**** BEGIN LOGGING AT Sat Jan 16 02:59:56 2010 Jan 16 03:13:58 cshore * r19160 /trunk/target/linux/brcm63xx/files/arch/mips/bcm63xx/boards/board_bcm963xx.c: Fix typo in AGPF_S0 reset button definition Jan 16 03:27:56 cshore * r19161 /packages/libs/gconv-modules/ (. Makefile): Added package gconv-modules, which is glibc-specific and need by many packages in that case. Jan 16 03:42:25 damn...revert didn't work Jan 16 03:42:47 How do I erase the last commit? Jan 16 03:43:15 I thought it was svn revert path/to/file Jan 16 03:43:50 no, that's only possible if you didnt commit Jan 16 03:44:20 oh, so there's not way to erase? Ok, how do I change the log message? Jan 16 03:44:35 I think you can "svn merge -r head:head-1 svn://..." Jan 16 03:44:43 cshore * r19162 /packages/libs/gconv-modules/Makefile: Revert commit; really gconv-modules shouldn't be in base and installed by default Jan 16 03:44:56 and changing the log isn't possible afaik Jan 16 03:45:29 ok, well, I'll have to keep that in mind then...dumb newbie error in the first place Jan 16 03:45:46 (unless you svndump and then reimport without the commits you didn't want) Jan 16 03:45:53 ick Jan 16 03:46:38 well the log message is almost correct...if I could get rid of revert commit it'd be right, if a little noisy (two commits for what should have been one) Jan 16 03:46:44 bah! Jan 16 03:46:50 but hey, it's svn, and errare humanum est, I wouldnt worry about it too much Jan 16 03:47:40 but I'm supposed to perfect, hadn't you heard ? !! :-/ Jan 16 03:47:59 I only heard that the bird is the word :P Jan 16 03:48:08 heh Jan 16 04:53:16 cshore: still need the relocate info? Jan 16 05:07:39 nope, figured it out, thanks Jan 16 07:04:57 swalker * r19163 /packages/utils/dosfstools/Makefile: [packages] dosfstools: update to 3.0.7 Jan 16 07:10:46 swalker * r19164 /packages/net/httping/ (Makefile patches/002-version.patch): [packages] httping: update to 1.4.1 Jan 16 07:31:50 cshore * r19165 /packages/lang/luasocket/Makefile: packages/luasocket: Added missing .lua files for socket functionality used by Prosody IM. Jan 16 07:59:13 can we push this on trunk: https://dev.openwrt.org/ticket/6484 Jan 16 08:19:47 * swalker wonders how much of openswan's init script is actually relevant when running on openwrt Jan 16 08:59:25 is there a way to set the ownership of a directory on firstboot? Jan 16 09:10:54 anyone has a clue for this: http://openwrt.pastebin.com/d7c1bb50d Jan 16 09:45:19 acoul: what is the Ticket #6484 (new enhancement) doing ??? Jan 16 10:08:34 adds 2-6.33 support Jan 16 13:36:29 jow * r19166 /trunk/target/linux/orion/ (5 files in 3 dirs): [orion] add sysupgrade support for the Netgear WNR854T and Linksys WRT350N v2 - thanks Matthias Buecher Jan 16 13:39:16 jow * r19167 /trunk/package/base-files/ (Makefile files/lib/upgrade/common.sh): Jan 16 13:39:16 [package] base-files: Jan 16 13:39:16 sysupgade library coomon.sh has "bug" in glibc enviorement. For some reason Jan 16 13:39:16 ld-linux.so.3 is not installed at ramfs and sysupgrade always fails. Jan 16 13:39:16 Signed-off-by: Zintis Petersons Jan 16 13:47:34 jow * r19168 /packages/libs/openldap/Makefile: [package] openldap: add slapd.conf to the package and mark it as config file Jan 16 13:51:59 jow * r19169 /packages/libs/openldap/Makefile: [packages] openldap: install an empty ldif directory (#6509) Jan 16 13:55:30 jow * r19170 /packages/net/krb5/ (Makefile patches/003-krb5kdc-com-err-fix.patch): [packages] krb5: ensure linking against supplied libcom_err (#6508) Jan 16 15:11:53 jow * r19171 /trunk/target/linux/brcm-2.4/ (4 files in 4 dirs): [brcm-2.4] fix serial flash support (#6442) Jan 16 15:45:14 jow * r19172 /trunk/package/broadcom-sdhc/ (. Makefile): [package] add broadcom-sdhc - successor of broadcom-mmc (#6343) Jan 16 15:46:45 jow * r19173 /trunk/package/broadcom-mmc/: [package] remove broadcom-mmc, superseded by broadcom-sdhc Jan 16 19:24:03 ping xMff Jan 16 19:24:28 pong cshore Jan 16 19:24:49 hi. Just wondering if you had a chance to look at the preinit stuff Jan 16 19:25:20 I skimmed through it, quite impressive also the documentation Jan 16 19:25:52 I wondered why you split up procedure declarations and their calling into multiple files in some cases Jan 16 19:26:12 e.g. for prcs that are called just once# Jan 16 19:26:31 usually it was because of the order the hooks should be added ... the idea is have drop in packages that insert new things in between Jan 16 19:26:57 I see, to override procedures? Jan 16 19:26:59 yes Jan 16 19:27:04 and to add new ones Jan 16 19:28:16 so that preinit doesn't have to be modified, e.g. for usb root or provider provisioning failsafe...just drop in a package and it's there Jan 16 19:28:25 I haven't gotten around to actually test it yet, guess I'll build a x86/vm image Jan 16 19:29:34 ok, cool. I did notice one mistake so far... I reversed the logic in the firstboot called as an executable with no arguments...e.g. to reset the jffs...I had the case of no mini_fo doing what with mini_fo should do and vice versa Jan 16 19:29:50 I'm working on that, and updated docs right now Jan 16 19:29:54 that's for firstboot Jan 16 19:30:19 is there a complete, non-incremental and up-to-date patch? Jan 16 19:30:39 the one that I last posted has everything, except that I need to fix that bug Jan 16 19:30:50 okay Jan 16 19:31:10 I noticed it when I was doing the docs Jan 16 19:31:30 do you write them by hand? Jan 16 19:31:56 yes, text file in Mararku using webgen to generate the web page Jan 16 19:32:24 Maruku I mean (PHP Markdown Extra) Jan 16 19:32:38 maybe you should consider inline documentation Jan 16 19:33:03 e.g. function prototypes in the code and then generate the docs from that Jan 16 19:33:23 well I'm not sure how to do that for shell scripts Jan 16 19:35:24 And the general discussion and such isn't really suitable for that, I don't think Jan 16 19:35:47 yes Jan 16 19:36:38 yes, which? Jan 16 19:37:27 It's really not to bad...mostly the only think autogen would save is typing some function names Jan 16 19:37:27 yes, it is obviously not suitable for general discussion of the mechanism, it could just help to build the file + procedure + description tables Jan 16 19:37:52 hmmm...I see Jan 16 19:37:58 and it is only a suggestion, no requirement :) Jan 16 19:38:05 what tools could I use for that, though? Jan 16 19:38:45 Certainly having the docs inline would be good for code documentation Jan 16 19:41:59 depends, the most tools I know are made for C-like syntax Jan 16 19:42:20 there are a few other choices like http://www.xs4all.nl/~rfsber/Robo/ Jan 16 19:42:47 but in the end it could be a simple script which extracts comments you marked specifically, like "##" or "#?" Jan 16 19:44:25 this is the way all other tools work like, you have the usual langauge-dependent comment syntax and some tags to markup comments Jan 16 19:44:44 right Jan 16 19:45:01 like Jan 16 19:45:01 # @function: foo Jan 16 19:45:01 # @param: $bar Jan 16 19:45:01 # @return: 1 or 0 Jan 16 19:45:01 # @desc: does foo on $bar Jan 16 19:45:49 they try to minimize the work by parsing the code and try to figure out stuff automatically Jan 16 19:45:57 right Jan 16 19:46:12 Doxygen is another popular solution Jan 16 19:47:03 I've used that before...but for C/C++ Jan 16 19:49:21 apparently it does not suport shell scripts natively but needs some kind of input filter Jan 16 19:50:13 don't put too much work into it, just consider it if you have some spare time Jan 16 19:55:28 I'll think about it, but trying google isn't turning up a lot of shell script tools for it Jan 16 19:57:25 oh, but robodoc is in Ubuntu...hmmm... Jan 16 19:58:27 the question is which way is more work Jan 16 19:59:20 right Jan 16 20:00:45 Hmmm...just pulling out a file/function name and comment would probably be easiest Jan 16 20:01:06 then the comment is the script so it's not duplicate Jan 16 20:01:12 more than that, I think is overkill for this Jan 16 20:01:40 yep Jan 16 20:02:09 if you stick to a comment block next to the procedure declaration it should be easy Jan 16 20:02:16 yes Jan 16 20:11:23 I think I might write the script in Perl though...I'm not sure how to do the same thing in a pure shell script...I suppose I could parse each newline seperated line Jan 16 20:11:47 in a for or whlie loop Jan 16 20:12:21 I'd prefer awk and sed though...but I don't know them well enough Jan 16 20:14:38 * xMff loves perl Jan 16 20:39:17 * cshore so does cshore Jan 16 20:40:02 * cshore means, as do I :-) Jan 16 22:18:58 nbd * r19174 /trunk/include/verbose.mk: make sure that processed directories are printed on make V=99 Jan 16 22:48:05 nbd: ping Jan 16 23:06:42 rtz: pong Jan 16 23:10:56 nbd: I was pretty busy the last 2 weeks, but I'm back now Jan 16 23:11:11 nbd: did you take a look at the phy stuff I send you? Jan 16 23:18:36 i don't remember receiving it Jan 16 23:18:53 hmmmm Jan 16 23:20:17 looks like I forgot to send it, sorry about that Jan 16 23:20:26 nbd: what was your mailadress again? Jan 16 23:21:18 nbd@nbd.name Jan 16 23:23:22 nbd: ok, i did send it to you Jan 16 23:23:48 there is lots of stuff missing, but the basics are there Jan 16 23:27:53 could you please send it to me as a patch? Jan 16 23:27:56 makes it easier to review Jan 16 23:29:50 nbd: I could, but besides in the r6040 driver, there is basically nothing of the original code still there Jan 16 23:30:02 and those bits are shuffeld around Jan 16 23:30:08 hm Jan 16 23:30:13 that'll make it hard to get it accepted upstream Jan 16 23:30:17 quite hard Jan 16 23:32:32 nbd: I tried in various ways, but i didn't find a good solution to add switch support to the current libphy :/ Jan 16 23:32:47 besides, it does some things in a rather complicated way Jan 16 23:50:45 would it be possible to get it into mainline, if it's only targeted at switchs for now and have it coexsist with the current libphy? Jan 16 23:51:23 probably not Jan 16 23:51:32 since ethernet drivers can only target one of the two Jan 16 23:51:48 some ethernet devices might be connected to either a phy or a switch Jan 16 23:54:03 no, I mean, the current drivers which are happy with the current libphy can continue to use it Jan 16 23:54:23 nbd: afaik, there are other more or less duplicated subsystems in the kernel Jan 16 23:55:42 and only use the new stuff for drivers which need switch support Jan 17 00:00:19 but the new ethernet drivers can't make use of the old phy drivers anymore then Jan 17 00:00:38 that creates a nasty compatibility mess Jan 17 00:09:07 well, there are only a handfull phydrivers mainline Jan 17 00:09:20 nbd: I'm really not sure, how else to do this Jan 17 00:09:49 getting switch support into the current libphy in a coherent fashion will be very difficult Jan 17 00:09:57 i'll take a look at your stuff when i have time Jan 17 00:10:42 nbd: this would be nice Jan 17 00:10:57 nbd: any idea when you may find the time? Jan 17 00:17:33 nope, sorry Jan 17 00:17:39 got a lot of stuff on my list at the moment Jan 17 00:18:21 ok Jan 17 00:18:56 I also have a lot of stuff to do, so I don't know when I would be able to work on it anyway Jan 17 00:19:58 nbd: maybe I can prepare some docs and send them to you, but not much beyond that Jan 17 01:07:19 nbd * r19175 /trunk/target/linux/ar71xx/patches-2.6.30/ (2 files): ar71xx: the lack of DQ* flash status checking for cfi cmdset 0002 does not only apply to erase commands, it also affects writes. extend the workaround to cover writes commands as well Jan 17 01:38:09 nbd * r19176 /trunk/package/mac80211/ (17 files in 2 dirs): mac80211: upgrade to compat-wireless 2010-01-15 Jan 17 01:52:47 nbd * r19177 /trunk/package/mac80211/files/lib/wifi/mac80211.sh: mac80211: fix mac address handling for multiple phy/vif Jan 17 01:53:19 nbd: oi Jan 17 01:53:19 nbd: oi Jan 17 01:53:20 nbd: oi Jan 17 01:53:25 wtf? Jan 17 01:53:30 main channel needs a ban plox Jan 17 01:55:40 thanks and sorry for the highlights Jan 17 01:55:49 no problem Jan 17 02:01:26 rtz: https://dev.openwrt.org/ticket/6492 Jan 17 02:09:29 AlexanderS: I'm already working on it, but it needs some groundwork, and I'm not sure, how this will look like Jan 17 02:09:55 AlexanderS: so don't hold your breath for a solution :( Jan 17 02:10:13 AlexanderS: another question, is usb working for you? Jan 17 02:17:20 AlexanderS: by the way, I have a really ugly fix for the switch problem, if it's urgent **** ENDING LOGGING AT Sun Jan 17 02:59:56 2010