**** BEGIN LOGGING AT Mon Nov 11 02:59:58 2013 Nov 11 10:19:16 morning all Nov 11 10:25:38 evening here - bye :-) Nov 11 10:27:12 morning Nov 11 10:39:35 Morning. Nov 11 10:39:46 bluelightning, BTW, I have your email, and I'll be looking at the issues you point out through this week. Nov 11 10:40:04 Stygia: ok, cool Nov 11 10:57:34 bluelightning, Sorry for being so slow with this whole thing, I'm too busy for my own good lately. Nov 11 10:57:37 Startups... Nov 11 10:57:58 Stygia: it's ok, we'll get there Nov 11 11:03:22 bluelightning, Yup, I know, but still, I sort of hope nobody's wasting time writing perl recipes when they already exist. Nov 11 11:03:45 I hope, at least, if someone comes here or on #oe asking about perl recipes, you'll refer them to me, maybe the recipes aren't 100% QA yet, but they work, mostly. Nov 11 11:04:01 Stygia: I have been, yes Nov 11 11:04:13 bluelightning, Good. Nov 11 11:04:28 bluelightning, I'd rather not waste other people's time if I can help it. Nov 11 13:39:39 Hey. Nov 11 13:40:02 Anyone know where I can find the setting for UTC? w.r.t time and timezones, on fedora it would be /etc/sysconfig/clock Nov 11 14:42:49 hi, could anyone help me with little misunderstanding? Nov 11 14:43:41 I don't really understand the meaning of LICENSE field, if in the end LIC_FILES_CHKSUM is used anyway? Nov 11 14:44:19 What would be the benefit of creating my own custom license, adding it LICENSE_PATH > Nov 11 14:44:23 What would be the benefit of creating my own custom license, adding it LICENSE_PATH ? Nov 11 14:57:07 chocis: LICENSE is a statement of what the licenses are Nov 11 14:57:16 chocis: the checksum is exactly that, a checksum Nov 11 14:57:41 if the license checksum changes you have to verify the licenses again Nov 11 14:57:57 rburton: it's so you can detect if upstream changes the license included in their source file, right? Nov 11 14:58:15 yes Nov 11 15:02:01 And you really never want to create a custom license, when starting a new project. There's always an existing one that will do, and which won't increase license proliferation/confusion :) Nov 11 15:02:05 * kergoth yawns Nov 11 15:05:39 We are already working for a while with our custom board and layers and then I saw the LICENSE_PATH variable, so I thought that it could be used for general license file, so I wouldn't need it in our software recipes. Nov 11 15:06:46 but I think I've got the meaning of the extra license folder. I can create a license for my company and refer to it in all our recipes (with our software). If later building images someone doesn't want to use our license he can exclude it in .conf file Nov 11 15:06:51 LICENSE_PATH is where canonical copies of the license text go for producing release manifests Nov 11 15:06:53 then there will e an error Nov 11 15:38:42 chocis, On a closed system / controlled by you? Nov 11 15:38:52 chocis, If it's closed you can just set the license to CLOSED Nov 11 15:39:29 chocis, On a FOSS recipe, you have LIC_FILES_CHKSUM, to ensure that the file has not changed (i.e. the license could have changed), and the LICENSE field to tell you which license you have - using the LIC_FILES_CHKSUM field to prove it hasn't changed since. Nov 11 16:17:43 Hmm, anyone know offhand if we can use lookahead and lookbehind assertions in MIRRORS/PREMIRRORS? I'm thinking it should be doable given its still just python re under teh hood Nov 11 16:17:46 * kergoth gets caffeine Nov 11 17:22:29 halstead: ping Nov 11 17:22:57 Hi scottrif. Good morning. Nov 11 17:24:31 halstead: morning...I suddenly have errors when trying to send an email out of my Mozilla Thunderbird account. It is saying "An error occurred sending mail: Unable to establish a secure link with SMTP server linux.intel.com using STARTTLS since it doesn't advertise that feature. Switch off STARTTLS for that server or contact your service provider" Any ideas? Nov 11 17:25:33 halstead: yes - I am connected Nov 11 17:25:53 halstead: I did try shutting down Thunderbird and restarting that Nov 11 17:26:34 halstead: ok Nov 11 19:43:37 howdy! i've got a sdk generated and the fitting qemu instance started. eclipses adt plugin is correctly setup, bu i kind of fail to find the way to use the tcf remote debugging. is this documented somewhere? Nov 11 19:45:34 I don't' know how much this'll help. Nov 11 19:45:34 https://lists.yoctoproject.org/pipermail/yocto/2011-March/003426.html Nov 11 19:45:55 https://www.yoctoproject.org/docs/current/adt-manual/adt-manual.html Nov 11 19:46:05 thanks Nov 11 19:46:27 i actually managed to progress a bit - i used the .launch file and set up a new remote connection for it Nov 11 19:47:23 look like thats the way Nov 11 19:48:33 always the same. you tinker around for a couple of hours, and when you finally bother someone, things fall into place. Nov 11 19:49:04 Stygia: but thanks a lot nevertheless! Nov 11 19:49:31 LetoThe2nd, No worries. Nov 11 19:49:36 LetoThe2nd, Good luck with it! Nov 11 20:04:33 Anyone here that deals with the buildbot implementation used with Yocto ? Nov 11 20:11:14 WarheadsSE: pidge, but she's not here Nov 11 20:12:38 K Nov 11 20:12:50 Looking at doing something inhouse for our bsp Nov 11 20:13:16 I've got one guy jumping about jenkins, see that Yocto itself is using Buildbot Nov 11 20:13:35 * WarheadsSE will hunt pidge down later Nov 11 20:13:53 WarheadsSE: she's in the UK right now, so just went home about half an hour ago Nov 11 20:14:07 k Nov 11 20:14:20 WarheadsSE: i've gone from checking it out to having a custom builder running this afternoon Nov 11 20:14:25 so i can endorse yocto-autobuilder Nov 11 20:14:55 We'd needa full stack in house, and I have no doubt that it's capable Nov 11 20:15:19 I've got a full cluster of master/slave/slave/ircd sitting in lxc atm Nov 11 20:16:34 but that is just the stock buildbot sources & configuration Nov 11 22:21:34 i can't seem to load Nov 11 22:22:37 sorry, pressed enter too early. i can't seem to load nameservers from /etc/network/interfaces. i have a dns-nameservers stanza, but nothing gets written to /etc/resolve.conf Nov 11 22:22:59 running ifdown/ifup manually after boot does nothing either Nov 11 22:23:30 any idea how to fix this? i'd rather not create resolv.conf manually, as i'm trying to keep all network configuration in one place Nov 11 22:25:25 BCMM: i'd suggest installing resolvconf, which manages resolv.conf for you, and then use its ifup hook, which provides dns-domain, dns-search, and dns-nameservers directives for interfaces Nov 11 22:25:34 BCMM_: i'd suggest installing resolvconf, which manages resolv.conf for you, and then use its ifup hook, which provides dns-domain, dns-search, and dns-nameservers directives for interfaces Nov 11 22:25:48 heh, sorry, was just about to move do another machine Nov 11 22:26:09 kergoth: so ifup doesn't do DNS settings on its own? that's a resolvconf feature? Nov 11 22:26:24 afaik stock interfaces / ifupdown doesn't provide it Nov 11 22:26:33 at least, the ubuntu man page on interfaces doens't mention anything about dns at all Nov 11 22:26:36 * kergoth shrugs Nov 11 22:26:47 its fundamentally a debian thing that we pull in Nov 11 22:27:07 kergoth: what is? the ifup/ifdown system? Nov 11 22:27:15 yep Nov 11 22:27:18 and thanks! i'd assumed all /etc/network/interfaces stuff was done directly by ifup Nov 11 22:27:21 the oe/yocto distros started as being debian based Nov 11 22:27:28 it is, but ifupdown is software that comes from debian ;) Nov 11 22:27:39 though i think busybox provides a mini implementation Nov 11 22:27:41 kergoth: i mean the resolvconf stuff Nov 11 22:27:43 not sure on that though Nov 11 22:28:01 ifupdown is extensible. software can add hooks to add support for new interfaces directives Nov 11 22:28:08 e.g. wpa-supplicant provides wireless config via interfaces Nov 11 22:28:21 see /etc/network/if-*.d/ Nov 11 22:29:15 * kergoth is getting rusty, spending too much time in buildsystems land and not enough dealing with the distros on the target Nov 11 22:29:30 kergoth: thanks, it's starting to make sense now Nov 11 22:29:39 i'm basically a debian noob you see... Nov 11 22:29:46 * kergoth nods Nov 11 22:29:47 np Nov 11 22:30:16 so basically, i should use resolvconf, and set up resolvconf to get its information from /etc/network/interfaces? Nov 11 22:30:36 well, its ifupdown that calls resolvconf, so more the other way around Nov 11 22:30:40 but basically yes :) Nov 11 22:30:53 resolvconf does provide the hook that ifupdown calls to run resolvconf, though ;) Nov 11 22:32:12 oh i think i get it, resolvconf includes a hook in ifupdown, such that ifup will send DNS information to resolvconf? Nov 11 22:32:13 resolvconf is nice because it lets you assemble resolv.conf from multiple pieces, so it lets things like dhcp clients add their dynamic pieces without blowing away user local customizations like search directives Nov 11 22:32:17 iirc, anyway Nov 11 22:32:18 yes, precisely Nov 11 22:32:43 when you call ifup, ifup calls all the if-up.d hooks with env vars set based on the content of /etc/network/interfaces, which then use that to do stuff Nov 11 22:32:45 and the resolvconf recipe will install that hook for me? i don't need to edit some config file somewhere? Nov 11 22:33:18 it should. if it doesn't, it's a bug in the recipe Nov 11 22:33:21 cat /etc/network/if-up.d/000resolvconf |grep /resolvconf Nov 11 22:33:21 [ -x /sbin/resolvconf ] || exit 0 Nov 11 22:33:21 echo -n "$R" | /sbin/resolvconf -a "${IFACE}.${ADDRFAM}" Nov 11 22:33:36 kergoth: thanks a lot! Nov 11 22:33:41 np Nov 11 22:33:56 basically install it, and add dns-nameservers to interfaces, and that *should* be sufficient, afaik Nov 11 22:33:59 in theory :) Nov 11 23:16:05 wb hollisb Nov 11 23:16:19 thanks mranostay :-) Nov 11 23:16:39 mranostay: enjoying the weather? Nov 11 23:28:11 hollisb: as long as it isn't raining i don't mind :) Nov 11 23:29:40 good Ingress conditions :-) Nov 11 23:31:03 i'm sure this state will the second longest one i've lived in :P Nov 11 23:32:04 hollisb: you live in PDX or in one of the burbs? Nov 11 23:33:43 kergoth: any idea why resolvconf pulls in bash? it seems to be written in #!/bin/sh Nov 11 23:36:27 not offhand, no, sorry to say Nov 11 23:36:41 that does sound questionable, unless it has implicit bash dependnecies, in which case its schebang should be fixed Nov 11 23:36:55 i'd say try getting rid of it and see if it explodes :) Nov 11 23:37:07 alternatively, check the git history for rthe original reasoning Nov 11 23:37:12 see e.g. git blame for the line that adds the dep Nov 11 23:39:47 kergoth: ah, found a debian bug saying that it's Nov 11 23:40:23 sorry, pressed enter... /etc/resolvconf/update.d/bind rather than the main script, and a long arguement as to whether it should use bashisms... http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=555735 Nov 11 23:42:38 and that file doesn't even seem to be installed by the recipe... so tomorrow i'll do the "see if it explodes" option Nov 11 23:43:14 good luck :) Nov 11 23:43:29 I'm sure folks woudl appreciate one less recipe sucking in bash, so that'd be good to go upstream Nov 11 23:44:26 gah, i still have to learn some git to work out how to actually upstream stuff... Nov 11 23:44:46 (i already have some modifications for the mpd recipe which probably belong upstream) Nov 11 23:48:40 um, how do i get the commit message out of git blame? Nov 11 23:48:57 git show Nov 11 23:49:03 git blame lists the commits in the left column Nov 11 23:49:19 see also http://www.git-scm.com/book Nov 11 23:50:08 kergoth: git blame is my favorite command :) Nov 11 23:52:01 kergoth: thanks, unfortunately it's "RDEPENDS, RRECOMMENDS -> RDPEPENDS_${PN}, RRECOMMENDS_${PN}"; i.e. a mass change of syntax in loads of files, and now i need to know why the line that *used* to be there was put in... Nov 11 23:52:43 before i mess with the recipe i'm going to manually delete /bin/bash and see what breaks... Nov 11 23:53:24 thing is, if nothing does explode, i won't know if that is jsut cause i don't use the feature that should have exploded... **** ENDING LOGGING AT Tue Nov 12 02:59:58 2013