**** BEGIN LOGGING AT Sat Apr 26 02:59:57 2008 Apr 26 07:13:21 wtf? why do i have a symlink packages/openwrt-packages pointing to ../feeds/packages? is that normal? Apr 26 07:19:04 delete it Apr 26 07:19:55 older revisions linked it there, now it's in feeds/openwrt or something like that (I don't use ./scripts/feeds) Apr 26 07:20:08 err, package/feeds/openwrt Apr 26 07:29:53 yep, i'v deleted it Apr 26 07:31:47 i'm not that familiar with svn but i guess it should delete this symlink? or was it created by a script? Apr 26 07:33:13 let me guess: it was created by make packages/symlinks? Apr 26 07:34:27 yeah, that runs ./scripts/feeds Apr 26 07:34:41 yes, i see: https://dev.openwrt.org/browser/trunk/Makefile?rev=9990 line 62 Apr 26 07:35:45 at least make distclean should remove that symlink. i'm pretty surprised that only i ran in that problem.. Apr 26 07:38:59 i think everyone upgrading from a revision prior changeset 10634 will have that symlink even after running make distclean, which lead to a corruption of the packageinfo cache in tmp/info and therefor to a broken menuconfig selection (all packages in feeds/ are doubled, because they are symlinked twice to packages/) Apr 26 07:40:46 do you by chance have package/xwrt-packages (or something similar)? Apr 26 07:41:55 no, i just have feeds/xwrt/ Apr 26 07:42:30 and packages/feeds/xwrt with the packages pointing to feeds/xwrt/ Apr 26 07:44:47 well i dont really now where the removing of the symlink should placed.. its completly redundant after removing that symlink and for everyone checking out the trunk today the first time.. Apr 26 07:45:14 fish_: I'm tossing it in distclean Apr 26 07:45:29 okay Apr 26 07:49:44 agb * r10941 /trunk/include/toplevel.mk: Apr 26 07:49:44 Remove old feeds symlink in distclean. Thanks fish Apr 26 07:49:44 - Fixes duplicate package entries when coming from pre r10634 Apr 26 07:50:19 ah crap. typo. Apr 26 07:51:35 :-( Apr 26 07:51:40 agb * r10942 /trunk/include/toplevel.mk: Fix typo in last commit Apr 26 08:11:08 juhosg * r10943 /trunk/target/linux/generic-2.6/patches-2.6.25/ (209-mini_fo.patch 210-mini_fo_2.6.25_fixes.patch): [kernel] fix mini_fo on 2.6.25 Apr 26 08:12:32 juhosg: \o/ Apr 26 08:12:48 :P Apr 26 08:13:12 juhosg: 1 Kaloz: 0 Apr 26 08:17:40 Bartman007: i have discussed it with Kaloz, so i would prefer the 1:1 Apr 26 08:20:44 juhosg: .5 Kaloz: .5 Apr 26 08:22:27 Bartman007: thanks Apr 26 08:42:15 * Kaloz slaps Bartman007 Apr 26 08:43:34 that'll be -1 points for Kaloz Apr 26 08:43:52 :) Apr 26 08:44:04 specially as I was again the moron who did the generic patches :P Apr 26 08:44:51 well, it's not so hard, is it? Apr 26 08:45:42 i mean quilt really takes out most of the annoying parts of that Apr 26 08:47:08 sometimes. anyways, I'm the moron who does this all the time since .21 Apr 26 08:47:35 i have a question: what do you think about http://freigeist.org/static/openwrt-patches/10-uuid-automount ? i think that such functionality would be nice to have for many people. but that depends on vol_id which is about 40k big, so i think only users who wants that feature should ne to install it. should i make a package out of this? and whats with the current usb storage automount stuff? should i replace it after installing the package? Apr 26 08:48:17 btw: the script is a hotplug script which mounts a usb storage device to a mountpoint specified by a uuid in /etc/config/fstab Apr 26 08:49:06 Talking of quilt - I seem to have found a problem with either quilt or the way its used by the make system Apr 26 08:50:06 If a quilt patch fails, it isn't backing out properly, and ends up deleting the files, not reverting them. I think it's because the reject file are going in the wrong place. Apr 26 08:50:07 http://freigeist.org/static/openwrt-patches/etc-config-fstab-automount-example.txt <- thats my /etc/config/fstab Apr 26 08:50:19 well, a part of it Apr 26 08:54:29 http://openwrt.pastebin.ca/998216 Apr 26 08:55:08 weird Apr 26 08:55:10 As you can see - the reject name is the same as the target, so when it cleans up, it removes the whole target Apr 26 08:55:13 never happened to me Apr 26 08:55:30 i think it's unlikely that this is a problem in the openwrt integration of quilt Apr 26 08:55:41 or maybe i should extend the functionality to check whether vol_id is installed and if not, implement the simple mounting like /sbin/usb-storage Apr 26 08:55:55 Indeed - I couldn't see any way you'd changed the reject file names Apr 26 08:56:23 fish_: That sounds better, if it's not too large Apr 26 08:56:43 fish_: i think it would be useful if vol_id was only one criterium for selecting disks for mounting Apr 26 08:57:03 fish_: just make it accept either device name, volume id or some other stuff Apr 26 08:57:11 without fixed dependency Apr 26 08:57:21 so if you install vol_id, the volume id check will work Apr 26 08:57:25 nbd: Which version of quilt are you using? Apr 26 08:57:25 if not, you can choose any other method Apr 26 08:57:42 yes, i thought about something like that Apr 26 08:57:46 noz: server has 0.45, laptop has 0.46 Apr 26 09:17:24 Hmm. Quilt push has the following snippet of code in: Apr 26 09:17:38 awk ' Apr 26 09:17:38 /^patching file / { filename = substr($0, 15) } Apr 26 09:17:38 { gsub(/ -- saving rejects to (file )?.*/, Apr 26 09:17:38 " -- rejects in file " filename) } Apr 26 09:17:38 { print } Apr 26 09:18:06 Don't really know what the substr ($0,15) is doing, but it looks bogus Apr 26 09:21:54 noz: i guess it extracts the filename after "patching file ..." Apr 26 09:22:08 but 15? Apr 26 09:23:51 noz: well, when that meas it starts with character 15 that would be correct i guess.. Apr 26 09:24:48 fish@phoibe ~ $ echo patching file foobar | awk '/^patching file / { print substr($0, 15) }' Apr 26 09:24:51 foobar Apr 26 09:25:09 its ugly but it works ;) Apr 26 09:29:03 It seems to be related to the condition when the patch fails. quilt rolls back, then tries to reverse-apply the patch, then rolls that back Apr 26 09:29:25 nbd: Have you used quilt in the condition that the patch fails to apply? Apr 26 09:36:00 short question: anyone know ad hoc what the '1' means: config_get_bool enabled "$cfg" "enabled" '1'? --> defaults to 1 if not set? Apr 26 09:46:23 kaloz * r10944 /trunk/target/linux/ixp4xx/patches-2.6.25/ (200-npe_driver.patch 202-npe_hss.patch): split npe ethernet and hss patch, sync with upstream Apr 26 09:53:38 http://freigeist.org/static/openwrt-patches/10-usb-storage <- not yet testet (and normaly my scripts need a lot of testing ;)) but does that met what you were thinking about? Apr 26 09:55:05 uhm, i should have to move that vol_id checking/getting uuid block out there, it is not necessary to check that everytime. uhm Apr 26 09:57:15 well, but the conecpt is: get whether there is vol_id installed, if so, get the uuid. that it checkes in every config section if either the uuid matches the device option of the device matches the current hotpluged device. in those cases it uses the config options for mounting. if not it tried to mount it to /mnt/$device Apr 26 10:00:07 kaloz * r10945 /trunk/package/ixp4xx-microcode/ (Makefile src/IxNpeMicrocode.h): generate and install NPE-A ethernet microcode as well Apr 26 10:02:56 kaloz * r10946 /trunk/target/linux/ixp4xx/patches-2.6.24/402-ixp4xx_ehci_backport.patch: backported ixp4xx ehci driver Apr 26 10:08:52 kaloz * r10947 /trunk/target/linux/generic-2.6/patches-2.6.24/950-mtd_cfi_intel_p33_compatbility.patch: Intel P3x chips use CFI 1.5 -- treat them the same way as 1.4 ones Apr 26 10:12:10 kaloz * r10948 /trunk/toolchain/Config.in: PXA has Xscale core as well, so optimize the same way as we do on IXP/IOP Apr 26 10:14:56 kaloz * r10949 /trunk/target/linux/pxa/Makefile: PXA can use EABI, too Apr 26 14:04:12 the source file in https://dev.openwrt.org/browser/trunk/package/mac80211/Makefile Apr 26 14:04:24 does not exist at http://linuxwireless.org/download/compat-wireless-2.6/ Apr 26 14:16:59 florian * r10950 /trunk/target/linux/rb532/config-2.6.23: Re-enabled the CompactFlash driver for rb532 Apr 27 00:04:27 anyone working on the airlink ar680w? Apr 27 01:47:35 florian * r10951 /trunk/package/mac80211/Makefile: Use the locally mirrored compat-wireless tarblall (#3377) **** ENDING LOGGING AT Sun Apr 27 02:59:57 2008