[05:09] <CIA-17> nbd * r4323 /branches/buildroot-ng/openwrt/package/ (42 files in 27 dirs): rewrite of the network scripts and configuration
[05:09] <nbd> finally... :)
[05:09] <groz> oh my, so this is gonna really break things now i guess
[05:09] <nbd> it works much better for me
[05:10] <nbd> but the config file format is completely different
[05:10] <groz> I'm just getting ready to go get some dinner, but, i'll look at it when i get back
[05:11] <nbd> k
[05:18] <CIA-17> nbd * r4324 /branches/buildroot-ng/openwrt/package/dnsmasq/files/dnsmasq.init: only set dhcp_enable if lan proto is set to static
[12:22] <florian__> good morning !
[12:23] <nbd> good morning
[12:40] <florian__> nbd: what kind of usb fixes do you have ?
[12:43] <nbd> asus has backported some stuff from linux 2.6
[12:43] <nbd> in their wl-500gP gpl release
[12:44] <nbd> i have isolated the actual fixes and ported them to whiterussian
[12:46] <florian__> ah great
[12:48] <nbd> i just don't know if they actually work :)
[12:48] <florian__> yes that's why you wanted some help
[12:49] <florian__> there was a guy yesterday, who had a wl-500gP
[12:58] <florian__> I don't remember his pseudo
[12:59] <nbd> i'll just ask in the channel again later
[12:59] <florian__> sure
[13:30] <florian__> nbd: I wrote to johannes berg for the brcm63xx adsl driver reverse engineering and he seems pretty motivated too :)
[13:30] <nbd> great
[13:31] <nbd> talking to him about the usb net stuff and other bcm963xx gpl violations right now
[13:31] <nbd> i mentioned the nasty EXPORT_SYMBOL_GPL hack
[13:31] <nbd> he says gregkh won't like this :)
[13:35] <florian__> I just wrote to gregkh :)
[13:36] <nbd> :D
[13:37] <florian__> I doubt he will appreciate knowing this :)
[13:38] <florian__> the good point is that in some tarballs, including a comtrend one, we have object files that are not relocated yet, and not kmod
[13:39] <florian__> I think it's easier when it comes to reversing this kind of files, rather than a stripped kmod right ?
[13:40] <nbd> yeah, i think so
[13:41] <nbd> btw. i'll get a 63xx board soon
[13:41] <nbd> then i can help with the kernel stuff
[13:42] <florian__> great
[13:42] <florian__> currently I have added the bootloader specific code, and I just need to test it :)
[13:42] <florian__> then I could focus on this physmap driver
[13:43] <nbd> sebastian told me that the serial console is still not working correctly
[13:52] <florian__> I did not mentionned any issue with my device
[13:52] <florian__> and as I don't know what he merged, how he merged it ...
[13:53] <nbd> he probably added lots of weird hacks
[13:53] <florian__> yes
[13:54] <florian__> this board is so, so crappy
[13:54] <florian__> I think it's even worst than the 947xx
[13:54] <nbd> at least it doesn't have this annoying SiliconBackplane mess
[13:54] <nbd> so cleaning up the system code might be a bit easier
[13:58] <florian__> yep
[13:59] <florian__> the pcmcia driver seems to be same as far as I have read the patchs
[14:19] <nbd> seems like the workarounds in the usb patch did enable
[14:19] <nbd> and it didn't break in any obvious way
[14:19] <nbd> so there's a good chance it'll work
[14:20] <nbd> 8 parallel read processes in usb2 mode on an usb stick also worked
[14:21] <nbd> trying 4 reads and 4 writes at the same time ...
[14:35] <nbd> cool.. the patch also seems to increase performance
[14:35] <nbd> i just wrote data to an usb stick with >4 MB/s and only 30% cpu loaed
[14:39] <florian__> ah great
[14:47] <CIA-17> nbd * r4325 /branches/ (2 files in 2 dirs): usb fixes and performance enhancements
[14:47] <nbd> :)
[15:21] <CIA-17> nbd * r4326 /branches/buildroot-ng/openwrt/package/ (5 files in 4 dirs): fix more instances of '' abuse
[15:23] <nbd> oops... that should have been '$type' instead of ''
[15:24] <nbd> well, trac even leaves out the ''
[16:42] <CIA-17> nbd * r4327 /branches/buildroot-ng/openwrt/package/ipsec-tools/Makefile: fix ipsec-tools (missing include, missing dependencies)
[16:51] <CIA-17> nbd * r4328 /branches/buildroot-ng/openwrt/package/fuse/Makefile: fuse: put PKG_BUILD_DIR in KERNEL_BUILD_DIR
[17:27] <CIA-17> nbd * r4329 /branches/buildroot-ng/openwrt/ (5 files in 4 dirs): port ppp modules to package/kernel and fix ppp package dependencies
[17:51] <CIA-17> florian * r4330 /packages/net/aiccu/ (. Config.in Makefile ipkg/): Port aiccu to buildroot-ng
[18:12] <CIA-17> florian * r4331 /packages/net/aircrack-ng/ (. Config.in Makefile ipkg/): Port aircrack-ng to -ng, also update to latest version : 0.6
[18:23] <CIA-17> nbd * r4332 /branches/buildroot-ng/openwrt/package/ (diag/Makefile shfs/Makefile spca5xx/Makefile zd1211/Makefile): BUILD_DIR -> KERNEL_BUILD_DIR
[18:25] <florian__> champagne for the number 5000 commit !
[18:25] <nbd> hehe
[18:25] <nbd> l
[18:36] <florian__> I really wonder why uclibc keep using their buildroot where they can use our
[18:36] <db90h> ?
[18:37] <db90h> is uclibc's buildroot shitty or something?
[18:38] <florian__> it's pretty non abstracted
[18:38] <florian__> usually, when you build a system images, the linux kernel compilation is done after, so that modules are not included in the final image
[18:38] <florian__> it uses too much fakeroo
[18:38] <florian__> fakeroot sorry
[18:40] <florian__> the only advantage is that is it usually well known by people, so that you don't spend too much time on explaining them how it works
[18:40] <florian__> and how to port packages to it
[18:42] <florian__> so yes db90h for me it's highly perfectible
[18:42] <db90h> ic
[18:43] <CIA-17> florian * r4333 /packages/net/amwall/ (. Config.in Makefile ipkg/): Port amwall to -ng
[18:44] <florian__> db90h: can you explain what you wanted to do with a stripping symbol proxy ;) ?
[18:44] <db90h> hmm, BusyBox 1.2.0 on June 30th
[18:45] <florian__> {Nico}: has pending patches for it
[18:46] <db90h> well it's really kinda silly and useless, I was just daydreaming.. now i think it's better to just go through the filesystem and strip all ELFs, with an exclusions.. but the 'proxy', which i mis-labelled completely, was just an sstrip replacement that queued all sstrip requests, so at the end of the build you go and apply the queued sstrip requests (with any needed path adjustments)
[18:48] <db90h> i have a lot of crazy ideas, i like to think of sideways approaches to problems.. most of the time they don't work out, but everyone once and a while i hit upon something useful ;p
[18:48] <florian__> yes
[18:48] Action: nbd already knows how to implement the function stripping in buildroot-ng
[18:48] <florian__> actually I did not see the point in queuing stripping requests
[18:48] <db90h> yea, it's not a problem of how to do it at all
[18:49] <db90h> yea, there isn't a point at all
[18:49] <florian__> because unless you modifiy the source of the program to share code, it's quite hard to gain space
[18:49] <florian__> the only thing a proxy will do is suppress doubles
[18:50] <db90h> hu?
[18:50] <db90h> i should have never used the word proxy, it's really wrong in this case
[18:50] <florian__> I think I see what you mean
[18:50] <florian__> by proxy you mean a queue right ?
[18:50] <florian__> or a buffer ?
[18:50] <db90h> yea
[18:51] <florian__> that's also how I understood the thing
[18:51] <florian__> so, the queue will only suppress doubles
[18:51] <db90h> doubles?
[18:51] <florian__> oups, doublons ?
[18:52] <db90h> i was just talking about queue'ing the sstrip requests so they are delayed until the end.. the reason to do this because binutils doesn't work with post-sstrip'd modules, so u can't enumerate symbols and discover what can be pruned from the runtime library
[18:53] <db90h> i still don't understand the doubles part :( the queue'ing is so useless we need to just forget it now, i feel dumber everytime its mentioned for even thinking of it to start with
[18:53] <florian__> when you find the same thing twice, it something that is "doubled" right ?
[18:54] <db90h> yea, but presumably the build system doesn't invoke sstrip mulitiple times for a single binary..
[18:54] <db90h> sstrip doesn't have anything to do with actually enumerating symbols
[18:55] <db90h> or maybe i'm all confused, that happens frequently
[18:55] <florian__> no it's me
[18:55] <florian__> I see what you mean
[18:55] <CIA-17> nbd * r4334 /branches/buildroot-ng/openwrt/package/kernel/modules.mk: fix typo
[18:55] <db90h> at first i thought sstrip was 'symbol strip', but ot
[18:55] <db90h> its intention is not that
[18:56] <florian__> symbol strip is the real strip
[18:56] <db90h> ah, ok, so my early confusion was reasonable then ;)
[18:56] <CIA-17> nbd * r4335 /branches/buildroot-ng/openwrt/package/kernel/modules.mk: oops... another one
[19:06] <CIA-17> nbd * r4336 /branches/buildroot-ng/openwrt/package/kernel/Makefile: fix kernel postinst scripts
[19:09] <CIA-17> nbd * r4337 /branches/buildroot-ng/openwrt/ (Makefile package/Makefile): create a package index
[19:11] <db90h> florian__: keep in mind that what everyone is calling 'symbol stripping' is really apparently runtime library pruning (removing unused objects from uclibc or glibc)
[19:11] <db90h> i had misunderstood this for some time and i don't think 'symbol stripping' accurate describes library pruning
[19:22] <CIA-17> nico * r4338 /packages/utils/lcd4linux/ (. Config.in Makefile Makefile ipkg/ patches/ patches/): port lcd4linux to buildroot-ng
[19:40] <CIA-17> florian * r4339 /packages/net/arpd/ (. Config.in Makefile ipkg/): Port arpd to -ng
[19:42] <florian__> arptables should directly go in buildroot-ng right ?
[19:43] <nbd> yep
[19:43] <florian__> ok
[19:53] <CIA-17> florian * r4340 /branches/buildroot-ng/openwrt/package/arptables/ (. Makefile): Port arptables to -ng
[19:55] <florian__> nbd: do you have a package example of how I can handle conffiles ?
[19:56] <nbd> define Package/<foobar>/conffiles
[19:56] <nbd> then the contents of conffilers
[19:56] <nbd> and then just endef
[19:56] <florian__> oki, thanks !
[20:00] <CIA-17> nico * r4341 /packages/utils/lcd4linux/Makefile: add missing copyright notice, remove unneeded PKG_INSTALL_DIR
[20:03] <crazy_imp> {Nico} do you never need to download new sources?
[20:06] <florian__> crazy_imp: you mean update package at the same time ?
[20:07] <crazy_imp> florian__: no i rewrote a part of the downloadscript for the sourecs, so resume is supported now, and {Nico} said he will take a look at it
[20:17] <florian__> ah ok
[20:19] <CIA-17> florian * r4342 /packages/net/arpwatch/ (. Config.in Makefile ipkg/): Port arpwatch to -ng
[20:20] <florian__> {Nico}: I think I will let asterisk for later ;)
[22:01] <[g2]> Kaloz what ever happend with the xscale port ?
[22:02] <Bartman007> [g2]: it is still being worked on
[22:02] <Bartman007> not ready yet
[22:02] <Bartman007> it is possible to build images for xscale targets, but they aren't very useful yet except for a limited number of devices
[22:03] <[g2]> Bartman007 thx, could you be more specific ?
[22:04] <[g2]> Bartman007 you probably don't know, but I've basically have my own xscale platform
[22:05] <Bartman007> I have a gateway 7001 and I know the ethernet driver isn't commited for that deviec
[22:06] <[g2]> last I spoke to Kaloz he was futzing with the 2.1.1 drivers and they were unhappy
[22:06] <Bartman007> there are people here working w/ wrv54's but I don't know if ethernet is working yet for those either.
[22:06] <[g2]> I've been running ethernet on 2.6.16 since 9/04
[22:06] <[g2]> on the XScale NPE's that is
[22:06] <Bartman007> ... 2.6.16 in 9/04 ?
[22:07] <[g2]> ok it started at 2.6.7
[22:07] <Bartman007> hehe, :-P
[22:07] <[g2]> I'm running on 16 now with the 2.1 stuff a
[22:08] <[g2]> there's also some guy doing open drivers
[22:08] <[g2]> there were some posts on L-A-K
[22:09] <[g2]> so if it's just kernel/wifi/board support I've got all that stuff
[22:10] <[g2]> I was looking at firstboot the other week at how the jffs2 was setup from the squashfs
[22:10] <Bartman007> I'm even having trouble getting RedBoot to load my kernel but I would assume it is due to ignorance on my behalf as I've never worked with it before
[22:11] <[g2]> what problem are you having ? are you running BE or LE ?
[22:13] <Bartman007> I'm not sure what those are, but I can build a kernel and upload it to the device, but the device hangs as sson as it attempts to boot the kernel
[22:13] <[g2]> Big Endian or Little Endian
[22:13] <Bartman007> oh, heh.
[22:13] <[g2]> :)
[22:14] <[g2]> are you running 1.92 or 2.x Redboot ?
[22:14] <[g2]> I'd guess that's in BE
[22:14] <Bartman007> 1.92
[22:14] <[g2]> most likely BE, that's the Redboot from the NSLU2 too
[22:15] <[g2]> are you building for BE ?
[22:15] <Bartman007> yes
[22:16] <[g2]> a 2.6.x kernel ?
[22:17] <Bartman007> 2.6.16.7
[22:18] <[g2]> Ok are you using the toolchain from svn ?
[22:18] <Bartman007> I'm using the trunk as xscale isn't supported in buildroot-ng yet
[22:18] <Bartman007> so yes.
[22:20] <[g2]> does the buildroot-ng build any working images yet ?
[22:21] <Bartman007> for other targets
[22:21] <[g2]> yeah that's fine
[22:22] <[g2]> basically it seems like there hasn't been a working xscale image yet
[22:22] <Bartman007> kaloz has been making them from trunk for his 7001, I just can't get it working on mine yet
[22:23] <[g2]> mine == ??
[22:23] <Bartman007> my Gateway 7001
[22:23] <Bartman007> http://wiki.openwrt.org/OpenWrtDocs/Hardware/Gateway/7001
[22:25] <[g2]> Bartman007 what part of the world are you in ?
[22:25] <[g2]> which continent
[22:25] <Bartman007> North America
[22:25] <[g2]> Ok cool. I'm in NC
[22:26] <Bartman007> California here
[22:27] <[g2]> did anything every happen with the DSM-G600 R1s ?
[22:30] <Bartman007> I've heard some minor talk about the g600's and AFAIK one of the NSLU2 team's have gotten it booting
[22:30] <[g2]> yeah that was me
[22:30] <Bartman007> oh, lol.
[22:31] <[g2]> I put a Debian rootfs and kernel online back in april
[22:31] <[g2]> as the Velocity is on the G600 and there's no Intel NPE coded needed
[22:31] <Bartman007> I'll be back later, but I gotta run right now.
[22:32] <[g2]> ok cheers
[22:32] <[g2]> http://www.giantshoulderinc.com/g600/
[22:47] <[g2]> what's the option to enable printing everything that the trunk make file does ? it's ??=1
[22:48] <[g2]> s/option/env var/
[23:06] <Bartman007> well, there is make V=99
[23:09] <[g2]> thx that's the one I was looking for
[23:10] Action: [g2] kicks off xscale build from trunk
[23:56] <florian__> nbd: how do I handle postinst files ?
[23:57] <florian__> should I do everything in Package/foo/install ?
[00:00] --- Mon Jul 31 2006
[05:09] <nbd> finally... :)
[05:09] <groz> oh my, so this is gonna really break things now i guess
[05:09] <nbd> it works much better for me
[05:10] <nbd> but the config file format is completely different
[05:10] <groz> I'm just getting ready to go get some dinner, but, i'll look at it when i get back
[05:11] <nbd> k
[05:18] <CIA-17> nbd * r4324 /branches/buildroot-ng/openwrt/package/dnsmasq/files/dnsmasq.init: only set dhcp_enable if lan proto is set to static
[12:22] <florian__> good morning !
[12:23] <nbd> good morning
[12:40] <florian__> nbd: what kind of usb fixes do you have ?
[12:43] <nbd> asus has backported some stuff from linux 2.6
[12:43] <nbd> in their wl-500gP gpl release
[12:44] <nbd> i have isolated the actual fixes and ported them to whiterussian
[12:46] <florian__> ah great
[12:48] <nbd> i just don't know if they actually work :)
[12:48] <florian__> yes that's why you wanted some help
[12:49] <florian__> there was a guy yesterday, who had a wl-500gP
[12:58] <florian__> I don't remember his pseudo
[12:59] <nbd> i'll just ask in the channel again later
[12:59] <florian__> sure
[13:30] <florian__> nbd: I wrote to johannes berg for the brcm63xx adsl driver reverse engineering and he seems pretty motivated too :)
[13:30] <nbd> great
[13:31] <nbd> talking to him about the usb net stuff and other bcm963xx gpl violations right now
[13:31] <nbd> i mentioned the nasty EXPORT_SYMBOL_GPL hack
[13:31] <nbd> he says gregkh won't like this :)
[13:35] <florian__> I just wrote to gregkh :)
[13:36] <nbd> :D
[13:37] <florian__> I doubt he will appreciate knowing this :)
[13:38] <florian__> the good point is that in some tarballs, including a comtrend one, we have object files that are not relocated yet, and not kmod
[13:39] <florian__> I think it's easier when it comes to reversing this kind of files, rather than a stripped kmod right ?
[13:40] <nbd> yeah, i think so
[13:41] <nbd> btw. i'll get a 63xx board soon
[13:41] <nbd> then i can help with the kernel stuff
[13:42] <florian__> great
[13:42] <florian__> currently I have added the bootloader specific code, and I just need to test it :)
[13:42] <florian__> then I could focus on this physmap driver
[13:43] <nbd> sebastian told me that the serial console is still not working correctly
[13:52] <florian__> I did not mentionned any issue with my device
[13:52] <florian__> and as I don't know what he merged, how he merged it ...
[13:53] <nbd> he probably added lots of weird hacks
[13:53] <florian__> yes
[13:54] <florian__> this board is so, so crappy
[13:54] <florian__> I think it's even worst than the 947xx
[13:54] <nbd> at least it doesn't have this annoying SiliconBackplane mess
[13:54] <nbd> so cleaning up the system code might be a bit easier
[13:58] <florian__> yep
[13:59] <florian__> the pcmcia driver seems to be same as far as I have read the patchs
[14:19] <nbd> seems like the workarounds in the usb patch did enable
[14:19] <nbd> and it didn't break in any obvious way
[14:19] <nbd> so there's a good chance it'll work
[14:20] <nbd> 8 parallel read processes in usb2 mode on an usb stick also worked
[14:21] <nbd> trying 4 reads and 4 writes at the same time ...
[14:35] <nbd> cool.. the patch also seems to increase performance
[14:35] <nbd> i just wrote data to an usb stick with >4 MB/s and only 30% cpu loaed
[14:39] <florian__> ah great
[14:47] <CIA-17> nbd * r4325 /branches/ (2 files in 2 dirs): usb fixes and performance enhancements
[14:47] <nbd> :)
[15:21] <CIA-17> nbd * r4326 /branches/buildroot-ng/openwrt/package/ (5 files in 4 dirs): fix more instances of '' abuse
[15:23] <nbd> oops... that should have been '$type' instead of ''
[15:24] <nbd> well, trac even leaves out the ''
[16:42] <CIA-17> nbd * r4327 /branches/buildroot-ng/openwrt/package/ipsec-tools/Makefile: fix ipsec-tools (missing include, missing dependencies)
[16:51] <CIA-17> nbd * r4328 /branches/buildroot-ng/openwrt/package/fuse/Makefile: fuse: put PKG_BUILD_DIR in KERNEL_BUILD_DIR
[17:27] <CIA-17> nbd * r4329 /branches/buildroot-ng/openwrt/ (5 files in 4 dirs): port ppp modules to package/kernel and fix ppp package dependencies
[17:51] <CIA-17> florian * r4330 /packages/net/aiccu/ (. Config.in Makefile ipkg/): Port aiccu to buildroot-ng
[18:12] <CIA-17> florian * r4331 /packages/net/aircrack-ng/ (. Config.in Makefile ipkg/): Port aircrack-ng to -ng, also update to latest version : 0.6
[18:23] <CIA-17> nbd * r4332 /branches/buildroot-ng/openwrt/package/ (diag/Makefile shfs/Makefile spca5xx/Makefile zd1211/Makefile): BUILD_DIR -> KERNEL_BUILD_DIR
[18:25] <florian__> champagne for the number 5000 commit !
[18:25] <nbd> hehe
[18:25] <nbd> l
[18:36] <florian__> I really wonder why uclibc keep using their buildroot where they can use our
[18:36] <db90h> ?
[18:37] <db90h> is uclibc's buildroot shitty or something?
[18:38] <florian__> it's pretty non abstracted
[18:38] <florian__> usually, when you build a system images, the linux kernel compilation is done after, so that modules are not included in the final image
[18:38] <florian__> it uses too much fakeroo
[18:38] <florian__> fakeroot sorry
[18:40] <florian__> the only advantage is that is it usually well known by people, so that you don't spend too much time on explaining them how it works
[18:40] <florian__> and how to port packages to it
[18:42] <florian__> so yes db90h for me it's highly perfectible
[18:42] <db90h> ic
[18:43] <CIA-17> florian * r4333 /packages/net/amwall/ (. Config.in Makefile ipkg/): Port amwall to -ng
[18:44] <florian__> db90h: can you explain what you wanted to do with a stripping symbol proxy ;) ?
[18:44] <db90h> hmm, BusyBox 1.2.0 on June 30th
[18:45] <florian__> {Nico}: has pending patches for it
[18:46] <db90h> well it's really kinda silly and useless, I was just daydreaming.. now i think it's better to just go through the filesystem and strip all ELFs, with an exclusions.. but the 'proxy', which i mis-labelled completely, was just an sstrip replacement that queued all sstrip requests, so at the end of the build you go and apply the queued sstrip requests (with any needed path adjustments)
[18:48] <db90h> i have a lot of crazy ideas, i like to think of sideways approaches to problems.. most of the time they don't work out, but everyone once and a while i hit upon something useful ;p
[18:48] <florian__> yes
[18:48] Action: nbd already knows how to implement the function stripping in buildroot-ng
[18:48] <florian__> actually I did not see the point in queuing stripping requests
[18:48] <db90h> yea, it's not a problem of how to do it at all
[18:49] <db90h> yea, there isn't a point at all
[18:49] <florian__> because unless you modifiy the source of the program to share code, it's quite hard to gain space
[18:49] <florian__> the only thing a proxy will do is suppress doubles
[18:50] <db90h> hu?
[18:50] <db90h> i should have never used the word proxy, it's really wrong in this case
[18:50] <florian__> I think I see what you mean
[18:50] <florian__> by proxy you mean a queue right ?
[18:50] <florian__> or a buffer ?
[18:50] <db90h> yea
[18:51] <florian__> that's also how I understood the thing
[18:51] <florian__> so, the queue will only suppress doubles
[18:51] <db90h> doubles?
[18:51] <florian__> oups, doublons ?
[18:52] <db90h> i was just talking about queue'ing the sstrip requests so they are delayed until the end.. the reason to do this because binutils doesn't work with post-sstrip'd modules, so u can't enumerate symbols and discover what can be pruned from the runtime library
[18:53] <db90h> i still don't understand the doubles part :( the queue'ing is so useless we need to just forget it now, i feel dumber everytime its mentioned for even thinking of it to start with
[18:53] <florian__> when you find the same thing twice, it something that is "doubled" right ?
[18:54] <db90h> yea, but presumably the build system doesn't invoke sstrip mulitiple times for a single binary..
[18:54] <db90h> sstrip doesn't have anything to do with actually enumerating symbols
[18:55] <db90h> or maybe i'm all confused, that happens frequently
[18:55] <florian__> no it's me
[18:55] <florian__> I see what you mean
[18:55] <CIA-17> nbd * r4334 /branches/buildroot-ng/openwrt/package/kernel/modules.mk: fix typo
[18:55] <db90h> at first i thought sstrip was 'symbol strip', but ot
[18:55] <db90h> its intention is not that
[18:56] <florian__> symbol strip is the real strip
[18:56] <db90h> ah, ok, so my early confusion was reasonable then ;)
[18:56] <CIA-17> nbd * r4335 /branches/buildroot-ng/openwrt/package/kernel/modules.mk: oops... another one
[19:06] <CIA-17> nbd * r4336 /branches/buildroot-ng/openwrt/package/kernel/Makefile: fix kernel postinst scripts
[19:09] <CIA-17> nbd * r4337 /branches/buildroot-ng/openwrt/ (Makefile package/Makefile): create a package index
[19:11] <db90h> florian__: keep in mind that what everyone is calling 'symbol stripping' is really apparently runtime library pruning (removing unused objects from uclibc or glibc)
[19:11] <db90h> i had misunderstood this for some time and i don't think 'symbol stripping' accurate describes library pruning
[19:22] <CIA-17> nico * r4338 /packages/utils/lcd4linux/ (. Config.in Makefile Makefile ipkg/ patches/ patches/): port lcd4linux to buildroot-ng
[19:40] <CIA-17> florian * r4339 /packages/net/arpd/ (. Config.in Makefile ipkg/): Port arpd to -ng
[19:42] <florian__> arptables should directly go in buildroot-ng right ?
[19:43] <nbd> yep
[19:43] <florian__> ok
[19:53] <CIA-17> florian * r4340 /branches/buildroot-ng/openwrt/package/arptables/ (. Makefile): Port arptables to -ng
[19:55] <florian__> nbd: do you have a package example of how I can handle conffiles ?
[19:56] <nbd> define Package/<foobar>/conffiles
[19:56] <nbd> then the contents of conffilers
[19:56] <nbd> and then just endef
[19:56] <florian__> oki, thanks !
[20:00] <CIA-17> nico * r4341 /packages/utils/lcd4linux/Makefile: add missing copyright notice, remove unneeded PKG_INSTALL_DIR
[20:03] <crazy_imp> {Nico} do you never need to download new sources?
[20:06] <florian__> crazy_imp: you mean update package at the same time ?
[20:07] <crazy_imp> florian__: no i rewrote a part of the downloadscript for the sourecs, so resume is supported now, and {Nico} said he will take a look at it
[20:17] <florian__> ah ok
[20:19] <CIA-17> florian * r4342 /packages/net/arpwatch/ (. Config.in Makefile ipkg/): Port arpwatch to -ng
[20:20] <florian__> {Nico}: I think I will let asterisk for later ;)
[22:01] <[g2]> Kaloz what ever happend with the xscale port ?
[22:02] <Bartman007> [g2]: it is still being worked on
[22:02] <Bartman007> not ready yet
[22:02] <Bartman007> it is possible to build images for xscale targets, but they aren't very useful yet except for a limited number of devices
[22:03] <[g2]> Bartman007 thx, could you be more specific ?
[22:04] <[g2]> Bartman007 you probably don't know, but I've basically have my own xscale platform
[22:05] <Bartman007> I have a gateway 7001 and I know the ethernet driver isn't commited for that deviec
[22:06] <[g2]> last I spoke to Kaloz he was futzing with the 2.1.1 drivers and they were unhappy
[22:06] <Bartman007> there are people here working w/ wrv54's but I don't know if ethernet is working yet for those either.
[22:06] <[g2]> I've been running ethernet on 2.6.16 since 9/04
[22:06] <[g2]> on the XScale NPE's that is
[22:06] <Bartman007> ... 2.6.16 in 9/04 ?
[22:07] <[g2]> ok it started at 2.6.7
[22:07] <Bartman007> hehe, :-P
[22:07] <[g2]> I'm running on 16 now with the 2.1 stuff a
[22:08] <[g2]> there's also some guy doing open drivers
[22:08] <[g2]> there were some posts on L-A-K
[22:09] <[g2]> so if it's just kernel/wifi/board support I've got all that stuff
[22:10] <[g2]> I was looking at firstboot the other week at how the jffs2 was setup from the squashfs
[22:10] <Bartman007> I'm even having trouble getting RedBoot to load my kernel but I would assume it is due to ignorance on my behalf as I've never worked with it before
[22:11] <[g2]> what problem are you having ? are you running BE or LE ?
[22:13] <Bartman007> I'm not sure what those are, but I can build a kernel and upload it to the device, but the device hangs as sson as it attempts to boot the kernel
[22:13] <[g2]> Big Endian or Little Endian
[22:13] <Bartman007> oh, heh.
[22:13] <[g2]> :)
[22:14] <[g2]> are you running 1.92 or 2.x Redboot ?
[22:14] <[g2]> I'd guess that's in BE
[22:14] <Bartman007> 1.92
[22:14] <[g2]> most likely BE, that's the Redboot from the NSLU2 too
[22:15] <[g2]> are you building for BE ?
[22:15] <Bartman007> yes
[22:16] <[g2]> a 2.6.x kernel ?
[22:17] <Bartman007> 2.6.16.7
[22:18] <[g2]> Ok are you using the toolchain from svn ?
[22:18] <Bartman007> I'm using the trunk as xscale isn't supported in buildroot-ng yet
[22:18] <Bartman007> so yes.
[22:20] <[g2]> does the buildroot-ng build any working images yet ?
[22:21] <Bartman007> for other targets
[22:21] <[g2]> yeah that's fine
[22:22] <[g2]> basically it seems like there hasn't been a working xscale image yet
[22:22] <Bartman007> kaloz has been making them from trunk for his 7001, I just can't get it working on mine yet
[22:23] <[g2]> mine == ??
[22:23] <Bartman007> my Gateway 7001
[22:23] <Bartman007> http://wiki.openwrt.org/OpenWrtDocs/Hardware/Gateway/7001
[22:25] <[g2]> Bartman007 what part of the world are you in ?
[22:25] <[g2]> which continent
[22:25] <Bartman007> North America
[22:25] <[g2]> Ok cool. I'm in NC
[22:26] <Bartman007> California here
[22:27] <[g2]> did anything every happen with the DSM-G600 R1s ?
[22:30] <Bartman007> I've heard some minor talk about the g600's and AFAIK one of the NSLU2 team's have gotten it booting
[22:30] <[g2]> yeah that was me
[22:30] <Bartman007> oh, lol.
[22:31] <[g2]> I put a Debian rootfs and kernel online back in april
[22:31] <[g2]> as the Velocity is on the G600 and there's no Intel NPE coded needed
[22:31] <Bartman007> I'll be back later, but I gotta run right now.
[22:32] <[g2]> ok cheers
[22:32] <[g2]> http://www.giantshoulderinc.com/g600/
[22:47] <[g2]> what's the option to enable printing everything that the trunk make file does ? it's ??=1
[22:48] <[g2]> s/option/env var/
[23:06] <Bartman007> well, there is make V=99
[23:09] <[g2]> thx that's the one I was looking for
[23:10] Action: [g2] kicks off xscale build from trunk
[23:56] <florian__> nbd: how do I handle postinst files ?
[23:57] <florian__> should I do everything in Package/foo/install ?
[00:00] --- Mon Jul 31 2006