**** BEGIN LOGGING AT Sun Jul 01 02:59:56 2007 Jul 01 03:42:50 nbd * r7815 /trunk/package/busybox/patches/270-ash_performance.patch: Jul 01 03:42:50 massively reduce the number of forks that busybox ash has to do while processing our scripts Jul 01 03:42:52 this is accomplished simply by changing the order in which it looks for builtins and applets Jul 01 06:44:30 just woke up from a few hours of sleep, have coffee in left hand - morning to you all :) Jul 01 06:45:09 Morning Jul 01 06:45:25 We still in WIP-day? Jul 01 06:45:50 noz: yes + i have Beastie Boys on the soundsystem :) Jul 01 06:46:25 Ugh. Not sure I could work with that on, but to each his own.... Jul 01 06:47:58 noz: i need to wake up Jul 01 06:48:11 Fair enough! Jul 01 06:48:51 noz: you have any comments in regard to wip-day, suggestions, ideas whatever? Jul 01 06:50:15 Well, it's a damn good idea. It's been a bit quiet, but that shouldn't put us off. I've not been able to spend much time on it. Jul 01 06:50:22 (this time) Jul 01 06:51:59 noz: as one of the wellknown actors/politicians said = i'll be back :) Jul 01 06:52:25 It hasn't been very focussed, but it might actually be better that way. One way might be to agree some ticket numbers to work on in advance Jul 01 06:54:27 On the other hand, it could potentially work better if it's just devs responding to people on the day. The main driving force should come from "everyone", rather than the devs, since they tend to work on things anyway. Jul 01 06:55:46 I see the biggest hurdle to getting tickets closed as actually being able to test things on certain hardware, and that should be a lot easier on an organised day Jul 01 06:55:52 (I'll shut up now) Jul 01 07:02:56 noz: good suggestions Jul 01 07:03:09 noz: thanx Jul 01 07:20:55 <[mbm]> re. Jul 01 07:22:44 <[mbm]> right; main thing seems to be finding a time when everyone is available, as things tend to get done when there's communication going on Jul 01 07:27:37 probably impossible for everyone to be around at the same time - perhaps better that people beforehand announce they'll be around? Jul 01 07:28:37 <[mbm]> well, my experience has been that people are around but you'd never know it unless they happen to make a remark on irc Jul 01 07:29:10 Another idea could be to follow-up on the idea noz suggested 'about agreeing on ticket number' --> Jul 01 07:29:43 <[mbm]> well, to some extent we've done that by specifying certain tickets under the 7.07 category Jul 01 07:30:52 asking members to look for tickets which have being "laying" around for a while, put them in a list (depending on platform) and then use this as 'point-of-departure'? Jul 01 07:31:49 <[mbm]> the usual criteria for selecting tickets to resolve is how active they are Jul 01 07:32:00 <[mbm]> eg, the /reports/11 on dev Jul 01 07:33:15 <[mbm]> it's somewhat interesting that we still have whiterussian tickerts, mostly on the basis that unless the development model somehow changes, those tickets will never be fixed Jul 01 07:33:40 <[mbm]> so it becomes a question of - do we live the old whiterussian tickets open, with no hope of fixing them? Jul 01 07:33:49 <[mbm]> or do we close them? Jul 01 07:34:27 [mbm]: At some point, we'll close them as "wontfix", but I guess not until Kamikaze is stable and the most common used Jul 01 07:35:35 <[mbm]> we need to work if we want people to switch to kamikaze Jul 01 07:35:49 <[mbm]> even though we have a release, most people are still installing whiterussian Jul 01 07:36:10 [mbm]: What do you think about #1640? I've done the gpio.h fix, but do we want i2c mods in the mainstream? Jul 01 07:36:37 [mbm]: Eventually, we are going to want them to switch - the question is when.... Jul 01 07:37:07 <[mbm]> I've been wanting the users to switch for the last year Jul 01 07:37:20 <[mbm]> but from their view it's just impractical Jul 01 07:37:27 <[mbm]> so it's not about what "we" want Jul 01 07:38:55 <[mbm]> re #1640, I see no reason to combine the gpio with i2c Jul 01 07:39:27 <[mbm]> afaik there are no i2c devices on the routers Jul 01 07:39:49 <[mbm]> so I don't understand the purpose of the patch Jul 01 07:46:46 That's why I asked before putting it in. I'm sure it was intended for someone who has done a GPIO->I2C mod on their router. Jul 01 07:47:47 If you're going to add an I2C device, using 2 GPIO pins is the obvious way to do it. Jul 01 07:47:51 <[mbm]> not sure if I understand it, but it seems to be an i2c frontend for the gpio, such that you can toggle gpios just by sending i2c commands Jul 01 07:48:08 <[mbm]> question is why you'd want to do that Jul 01 07:49:33 Indeed. Means you can use in-kernel drivers for i2c devices, like sensors, but your question about why is very valid ;) Jul 01 07:51:58 So I don't think it belongs in trunk. After all, it would depend on which GPIO pins you'd chosen to connect to, and that would be different on each model. I suppose you could parameterise it, but it's not elegant Jul 01 07:53:30 <[mbm]> on a related note, did anyone ever parameterize the mmc stuff? I know that's been a peeve of mine for some time now Jul 01 07:53:41 Don't know. Jul 01 07:54:08 (Actually the i2c driver is properly parameterised) Jul 01 07:54:11 <[mbm]> I don't use the mmc stuff, and really don't think all that highly of it, or I'd fix it Jul 01 07:58:17 Has the mmc stuff gone in? Jul 01 07:59:02 <[mbm]> don't think so; I know I left it out for exactly that reason when I was last asked about it Jul 01 08:00:10 I guess we need policy on how much we're going to support various hardware hacks. In general, the people that are capable of them are capable of running a locally-modified version Jul 01 08:00:53 (although hardware skill != software skill, but they tend to be correlated in one direction) Jul 01 08:02:03 <[mbm]> suppose we need to determine who our audience really is Jul 01 08:02:19 <[mbm]> there are people that have no iea about linux or networking that seemt o run openwrt Jul 01 08:02:53 <[mbm]> which are the distant end users Jul 01 08:03:08 <[mbm]> then there's anotehr group of end users that prefers a web interface Jul 01 08:03:21 <[mbm]> having some basic skills Jul 01 08:03:40 <[mbm]> and it moves up the chain to linux admins and hardware hackers Jul 01 08:04:20 <[mbm]> I've always considered our mission to basically provide a commandline linux distro, allowing anyone to put whatever they wanted ontop Jul 01 08:04:32 <[mbm]> but se wee to compete with oterh distros like ddwrt for users Jul 01 08:04:41 <[mbm]> which I don't quite understand why Jul 01 08:06:27 <[mbm]> my stragegy of providing the basic framework and letting others develop ontop only goes so far Jul 01 08:06:41 <[mbm]> in other words, we have a kamikaze release Jul 01 08:07:14 <[mbm]> but the layers ontop of that like documetnation and a web interface are moving much slower than I would like Jul 01 08:10:53 [mbm]: it keeps looping back to a mess of dependencies Jul 01 08:11:40 <[mbm]> yeah I know Jul 01 08:11:58 <[mbm]> a lot of it is just preception Jul 01 08:12:22 <[mbm]> we need to restructure the site so the information is more organized and there are places to put things Jul 01 08:13:45 [mbm]: so this could be a discussion/working subject for july, and a framework could then be presented @ next wip-day? Jul 01 08:17:30 <[mbm]> it's not about saying thigns need work, it's about finding ways to improve them Jul 01 08:18:36 glp_home: yay, catching you once online :P so, just wanted to say that the paper is quite nice to say at least Jul 01 08:18:48 <[mbm]> paper? Jul 01 08:19:30 glp_home: oh, and probably [mbm] would love to read it, too :) may I send it over to him? Jul 01 08:21:04 Kaloz: just do it it's GPL :) Jul 01 08:21:21 yeah, but not a released version yet :) Jul 01 08:23:16 Kaloz: it's kind'o released as it is included as part of a workshop @ an academic conference - but it still needs a good work-over before it might move towards a journal Jul 01 08:24:03 [mbm]: check your inbox Jul 01 08:24:07 Kaloz: but people can quote it + circulate it Jul 01 08:25:23 i think i'll go a take a shower, so i'm back in a little while :) Jul 01 08:56:56 <[mbm]> glp_home: hmm paper is a bit misleading and gets some events wrong Jul 01 08:58:08 [mbm]: i'm open for comments/fixes Jul 01 08:59:45 <[mbm]> at the end of page 5 you seem to be suggesting that openwrt only cane into existance after sveasoft changed to a subscription model Jul 01 09:00:19 <[mbm]> while that may be a point of increased popularity, it was not the formation of the openwrt project Jul 01 09:01:21 <[mbm]> see http://forum.openwrt.org/viewtopic.php?pid=46022#p46022 Jul 01 09:01:52 [mbm]: probably not a good formulation - i know it was not the case, but there seems to enough evidence to suggest that it increased interest Jul 01 09:02:02 florian * r7816 /packages/net/wifidog/Makefile: Update wifidog to 1.1.3-rc1 (#1966) Jul 01 09:02:25 <[mbm]> let me know when you've read through t he above log Jul 01 09:02:31 * [mbm] still reading the paper Jul 01 09:04:12 <[mbm]> I wouldn't say sveasoft's downfall had anything to do with openwrt as much as it did jame's increasing need for control, which drove developers away from his project Jul 01 09:05:05 <[mbm]> openwrt's popularity on the other hand is probably tied to it's rapid development cycle Jul 01 09:05:19 <[mbm]> (although admittedly slow release cycle) Jul 01 09:07:58 florian * r7817 /packages/utils/dmidecode/ (. Makefile): Add dmidecode (#1776) Jul 01 09:19:58 florian * r7818 /packages/net/asterisk/files/asterisk.init: Correctly symlink astdb for first time running or upgrading (#1295) Jul 01 09:30:18 juhosg * r7819 /trunk/target/linux/ (5 files in 3 dirs): [adm5120] rewrite of memory detection code, should be fix #1909 Jul 01 09:45:24 florian * r7820 /trunk/ (8 files in 6 dirs): Remove patch-cmdline from the packages and move it tools tools/. adm5120 and rb532 do use it for procuding ready-to-use kernels (#1631) Jul 01 09:46:51 [mbm]: did 'johnf' ever post a continuation of the irc conversation Jul 01 09:47:23 florian * r7821 /trunk/target/linux/adm5120-2.6/profiles/RB1xx.mk: patch-cmdline is now gone Jul 01 09:47:52 <[mbm]> glp_home: there was some quick discussions about the slides he was going to use for his presentation, but those were pretty much just bullet points Jul 01 09:48:09 <[mbm]> I never did see any followup as to how well it went Jul 01 09:48:14 [mbm]: + would be interesting to know what i actually presented @ penguicon (found the program and title, but no slides) Jul 01 09:48:22 i = he Jul 01 09:48:38 <[mbm]> don't know Jul 01 09:49:03 [mbm]: but nice destillation of the history Jul 01 09:49:12 <[mbm]> had a number of people contact me and say they're working on articles or whatever but for the number of people that have done that, there have been very few actual articles Jul 01 09:49:30 <[mbm]> and rarely do they get the history right Jul 01 09:52:16 <[mbm]> honestly I'm somewhat sick of the "openwrt vs. ____" and the "turn a $60 router into a $600 router" Jul 01 09:52:41 [mbm]: history is always interpretation Jul 01 09:53:44 <[mbm]> a $60 router will always be a $60 router nomatter what you do to it; you might add more functionality but the resources and quality of the actual device remain the same, and the capacity you can add to it is rather limited Jul 01 09:54:37 <[mbm]> ratehr scares me when people suggest running large deployments using a wrt54g with modded firmware Jul 01 09:54:51 [mbm]: but the adding (despite its minor amount) makes a difference Jul 01 09:55:05 <[mbm]> well sure, the normal firmware sucks Jul 01 09:55:47 <[mbm]> but to say that a cheap off the shelf router running openwrt is the equivilant of a $600 enterprise access point is rather misleading Jul 01 09:57:02 <[mbm]> and as I said in the log files, openwrt really started as a way to unify development Jul 01 09:57:27 <[mbm]> so the whole openwrt vs mentality is counterproductive Jul 01 09:58:19 <[mbm]> it's more or less "this is what suits my needs, feel free to expand upon it", I don't feel like we're trying to compete with anyone Jul 01 09:58:43 <[mbm]> yet everyone seems interested in asking questions like "is openwrt better than ___ Jul 01 09:59:23 [mbm]: i think thats what i try to close in on, by using the reference to 'magic' - openwrt as project wants to be productive, but there are boundaries. Jul 01 09:59:39 <[mbm]> I used to be in the habit of telling those people "openwrt sucks, you probably shouldn't use it" just because if that's your mentality, I probably don't want to deal with you Jul 01 10:00:29 mbm: I seem to have missed a great deal of this conversation. Can you please tell me what do you think about openwrt in commercial use /products? Jul 01 10:00:35 <[mbm]> stopped doing that so much lately because my intentions obviously weren't clear, and people used it to reenforce the ieda that openwrt sucked Jul 01 10:01:59 <[mbm]> xrg_: imho openwrt provides a perfect starting point, but it shouldn't be "let's use openwrt as the base and build a proprietary application ontop of that" if openwrt is used, it would be nice to have some recongnition or contributions back to the project Jul 01 10:02:52 mbm: Of course.. I would rather suggest for a frontend, on top of a regular openwrt.. Jul 01 10:03:35 mbm: The idea would be to lure some embedded devels to base their products on openwrt, instead of having their own pathetic distros. Jul 01 10:03:43 [mbm]: having and maintaining a relationship is important Jul 01 10:04:16 xrg_: why should you lure them? Jul 01 10:04:27 florian * r7822 /packages/sound/streamripper/ (. Makefile): Add streamripper (#1803) Jul 01 10:04:40 And, on the other hand, to have those devels contribute back to the project. This gives openwrt more resources (man-hours).. Jul 01 10:05:27 glp_home: I have noted that many "closed" linux distros for embedded lack the developer power to keep up with the changes etc. Jul 01 10:06:07 So, open source will let them collaborate and share work.. (in an ideal world, of course) Jul 01 10:06:47 In commercial terms, everything has a motive, therefore I say "lure" = give them a motive to work that way. Jul 01 10:06:48 <[mbm]> there really isn't much incentive for the vendors to collaborate with open source Jul 01 10:07:18 <[mbm]> it's seen as a one way mechanism to allow them to fast track development Jul 01 10:07:43 <[mbm]> pull the latest gpl stuff and build your framework ontop of that Jul 01 10:08:00 <[mbm]> they forget that they're selling hardware, not creating a software monopoly as well Jul 01 10:08:06 xrg_: i'm somewhat sceptical in respect to this 'homo calculus' notion .. Jul 01 10:08:32 <[mbm]> and so there's the concept that you can't possibly release your firmware, that's a trade secret, what would prevent the competition from leveraging that work Jul 01 10:09:00 Yet, a few os projects have managed to balance with their commercial use (see. kernel for one) .. Jul 01 10:09:24 [mbm]: kind of trying to write a paper on the 'blackbox' problem. Jul 01 10:10:02 <[mbm]> yep; I'm getting somewhat annoyed with my employer about those issues, but not the type of thing I care to disuss Jul 01 10:11:14 [mbm]: i have this synology DS207 which is quite nice, except they try to hard to close the firmware off (debian), and i would just like to be able to add nfs without having to go through a lot of hassle. Jul 01 10:11:54 * [mbm] was one of the early tivo hackers and fully supports the "tivo clause" in gpl3 Jul 01 10:13:23 <[mbm]> unfortunately that causes issues with gpl adoption at a company level Jul 01 10:14:01 <[mbm]> basically companies don't want to fund development of code they can't claim some sort of ownership or control over Jul 01 10:15:08 [mbm]: but the thing is that there are different notions of ownership. Jul 01 10:17:29 <[mbm]> it wouldn't be too much of a stretch to say that my view of the corporate world is a screw your neighbour capitalistic greed, where there are clear lines drawn between motivation and profit Jul 01 10:18:23 [mbm]: :) Jul 01 10:18:29 <[mbm]> don't quote me on that however, as it's an exaggeration to illustrate a point Jul 01 10:19:40 [mbm]: will not :) Jul 01 10:19:41 <[mbm]> but I have yet to see any forum of ultristic behaviour beyond the startup phases Jul 01 10:19:52 <[mbm]> s/forum/form/ Jul 01 10:20:17 <[mbm]> there's a marked transition between a priovate company and a public company Jul 01 10:20:32 [mbm]: in what way? Jul 01 10:20:41 <[mbm]> where as soon as the ipo hits, everything gets seen through the filter of "will this increase my stock price?" Jul 01 10:21:32 .. and there is where it strongly conflicts with GPL usage.. Jul 01 10:22:16 <[mbm]> it turns into an internal audit of what are you doing and how does that drive the stock up Jul 01 10:22:58 <[mbm]> admittedly much of that pressure comes from the investors and just cascades down through the management chain Jul 01 10:23:55 [mbm]: yeah - felt the transition. talked to a company (in the bay area who uses openwrt) about coming and visit, but then the whole thing changed --> now we're to "big" to spend time on people visiting us. Jul 01 10:25:28 <[mbm]> xrg_: yep, which ties back to my earlier reference to issues with my employer; if you connect the dots it's probably not too hard to figure out who I work for and what I do Jul 01 10:25:46 <[mbm]> but, as I said, I'd rather keep my personal life out of this Jul 01 10:28:19 My whole point is that open source devels are doing something good and should be paid for that! Said that, we should not isolate OS away from what gives money. Jul 01 10:28:37 Of course, I don't know the way! But it's something we should work on. Jul 01 10:30:31 <[florian]> there are pretty succesful business who contributed a lot to linux like mvista Jul 01 10:30:47 <[florian]> or others doing good things like o-hand Jul 01 10:30:58 <[florian]> meaning that the business model can be viable Jul 01 10:32:01 surely. And, I would add that I wouldn't mind having say 20% of my code non-GPL'd so that one company can build their closed stuff on it.. Jul 01 10:32:18 .. rather than having them build 100% closed source on their own. Jul 01 10:33:17 <[florian]> the main problem with companies using gpl code, is that they think that taking a linux tarball and tuning it, makes it a proprietary code (to approximate) Jul 01 10:33:22 <[florian]> which is absolutely not the case Jul 01 10:33:49 <[florian]> and no one will complain about a proprietary user-space binary put on top of fully gpl'd firmware Jul 01 10:34:07 <[florian]> which perfectly makes sense, since you might want to add your own bricks Jul 01 10:34:21 <[florian]> maemo is a good exemple of mixing proprietary/gpl code I think Jul 01 10:35:17 But Mikrotik, on the other hand, is a counter-example of a violation.. Jul 01 10:35:28 They gave nothing back. Jul 01 10:45:42 <[florian]> xrg_: well, yes, they did not give their work back, they do release tarballs with the sourcecode of their firmware anyway Jul 01 10:46:53 <[florian]> xrg_: they could also assume it was too specific to be given back to the community Jul 01 10:47:24 <[florian]> xrg_: which is less the case with nokia's hildon and friends, according to the lot of spin-offs and initiatives talking about guis for embedded linux devices Jul 01 10:52:41 nokia is another case: they understand that giving away an API for 3rd party apps is to their benefit.. Jul 01 10:52:54 <[florian]> xrg_: right Jul 01 10:53:07 So, they don't really build embedded products, they rather build computers.. Jul 01 10:54:10 I've had a saying a few years ago: "Computer is an electronic device you can load Tetris on.." Jul 01 10:54:22 So, nokias are computers, routers aren't .. Jul 01 11:00:20 The first WIP-Day is about to reach it's end. It's been relatively quite, but this only makes it interesting to see what will happen at the next Wip-Day, which will be on the last saturday in July from UTC 12.00 and onwards Jul 01 11:01:25 I'll "stich" something together about what has taken place and post it on the dev-mailinglist + in the forum :) Jul 01 11:01:53 <[mbm]> ;) Jul 01 11:02:01 <[mbm]> hope it wasn't too disappointing Jul 01 11:02:05 And thanx for the support and just generally hanging around Jul 01 11:02:16 [mbm]: always gotta start somewhere Jul 01 11:14:30 olli * r7823 /packages/utils/scponly/Makefile: Fix scponly compilation (#1982) Jul 01 11:44:47 juhosg * r7824 /trunk/include/kernel.mk: reordered GENERIC_PATCH_DIR variable Jul 01 14:05:35 florian * r7825 /trunk/package/bcm43xx-mac80211/Makefile: Make bcm43xx-mac802xx bcm47xx-2.6 specific Jul 01 14:06:37 [florian]: why did you do that? Jul 01 14:12:15 because bcm43xx-mac802xx fails on x86-2.6 Jul 01 14:12:32 why? Jul 01 14:12:45 bcm43xx-mac80211 should be usable for more than just broadcom hardware Jul 01 14:12:46 nbd, wait a seconed i give you a log Jul 01 14:12:58 compiling fails... Jul 01 14:13:09 maybe some ssb changes need to be moved into generic-2.6/patches-2.6.22 Jul 01 14:13:15 noz? Jul 01 14:14:20 let me checkin the updated vim package first, than i'll give you the buillog for bcm43xx-mac802xx Jul 01 14:17:29 olli * r7826 /packages/utils/vim/ (3 files in 2 dirs): Update vim: 5.8 > 7.1 Jul 01 14:22:00 nbd, fbcm43xx--mac802xx buildlog: http://pastebin.ca/598633 Jul 01 14:22:36 yeah, it's missing ssb changes Jul 01 14:37:11 nbd, can you fix the missing ssb things Jul 01 14:47:47 <[florian]> nbd: try building bcm43xx-mac80211 on anything else that does not have ssb and you will see Jul 01 14:49:04 [florian], fixed scponly now Jul 01 14:49:09 <[florian]> forum2006: thanks Jul 01 14:49:26 no problem, btw. can you add cfdisk? Jul 01 14:49:31 <[florian]> forum2006: sure Jul 01 14:49:35 i have only write access to /packages Jul 01 14:49:36 [florian]: minipci cards on other platforms have ssb Jul 01 14:49:53 [florian]: ssb is part of the kernel in newer versions Jul 01 14:49:56 [florian], but i don't know how to handle the libncurses dependency Jul 01 14:50:00 on cfdisk Jul 01 14:50:06 <[florian]> forum2006: I will handle it Jul 01 14:50:12 [florian], thanks Jul 01 14:50:24 <[florian]> nbd: just it failed to build for adm5120, and it will for others too, and I was using 2.6.22-rc6 actually Jul 01 14:50:34 i know that it failsa Jul 01 14:51:02 fails on x86 too Jul 01 14:51:04 but that's just because some changes that were made in brcm47xx need to be moved Jul 01 14:51:28 <[florian]> nbd: ok Jul 01 14:51:50 <[florian]> nbd: that's mainly why I kept it bcm47xx specific for now Jul 01 14:57:03 <[florian]> nbd: I admit it's a bit rude Jul 01 15:02:22 olli * r7827 /packages/net/wifidog/Makefile: wifidog depends on libpthread Jul 01 17:43:29 florian * r7828 /trunk/package/util-linux/Makefile: Package cfdisk and bump release number (#1992) Jul 01 17:44:14 florian * r7829 /packages/net/openvpn/Makefile: Bump openvpn release number since we added an initscript to it Jul 01 17:45:50 noz * r7830 /trunk/package/bcm43xx-mac80211/src/bcm43xx/ (5 files): Update bcm43xx-mac80211 to latest git pull, mostly debugging improvements Jul 01 17:53:25 florian * r7831 /trunk/toolchain/Config.in: Use -Os for rdc and -O2 for the x86 target Jul 01 18:12:41 nbd * r7832 /tags/kamikaze_7.07/: branch 7.07 (copy of trunk) Jul 01 18:15:09 nbd * r7833 /tags/kamikaze_7.07/package/base-files/files/etc/ (banner ipkg.conf): update banner and ipkg.conf Jul 01 18:16:32 nbd * r7834 /tags/kamikaze_7.07/target/linux/ (13 files in 13 dirs): remove unsupported targets Jul 01 18:17:19 19:15 < CIA-6> nbd * r7834 /tags/kamikaze_7.07/target/linux/ (13 files in 13 dirs): remove unsupported targets Jul 01 18:17:26 shit sorry Jul 01 18:31:51 nbd * r7835 /tags/kamikaze_7.07/Makefile: put the version in the master makefile Jul 01 18:37:09 nbd * r7836 / (2 files in 2 dirs): nuke broadcom 2.6.21 patches - 2.6.22-rc6 is more stable and more up to date Jul 01 18:43:52 nbd * r7837 / (2 files in 2 dirs): bump busybox release number Jul 01 18:48:06 Kaloz: will you commit - https://dev.openwrt.org/ticket/1918 to 7.07 ? Jul 01 18:57:08 this change looks reasonably trivial. i will commit it Jul 01 18:57:40 nbd: He said no point as he would have the new ethernet driver in Jul 01 18:57:55 well, it's a bit too late to upgrade the kernel in 7.07 Jul 01 18:58:06 we want to release very soon Jul 01 18:58:20 thanks anyway - will it use 2.6.21 for ixp4xx ? Jul 01 18:58:31 i think so Jul 01 18:59:31 nbd: what about the inevitable bug reports for qos-scripts - i.e the l7-filters part (is that to be changed ?) Jul 01 19:02:20 nbd * r7838 / (4 files in 2 dirs): add gateway 7001 support patch from #1918 Jul 01 19:04:58 h3sp4wn: i'm sure there are some l7 patches for 2.6.21 out there Jul 01 19:05:05 h3sp4wn: i will try to merge them before the release Jul 01 19:05:09 2.6.22-rc will have to wait Jul 01 19:05:23 hi Jul 01 19:05:29 hi Jul 01 19:05:31 hey Jul 01 19:05:55 i have a problem setting up wds .. could you please help me Jul 01 19:06:15 what openwrt version, what platform? Jul 01 19:06:18 ok Jul 01 19:06:34 kamikaze 7.06 on wrap board with atheros minipci Jul 01 19:06:38 selfcompiled Jul 01 19:06:45 ok, what is the problem? Jul 01 19:07:18 i get the error: Error for wireless request "Set ESSID" (8B1A) Jul 01 19:07:35 show me the wireless config Jul 01 19:07:42 ok Jul 01 19:08:29 just a second please Jul 01 19:10:34 http://snofoo.de/ebay/wrap_config.txt Jul 01 19:12:58 http://snofoo.de/ebay/iwconfig.txt Jul 01 19:14:01 snofoo: remove the ssid from the wds wifi-iface Jul 01 19:15:25 you mean the line: option ssid wds? Jul 01 19:15:30 yep Jul 01 19:16:43 same error Jul 01 19:16:49 Error for wireless request "Set ESSID" (8B1A) Jul 01 19:16:54 SET failed on device ath1 ; Operation not supported. Jul 01 19:19:32 any other ideas? Jul 01 19:20:18 not yet. i will test it here and make sure that it will get fixed for the soon-to-be-released 7.07 Jul 01 19:20:30 great Jul 01 19:20:34 please make a ticket and assign it to the 7.07 milestone Jul 01 19:20:38 so that i don't forget it Jul 01 19:20:49 ok i'll do that Jul 01 19:20:52 thank you Jul 01 19:21:27 the strange thing is that from time to time a ping gets through the wds Jul 01 19:22:21 do i have to enable stp with brctl? Jul 01 19:22:54 and is it right that i dont have a wds.0 interface? Jul 01 19:29:54 yes, it's right Jul 01 19:29:58 the wds interface is ath1 Jul 01 19:31:43 btw. specifying ath0 and ath1 in the ifnames for lan is unnecessary Jul 01 19:31:52 that's what 'option network lan' in the wifi config does Jul 01 19:33:33 btw. one other thing you can try is making a second ap interface for wds on one node and a sta interface on the other and adding 'option wds 1' on both Jul 01 19:33:41 (replacing the wds mode interfaces) Jul 01 19:52:36 ok i'll give it a try Jul 01 19:52:45 thanks Jul 01 21:06:58 nbd: i tried sta mode but something isnt working right Jul 01 21:07:17 if i start: iwlist ath1 scanning Jul 01 21:08:07 i'm not seeing the ap of the other wrap board... just the wlans near me Jul 01 21:41:40 nbd * r7839 /tags/kamikaze_7.07/target/linux/ (adm5120-2.6/ amcc-2.6/ avr32-2.6/): 7.07 will add support for avr32, amcc and adm5120 targets Jul 01 21:45:14 crazy_imp * r7840 /packages/libs/apr/Makefile: update apr to 1.2.8 Jul 01 21:54:22 crazy_imp * r7841 /packages/libs/apr/Makefile: update apr to 1.2.8, sorry got the wrong file the last time Jul 01 21:58:26 crazy_imp * r7842 /packages/utils/apr-util/ (Makefile patches/): remove the unneeded patches Jul 01 22:03:04 ah, crazy_imp also got svn write access Jul 01 22:03:21 crazy_imp * r7843 /packages/net/subversion/Makefile: update subversion to 1.4.4, svnserver works now (problem fixed since 7804) Jul 01 22:03:55 forum2006: yupp, but only for my packages Jul 01 22:04:07 i see Jul 01 22:05:35 i had them waiting since yesterday, but a simple "ö" caused troubles :D Jul 01 22:05:49 lol Jul 01 22:06:14 nbd: never use them in a pw^^ Jul 01 22:06:32 since my keyboard layout is us_intl, i certainly won't Jul 01 22:06:58 i like neo as layout Jul 01 22:06:58 the german keyboard layout is unsuitable for any serious programming work Jul 01 22:07:16 neo? Jul 01 22:07:33 nbd: http://pebbles.schattenlauf.de/layout.php Jul 01 22:08:36 crazy_imp: that one is just as unusable for programming work as the regular german keyboard layout Jul 01 22:08:56 because all those useful characters lile / [] {} have been horribly misplaced Jul 01 22:09:23 nbd: i like it Jul 01 22:09:34 and putting < and > on the same key is very stupid Jul 01 22:09:49 so unintuitive Jul 01 22:10:38 nbd: <> are on f and j also on neo Jul 01 22:10:48 just take a look at http://www.eigenheimstrasse.de:8668/space/Computerecke/NEO-Tastaturlayout/Download/Grafik/NEO-Layout-1.1.png Jul 01 22:11:06 oh, right Jul 01 22:11:13 i just now noticed that they duplicated certain characters Jul 01 22:11:55 ok, it's a little bit better than the german one Jul 01 22:12:20 putting alt gr on capslock is a good idea, though Jul 01 22:12:22 indeed, and using the brackets "below" the normal chars is quite fast Jul 01 22:12:41 who needs capslocl these days anymore Jul 01 22:14:17 was there ever a need for one? (besides shouting :D) Jul 01 22:44:03 noz * r7844 /trunk/target/linux/ (14 files in 11 dirs): Move SSB out of brcm47xx target into generic-2.6 target Jul 01 22:58:05 <[florian]> crazy_imp: do not add crazy_imp instead of $ Id: $, rather put the right propset on those makefiles Jul 01 22:59:29 [florian]: how can i do that? (and i think i put $Id$ in there, like Kaloz told me :) Jul 01 23:00:17 <[florian]> crazy_imp: there is a special svn property to do that, which is svn:keywords Jul 01 23:00:51 <[florian]> doing this would be : svn propset svn:keywords "Author Date Id Revision" Jul 01 23:02:51 [florian]: ok, i'll try that the next time i have to change something :) Jul 01 23:04:04 <[florian]> crazy_imp: I am very likely to do a massive propseting ;) Jul 01 23:04:39 [florian]: good idea :) Jul 01 23:09:44 florian * r7845 /packages/ (60 files in 60 dirs): Massive propset of svn:keywords for packages newly addded, next time, we should not forget it ;) Jul 02 02:10:01 nbd * r7846 / (12 files in 4 dirs): upgrade layer7 to the latest version and add fixes for 2.6.21 and 2.6.22-rc - compile tested on .21, run time tested on .22-rc6 **** ENDING LOGGING AT Mon Jul 02 02:59:56 2007