**** BEGIN LOGGING AT Thu Nov 12 02:59:58 2015 Nov 12 08:13:00 good morning Nov 12 08:13:39 good morning Nov 12 09:15:10 mornin Nov 12 10:07:17 notcing some spam on the oe wiki - how should I report that? Nov 12 10:24:48 joshuagl: thanks, I think I got it all now Nov 12 10:26:04 bluelightning: np Nov 12 11:33:57 why is oe still trying to patch even though I added inherit externalsrc Nov 12 11:36:52 it shouldn't be Nov 12 11:37:05 is SRCTREECOVEREDTASKS perhaps being overridden such that do_patch is not included? Nov 12 11:37:36 also, remind me which version of the build system are you using? and what kind of recipe is this - is it a kernel recipe? Nov 12 11:37:53 I have to step away - back later Nov 12 14:19:21 fray, https://access.redhat.com/security/cve/#/ Nov 12 14:19:38 or on on yocto-security, can you make sure they know about this? Nov 12 17:16:41 bluelightning: http://lwn.net/Articles/664073/ Nov 12 17:16:57 "Developers on Linux, Windows, or Mac systems can build packages for their Yocto-based embedded Linux targets and push them to the devices (or to the corresponding virtual-machine image)" Nov 12 17:17:13 does the extensible SDK works on Windows? Nov 12 17:17:34 it doesn't, no Nov 12 17:17:53 I need to reply to that (or send them a correction, I'm not sure) Nov 12 17:18:17 ah ok, yeah, that's what I thought Nov 12 17:18:20 linux vm would be the best bet for actually building anything on windows. i think the regular sdk would work there, but that would be just for application development, not building packages.. Nov 12 17:19:03 most of our clients are Windows-based, so what we give them is a toolchain + eclipse Nov 12 17:19:33 so reading the article I got excited, at least for a short while Nov 12 17:27:04 does anyone uses travis for ci in public layers? we are moving a bunch of our custom test to ptests, wondering if it's doable to use travis-ci to check each submission by running ptests Nov 12 17:28:05 sounds interesting, though potentially high overhead. not sure if anyone is using it, maybe someone will speak up Nov 12 17:28:15 i wonder if you could search for the travis ci config file in layers :) Nov 12 17:29:10 hehe, yeah Nov 12 17:36:16 Link that all can use to the article Nov 12 17:36:17 http://lwn.net/SubscriberLink/664073/1bd82d3892d11d4b/ Nov 12 17:53:54 naming is quite a confusing thing in yocto/OE/poky. Nov 12 17:55:39 And to make the situation even more confusing, there are still _branch_ names, which I personally find rather pointless, as they don't spacify any order, like numbers do. Nov 12 17:56:03 specify* Nov 12 17:57:31 plenty of projects have release codenames though... Nov 12 17:58:44 I personally don't see the point and really don't like that. It makes a useless indirection, specially when you are searching things on the internet. You never know if you should search for the codename or for the version number. Nov 12 17:59:00 searching for numbers doesn't work at all though Nov 12 17:59:32 I do understand people have a hard time remembering the version <-> name mapping, and that is a problem Nov 12 17:59:33 Seems to work for me. Nov 12 17:59:55 heck, sometimes I even have to look it up Nov 12 17:59:59 Another problem is that as time passes by, you don't know which codename is older. Nov 12 18:00:26 https://wiki.yoctoproject.org/wiki/Releases is what I usually refer to Nov 12 18:01:22 Yes. You have to know that the page exists, find it, then read it just to find the information which is not really useful for anything. Nov 12 18:01:33 I have the exact same problem with Debian, for example. Nov 12 18:01:44 right, as I said it's not special to us Nov 12 18:02:00 there are things we could do even without throwing out the names Nov 12 18:02:26 in theory we could use the versions in the branch names as well Nov 12 18:02:33 Yeah, like bitbake does. Nov 12 18:02:49 It's very clear and indirection-free. Nov 12 18:03:08 You know that bitbake 1.28 is newer than 1.26. Nov 12 18:03:25 But is dora newer than danzil? Nov 12 18:03:29 yes Nov 12 18:03:34 * rburton runs Nov 12 18:03:37 :-D Nov 12 18:04:20 the names all have numbers too, so you could start using numbers only to encourage others to do the same (or use both) Nov 12 18:04:48 Why using the name? Nov 12 18:05:11 why not? Nov 12 18:05:25 at least jethro isn't Beefy Miracle Nov 12 18:05:28 because it creates a pointless indirection. Nov 12 18:05:37 It's not useful at all. Nov 12 18:05:49 only creates indirection if that's how you primarily refer to the thing Nov 12 18:05:57 so use the number Nov 12 18:06:07 the latest release is version 2.0, codename jethro Nov 12 18:06:20 both uniquely identify the release Nov 12 18:06:23 How about creating a nickname for the codename then? Nov 12 18:06:26 there is one place where it does help - versioning across multiple layers Nov 12 18:06:31 So we have 3. Nov 12 18:06:47 e.g. if we just had "2.0" over in some distro - is that that distro's 2.0 version? or ours? Nov 12 18:07:00 if it says "jethro" it's a bit more obvious Nov 12 18:09:35 the fact is that for a while, searching for 2.6 with Google would correctly find results for Linux :) Nov 12 18:11:42 abelloni: well, google does tailor its results somewhat to the person connecting, so that may work for you but not someone else :) Nov 12 18:11:46 When I read "searching for numbers doesn't work at all though" I supposed one would look for the project name _and_ the version number. Obviously, searching for numbers only is not the smartest idea. Nov 12 18:11:55 (maybe in the 2.6 timeframe it wasn't doing that tailoring) Nov 12 18:12:16 right, point taken Nov 12 18:12:29 Well, I just wanted to let you know that there is at least one person that finds that the codename is pointless and only contributes to the confusion around names. Nov 12 18:13:03 understood, and I know you're not the only one Nov 12 18:18:39 I'm a fan of just using the numbers for tagging/branching too for what its worth, but it's a bit late for that now, unless we want to break any documentation already out there Nov 12 18:18:44 heh Nov 12 18:20:59 I think the worst was during the daisy/dizzy timeframe Nov 12 18:23:53 For a while I had the impression that releases must start with "d", for some reason (maybe to honor the global d variable?) Nov 12 18:24:53 mario-goulart: https://en.wikipedia.org/wiki/Dizzy_(series) Nov 12 18:25:30 oh! Nov 12 18:26:49 What's wrong with dozy? :-) Nov 12 18:27:24 I remember asking about it at ELCE 2014, I think belen1 was the one telling me to look for the names in wikipedia Nov 12 18:27:42 And why not Yolkto? :-) Nov 12 21:07:47 #microchip Nov 12 22:37:08 #espressomachine Nov 12 22:43:21 #cricketbat Nov 12 23:50:10 Why is it I always miss something in my patches and not spot it until it's in my inbox? I reviewed this carefully locally with git already, yet somehow miss things until they hit the list, it never fails Nov 12 23:50:13 often something trivial Nov 12 23:53:05 oh well, at least these are submitted, need to whittle down meta-mentor-staging, it has way too much crap in it Nov 12 23:58:30 next.. Nov 12 23:58:45 hmm, some of this does require more thought / rework to be suitable **** ENDING LOGGING AT Fri Nov 13 02:59:59 2015