is no make option for that
[17:12] <nbd> that's what i meant
[17:12] <tziOm> can I cd to a dir and make to make it work then? or do I have to make a script to do it for me?
[17:13] <nbd> the build system was not designed for having its temporary directories messed with
[17:13] <tziOm> in the last case, could you please tell me commands to run to 1) make squashfs filesystem 2) make kernel and modules and 3) combine those into .trx image
[17:14] <nbd> i don't remember the exact command
[17:14] <nbd> commands
[17:14] <nbd> you have to look them up yourself
[17:14] <nbd> i never build images manually
[17:15] <tziOm> nbd - what is the common way of attacking it when you want to manually change kernel options and filesystem stuff then? I can probably copy the .config somewhere after my menuconfig ARCH=mips?
[17:16] <nbd> you can edit the kernel config in target/linux/brcm-2.6/config (for the broadcom 2.6 target)
[17:17] <tziOm> nbd - that sounds like the same as using menuconfig then copying .. better..
[17:18] <tziOm> nbd, im using 2.4 .. since 2.6 seems to be having problems with its bcm43xx support in openwrt
[17:21] <tziOm> nbd, are You familiar with the bootup process?
[17:21] <nbd> sure
[17:21] <tziOm> ok, what is run and in what order?
[17:22] <nbd> just look at the init scripts
[17:22] <tziOm> what is run first then
[17:24] <tziOm> You dont know?
[17:24] <milan_h> everything starts at /etc/preinit I think
[17:24] <tziOm> kernel calls preinit? .. no!
[17:24] <tziOm> or
[17:25] <milan_h> root@OpenWrt:~# cat /proc/cmdline
[17:25] <milan_h> [...] init=/etc/preinit [...]
[17:26] <nbd> yes, the kernel calls preinit
[17:26] <tziOm> yeah, ok. thanks milan_h
[17:26] <nbd> sorry, if i don't go into detail on questions like that, but i have better things to do
[17:26] <nbd> some things you can just look up yourself
[17:26] <tziOm> nbd, bah
[17:27] <tziOm> hmm.. but how is the squashfs stuff handeled?.. this happends after squashfs image in uncompressed to ram..
[17:28] <nbd> the squashfs image is not uncompressed
[17:28] <nbd> it's being mapped to a mtd partition and mounted directly
[17:29] <tziOm> nbd, so it will be used as is from flash?
[17:29] <nbd> right
[17:30] <tziOm> ok...
[17:33] <tziOm> the mtd is only binary? where can i find any docs on it?
[17:33] <nbd> it's a kernel subsystem
[17:33] <tziOm> not the util
[17:34] <nbd> do you even try to look in the build system to figure out things like that yourself?
[17:34] <tziOm> yep
[17:34] <tziOm> but found only mtd and mtd-static in build_mipsel/mtd
[17:34] <nbd> and obvious place to look would have been package/mtd
[17:34] <nbd> build_mipsel is a temporary directory
[17:35] <tziOm> there are other sources there, but thaks.
[17:36] <tziOm> is that mtd documented anywhere?
[17:36] <nbd> it has a command line help
[17:40] <db90h> i wish i had your patience nbd
[17:42] <tziOm> hmm.. so its running preinit to figure out if it should run failsafe and telnetd (without checking existance of telnetd..) .. then if not, its running mount_root (maby add /sbin/ here to avoid possible problems) .. mount root mounts tmpfs and does some mtd unlock linux stuff.. then it checks what fs it is, and if like my case squashfs it runs includes /bin/firstboot
[17:42] <tziOm> nbd, am I right this far
[17:47] <tziOm> you require hexdump actually... hmm..
[17:49] <nbd> it doesn't always mount tmpfs
[17:49] <nbd> if the jffs2 part is valid, it mounts this instead
[17:50] <tziOm> nbd, should the . /etc/preinit.arch in preinit be preinit.${arch} after setting arch ?
[17:51] <nbd> no
[17:55] <tziOm> nbd, I dont understand exactly what your is_dirty function does.. (and be warned, you are using a bin (hexdump) that ppl easily drop in config.. maby add comment)
[17:55] <nbd> you're the first one i know that actually wanted to drop it
[17:56] <tziOm> wont need it after the bootup...
[17:56] <nbd> maybe you don't need it
[17:57] <nbd> but it can be useful sometimes
[17:57] <tziOm> who needs it?
[17:57] <tziOm> it can be.. but not on the system im making,
[17:57] <nbd> feel free to write equivalent shell code if you like
[17:57] <tziOm> but anyway.. if there are other ways of getting the info you need, maby you should consider keeping the bootup requirements down
[17:58] <nbd> since there's lots of other stuff to do, i won't bother doing something about this right now
[17:59] <tziOm> nbd, can you please explain what it is you are doing in the is_dirty?
[17:59] <nbd> tziOm: checking a flag in a byte that we reserved inside the trx
[17:59] <tziOm> nbd, for?
[18:00] <nbd> read the code
[18:00] <nbd> don't make me explain every single line
[18:00] <nbd> because i won't do it
[18:00] <tziOm> if you should switch to jffs2..
[18:02] <tziOm> nbd, another thing maby ppl are more likely to do is dropping telnetd and including sshd (dropbear) .. this will cause boot failure in preinit, its not checking existance of telnetd before executing... nor checking for dropbear..
[18:03] <nbd> please take some more time to figure out how the system works
[18:03] <db90h> lol, leave him alone trzi0m.. these guys put enough uncompensated time into openwrt w/o being harassed
[18:03] <tziOm> its advice.. i could do it.. its 2 lines of shellscripting
[18:03] <tziOm> but i dont have access to code modifications
[18:03] <nbd> tziOm: you're making wrong assumptions ... lots of them
[18:04] <tziOm> nbd, for instance?
[18:04] <nbd> that dropbear is not included by default
[18:04] <nbd> that the dropbear init script is there even if you deselect it
[18:05] <tziOm> what?
[18:05] <nbd> or that telnet is always enabled
[18:05] <tziOm> i was talking about the bootup, and spesifically the FAILSAFE mode.
[18:05] <nbd> then say that
[18:05] <nbd> anyway
[18:05] <nbd> dropbear is unsuitable for failsafe mode
[18:06] <nbd> telnetd is much simpler
[18:06] <tziOm> so is telnetd if its not there, and why unsuitable?
[18:06] Action: nbd is getting more and more annoyed
[18:06] <db90h> trizi0m: even if u make valid arguments, its easy to be a critic
[18:06] <nbd> just be silent for a while
[18:06] <db90h> ;p
[18:07] <nbd> db90h: lots of 'maybe i want this', 'maybe i want that' ...
[18:07] <nbd> maybe there are lots of other places that are way more important...
[18:07] <db90h> yep ;)
[18:07] <nbd> or 'maybe' people that write code are more useful than people that complain about code :)
[18:07] <db90h> hehehe
[18:08] <nbd> anyway, next time i won't be so patient with this guy
[18:08] <db90h> i dunno how you are so patient to begin with.. i'd lose my mind quite quickly
[18:09] <db90h> you are a much better man than I ;)
[18:09] <nbd> ;p
[18:13] Action: nbd is happy with the progress that buildroot-ng is making
[18:13] <nbd> almost done rewriting most of the code in target/linux/image...
[18:53] <common> Kaloz: ACK
[18:53] <db90h> SYN
[18:53] <common> db90h: RST
[18:54] <db90h> ;p
[19:29] <Kaloz> common: so?
[19:32] <common> whats zup with wrv support?
[19:33] <common> nbd told me you were about to merge the 2.6 diffs from openixp, but i saw little to nothing in that direction
[19:49] <nbd> i didn't say that
[19:50] <nbd> i said that kaloz was working on xscale stuff
[19:53] <nbd> let's see if he can behave now
[20:03] <tziOm> no hardware watchdog in the wlhdd?
[20:03] <nbd> no
[20:09] <florian> is there a developper meeting tonight ?
[20:10] <nbd> no idea. haven't heard anything
[20:10] <florian> ok
[20:10] <nbd> florian: btw. did you read the email thread that i started on -devel?
[20:11] <florian> yes I did, did not take time to answer yet
[20:11] <nbd> ok. i'd like to have some feedback from all the devs, that's why i'm asking
[20:11] <florian> right
[20:12] <common> is it possible to get a ro account on the ml for non dev lurkers?
[20:13] <nbd> ro account? everybody can already send, so if someone were to get read access, he'd have an rw account already :)
[20:13] <nbd> anyway, this issue is still being discussed
[20:13] <common> everybody can send to the ml?
[20:14] <nbd> yeah
[20:15] <common> just subscribed to users and devel
[20:17] <common> devel is pending, users empty
[20:58] <[mbm]> wow, is it 20,000 questions day?
[20:59] <[mbm]> biggest problem with tzi0m is that he doesn't understand anything, yet it's all wrong and needs to be changed
[21:10] <[mbm]> the worst part is that all this work that was done explaining to tzi0m will just be lost in the irc archives
[21:40] <nbd> [mbm]: btw. i think we should have an official freeze for the work on kamikaze packages
[21:40] <nbd> until everything's ported over to buildroot-ng
[21:43] <[mbm]> hrm ..
[21:43] <[mbm]> you know what happened with the last freeze
[21:43] <[mbm]> but yeah, probably should
[21:46] <nbd> well, this is not a stability freeze
[21:46] <nbd> it's a 'force people to convert stuff to the new format' freeze
[21:46] <nbd> so the result will most certainly be different :)
[21:47] <[mbm]> btw about the MAKEFLAGS do we really need to pass those via the commandline? everything includes rules.mk so we should just be able to set it there
[21:48] <florian__> [mbm]: do we make a developper meeting tonight ?
[21:48] <[mbm]> florian__: I hadn't planned one
[21:48] <florian__> ok
[21:49] <nbd> [mbm]: you can't override it from rules.mk
[21:49] <nbd> [mbm]: make doesn't allow it for some reason - i tried it
[21:51] <[mbm]> weird
[21:51] <[mbm]> just looks nasty on the commandline
[22:00] <florian__> there is a lot of new people here, welcome !
[22:00] <nbd> yeah, ever since this channel was made public
[22:47] <crazy_imp> has someone running kamikaze build (2.6 kernel) on a wl-hdd2.5? my shows strange behavior like bringing the network interface up and down in about 30 sec distance and its still responding to the nvram ip (but there are still ~80% lost) and to the default ip there are 100% lost. telnet is imposible :(. any ideas?
[22:57] <crazy_imp> or can someone say, that it's an hopless idea and i should try another 2.4 or stay by rc5?
[23:29] <tziOm> nbd, there, duud?
[23:35] <nbd> no
[23:36] <tziOm> ok
[23:37] <tziOm> I know I am doing alot of strange stuff.. but maby you can gain on that too.. making future openwrt more failsafe/idiotproof...
[23:38] <tziOm> or if it is only ment to work with the default confuguration (busybox++) you should maby disable user-configurability of required stuff..
[23:43] <tziOm> What i am about to bug about now is I (hehe..) disabled ifconfig/route and went for ip only.. then looking at the scripts you use only ifconfig/route with no checking..
[23:45] <tziOm> ..even where you dont need to use either.. i.e for iface in $(/sbin/ifconfig -a | awk '{print $1}' | grep eth); => for iface in $(grep eth /proc/net/dev | cut -f1 -d:)
[23:47] <milan_h> how about providing a patch for that?
[23:48] <tziOm> I will rewrite most or all the init system and perhaps post it somewhere
[23:49] <tziOm> alot of stuff is done alot harder than needed, and slower
[23:49] <tziOm> ban me if you want!
[23:50] <blux> someone know if CFE is available in Srr?
[23:50] <blux> src sry
[23:52] <[mbm]> tziOm: so send patches
[23:53] Action: [mbm] sees a lot of questions and complaining but nothing useful
[23:53] <tziOm> When asking questions (have to understand to be able to help) you just harass me.
[23:54] <[mbm]> that's because you claim it's wrong without even knowing what it us
[23:54] <tziOm> alot of the init system is quite complicated (the squashfs => jffs2 mtd stuff.)
[23:54] <[mbm]> right
[23:54] <[mbm]> but it works
[23:54] <tziOm> [mbm], thats not true.. anyway. I will contribute -- lets just get a better tone here.
[23:55] <[mbm]> :P
[23:56] <[mbm]> there's tons of stuff we already know is broken
[23:56] <tziOm> its not nessesarly wrong, but it can be done simpler/better, with less requirements and less commands.. alot of things.. getting same functionality..
[23:56] <crazy_imp> btw, does someone have an idea whats the problem i described?
[23:56] <blux> for something there is no need to be done faster IMHO
[23:57] <CIA-16> nbd * r3904 /branches/buildroot-ng/openwrt/target/linux/au1000-2.6/Makefile: fix unnecessary kernel rebuild
[23:57] <[mbm]> if we already know it's broken then you repeating that it's broken without offering to fix it is just the same as you complaining without helping
[23:58] <CIA-16> nbd * r3905 /branches/buildroot-ng/openwrt/target/linux/ (17 files in 12 dirs): target/linux/image rewrite
[23:58] <[mbm]> and there's nothing we hate more than people who complain
[23:58] <nbd> [mbm]: now the stuff in target/linux/image looks much more like what we have in package/*/Makefile
[23:58] <[mbm]> nbd: cool.
[00:00] --- Mon Jun 5 2006