<div>CORRECTED: This meeting took place on March 17, not March 10<br><br>Attendance: Mark Hatle, Richard Purdie, Khem Raj, Tom Rini, Stefan Schmidt, Koen Kooi, <br>Secretary: Jeff Osier-Mixon (non-voting)<br>_____________________________________________________________________________</div>
<div>AGENDA WITH DISCUSSION RESULTS</div><div>
<br></div><div> 01) Agree on meeting chair</div><div> </div><div>Koen this time, Khem volunteered next time</div><div>___________________________________</div><div> 02) Status report on oe-core</div><div><br></div><div>Mark: changes going in, still looking for rpm build issue reported by Khem & Denys, yocto bug 890</div>
<div>Khem is able to use oe-core with angstrom, has blogged on setup and also created wiki page on oe->oe-core sync</div><div>Mark: yocto master syncing regularly</div><div>(1:12:39 PM) koen: <a href="http://wiki.openembedded.org/index.php/List_of_OE_Features_For_oe-core" target="_blank">http://wiki.openembedded.org/index.php/List_of_OE_Features_For_oe-core</a></div>
<div>(1:12:51 PM) koen: <a href="http://sakrah.homelinux.org/blog/2011/03/using-openembedded-core-to-build-angstrom-for-qemu/" target="_blank">http://sakrah.homelinux.org/blog/2011/03/using-openembedded-core-to-build-angstrom-for-qemu/</a></div>
<div>Koen will poke Khem for gcc update re meta-toolchain issue</div><div>outstanding issues:</div><div> 1-bitbake sync</div><div> 2-s/yocto/oe-core/ and s/poky/??/ (Khem thinks ?? = oe)</div><div>take discussion on that to the ml</div>
<div>AI: Jefro volunteers to wikify checklist if someone sends</div><div>AI: Koen to send email gathering input for checklist (sent before meeting ended)</div><div>Mark: we've still got things we have to do before oe-core is "good enough" to present as 'done' IMHO</div>
<div> not gotten the branding and cleanup which I think we have to do before it's 'good enough'</div><div>AI: discussion for branding & recipe split to mailing list, Richard responded to Koen's post already</div>
<div>
<br></div><div>___________________________________</div><div> 03) Status report on pull model, contrib repo and guidelines</div><div> </div><div>Mark hopes to post commit/patch policy guidelines tomorrow, has feedback to incorporate</div>
<div>contrib repo now visible, pull model seems to be working</div><div> devs free to use own public repos</div><div> nice to do, pull req script to figure out url dynamically</div><div>Koen would like to copy oe-core guidelines as much as possible for meta-oe layer</div>
<div><br></div><div>___________________________________</div><div> 04) Status report on board support layer guidelines</div><div><br></div><div>no feedback from Richard on yocto side yet, agreed to use yocto guidelines minus scary linux-yocto requirement</div>
<div>need to find out lf position on bsp postings, feed to oe board</div><div> non-LF memeber postings</div><div>AI: continued, Richard to report back, send to board</div><div>AI: Richard has action to write proposal for YP steering group, won't have answer until their meeting</div>
<div>___________________________________</div><div> 05) Status report on version retention policy and interaction with</div><div> oe-core / meta-oe / $distro layers</div><div><br></div><div>Tom posted to ml, Koen likes doc from Mark & Tom</div>
<div> +/- ready to wikify</div><div>Koen doesn't like micromanaging & docing all use cases</div><div>TSC needs to define what is in what category</div><div>AI: Tom will wiki it</div><div>___________________________________</div>
<div> 06) Status on layer splitting of metadata</div><div><br></div><div>Khem using meta-oe layer, need to keep working on it - kitchen sink?</div><div>Koen - plan is to populate initially & evaluate</div><div>make pull strategy offiical by copying oe-core guidelines</div>
<div>Koen volunteers to be pullmaster</div><div>Richard reasonable to have clear areas of ownership, doc'd in meta-oe</div><div>must plan for higher activity in future</div><div>Koen: split recipes ok, but layers inside meta-oe not necessary yet</div>
<div>splitting criteria?</div><div> 1. willing maintainers (koen) following established quality guidelines (tom)</div><div> 2. systems/related functionality (fray)</div><div> 3. layer non-interdependence (koen)</div><div>
split into layers now? Koen says little experience with layers in oe community & with maintainers</div><div>Richard if structure could accomodate now, better for future</div><div>Khem - arrange like oe-core, meta and meta-xxx structure?</div>
<div>AI: Koen will move data in repo down a level into meta-oe now</div><div>___________________________________</div><div> 07) DISTRO_FEATURE/MACHINE_FEATURE/libc features and other</div><div> "flags" discussion</div>
<div> [ Proposed: Tom Rini ]</div><div><br></div><div>Tom: agree on problem scope & move to ml</div><div>need OE to meet needs of distros that maintain a feed</div><div>as well as useers that care more about initial build</div>
<div>important that feed people don't waste time debugging things</div><div>that nonfeed people introduce</div><div>fray: documenting flags as part of feed is necessary</div><div>signature code is starting point for both use cases</div>
<div>Mark is also concerned about namespace pollution btw flags</div><div>Richard - this is also on post-1.0 yocto list</div><div>AI: Tom to move to ml</div><div>___________________________________</div><div> 07) Continue discussion on posting of policies and guidelines</div>
<div><br></div><div>Mark: where to post and when?</div><div>AI: Khem to highlight the oe-core ml on oe-dev, tell poeple about discussion</div><div>suggest wiki for text, ml for discussion</div><div><br></div><div>___________________________________</div>
<div> 08) Continue discussions on infrastructure items</div><div><br></div><div>issues were to be raised on ml, no one did</div><div>Koen: admins sent doc requesting hw/funding, kick to board & LF?</div><div>Richard did pass it on to LF and is discussing hw options</div>
<div>Tom King made a draft on what is needed for future to replace oestats</div><div>Khem: autobuilder or tinderbox functionality still missing</div><div>oe infrastruc on machines shared with 4 other projects</div><div>tom requesting all 5 projects find funding</div>
<div><br></div><div>___________________________________</div><div> 09) How are we going to deal with the parts of OE that are not part of OE-core. (Frans)</div><div> </div><div> discussed in earlier line items</div><div>
<br>
</div><div>_____________________________________________________________________________</div><div>TRANSCRIPT</div><div><br></div><div><div>(12:59:13 PM) Tartarus: Just about now'ish right?</div><div>(12:59:59 PM) Jefro: 1 minute</div>
<div>(1:00:01 PM) Jefro: Khem sends apologies, so we are just waiting for tom and stefan</div><div>(1:00:08 PM) Jefro: oop, hi Tom :)</div><div>(1:01:10 PM) Jefro: until stefan arrives, I suggest starting on 01) Agree on meeting chair</div>
<div>(1:01:37 PM) Tartarus: First can we get the "final" agenda posted/pastebined?</div><div>(1:01:44 PM) Jefro: 1 second</div><div>(1:02:45 PM) Jefro: agenda is at <a href="http://pastebin.com/vjaiK3K3" target="_blank">http://pastebin.com/vjaiK3K3</a></div>
<div>(1:03:42 PM) fray: works for me</div><div>(1:03:50 PM) fray: before we begin we need someone to log</div><div>(1:04:14 PM) Jefro: I will log, if that is ok with everyone</div><div>(1:04:20 PM) khem [~<a href="mailto:khem@99-57-141-118.lightspeed.sntcca.sbcglobal.net" target="_blank">khem@99-57-141-118.lightspeed.sntcca.sbcglobal.net</a>] entered the room.</div>
<div>(1:04:20 PM) mode (+v khem) by ChanServ</div><div>(1:04:28 PM) Tartarus: Works for me</div><div>(1:04:37 PM) Jefro: alert: khem no longer sends apologies and has arrived on time</div><div>(1:05:01 PM) khem: crazy work</div>
<div>(1:05:45 PM) Jefro: khem: agenda is at <a href="http://pastebin.com/vjaiK3K3" target="_blank">http://pastebin.com/vjaiK3K3</a></div><div>(1:06:26 PM) Tartarus: So, I'd suggest koen or khem as everyone else has had a turn as the chair so far</div>
<div>(1:06:40 PM) koen: I'll chair</div><div>(1:06:51 PM) Jefro: 01) completed</div><div>(1:06:56 PM) koen: let's wait 4 minutes for stefan and then begin</div><div>(1:07:12 PM) koen: Jefro: double check: you'll be doing the minutes and logging, right?</div>
<div>(1:07:20 PM) fray: I thought Stefan said last week he wouldn't make it..</div><div>(1:07:26 PM) Jefro: koen - yes, I have volunteered</div><div>(1:07:30 PM) ***fray double checks</div><div>(1:08:28 PM) koen: fray: I don't really recall, so let's wait till ten past the hour :)</div>
<div>(1:08:34 PM) khem: next week I will chair so 01 for next week is done too</div><div>(1:08:39 PM) fray: can't find it</div><div>(1:09:21 PM) Jefro: with 8 more agenda items, I suggest continuing</div><div>(1:09:26 PM) koen: should the chairman send out an agenda for next week, or can we talk Jefro into doin gthat?</div>
<div>(1:09:45 PM) khem: chairman can do it</div><div>(1:10:06 PM) koen: OK, let's get started with 02) Status report on oe-core</div><div>(1:10:58
PM) fray: Changes are still going in... but I've not had a lot of time
to work with it.. I'm looking for the RPM build issue reported by both
Khem and Denys.. bug 890 in the yocto bug system</div>
<div>(1:11:08 PM) khem: lets see. I am finally able to use OE-core effectively with angstrom</div><div>(1:11:34 PM) khem: I have done a small weblog on how to get setup</div><div>(1:11:35 PM) fray: Yocto (master) is syncing regularly..</div>
<div>(1:11:53 PM) khem: and also created a page on wiki for outstanding OE -> oe-core syncs</div><div>(1:12:00 PM) khem: I dont think anyone else filled in anything</div><div>(1:12:13
PM) koen: I only have a problem with meta-toolchain on the buildhost at
work, but I'll poke khem to update gcc in meta-oe to see if that will
fix it</div>
<div>(1:12:16 PM) Tartarus: Where is your blog post and that page?</div><div>(1:12:30 PM) khem: on the ml</div><div>(1:12:32 PM) khem: oe-devel</div><div>(1:12:39 PM) koen: <a href="http://wiki.openembedded.org/index.php/List_of_OE_Features_For_oe-core" target="_blank">http://wiki.openembedded.org/index.php/List_of_OE_Features_For_oe-core</a></div>
<div>(1:12:51 PM) koen: <a href="http://sakrah.homelinux.org/blog/2011/03/using-openembedded-core-to-build-angstrom-for-qemu/" target="_blank">http://sakrah.homelinux.org/blog/2011/03/using-openembedded-core-to-build-angstrom-for-qemu/</a></div>
<div>(1:13:22 PM) koen: From the top of my head, outstanding items:</div><div>(1:13:23 PM) khem: we can collect stuff in the wiki and let people chose on what they would chose to do</div><div>(1:13:28 PM) koen: 1) bitbake sync</div>
<div>(1:13:30 PM) khem: add their name for ownership</div><div>(1:13:41 PM) koen: 2) s/yocto/oecore/ and s/poky/something/</div><div>(1:13:55 PM) khem: it should be oe I think</div><div>(1:13:58 PM) khem: in all cases</div>
<div>(1:14:14 PM) khem: or OpenEmbedded where elaboration is needed</div><div>(1:14:19 PM) koen: let's that take discussion to the ml</div><div>(1:14:19
PM) fray: ya, I think thats the direction.. but I agree.. we need to
start posting and updating a checklist.. or we're going to miss stuff..</div>
<div>(1:14:56 PM) koen: who wants the action item to elaborate teh checklist and post it?</div><div>(1:15:11 PM) Jefro: if someone sends me a list, I'll wikify it</div><div>(1:15:13 PM) ***khem volunteers Tartarus</div>
<div>(1:15:19 PM) fray: I'm swamped until early next week</div><div>(1:15:29 PM) ***Tartarus is also swamped</div><div>(1:15:42 PM) koen: should we put it in the 1.0 or 1.1 MS for yocto?</div><div>(1:16:19
PM) fray: I'm thinking to OE specific items should be in the oe wiki..
and we can work with the yocto side to track whatever they feel is
necessary as well..</div>
<div>(1:16:21 PM) khem: yocto ?</div><div>(1:16:34 PM) khem: ah right</div><div>(1:16:35 PM) fray: we've still got things we have to do before oe-core is "good enough" to present as 'done' IMHO</div>
<div>(1:16:56 PM) khem: yes we are not there yet</div><div>(1:17:25 PM)
fray: (I do see progress -- just we've not gotten the branding and
cleanup which I think we have to do before it's 'good enough')</div>
<div>(1:18:06 PM) koen: the discussion for the branding and recipe split is best done on the core ml IMO</div><div>(1:18:10 PM) koen: any objections to that?</div><div>(1:18:11 PM) fray: koen agreed</div><div>(1:18:14 PM) Tartarus: agree</div>
<div>(1:18:20 PM) khem: agreed</div><div>(1:18:34 PM) Jefro: does someone have an action item to write to the ml for that?</div><div>(1:18:59 PM) koen: RP responded to my post about it, so it is already going on</div><div>
(1:19:03 PM) Jefro: ok</div><div>(1:19:40 PM) koen: I'll send an email to gather input for the checklist</div><div>(1:19:59 PM) koen: unless there are any other items, let's move to 03) Status report on pull model, contrib repo and guidelines</div>
<div>(1:20:35 PM) fray: I did not get my AI of posting the commit/patch
policy/guidelines.. I'll get to it hopefully tomorrow.. (I did get
additional feedback though which I'd like to incorporate)</div><div>(1:21:21 PM) Jefro: fray: I am working on similar contribution guidelines for yocto, I'd be glad to sync with yours</div>
<div>(1:21:56 PM) fray: yes, we should.. we discussed last week using
the guidelines as a base.. then layers can deviate from the in
documented ways (documented by the layer that is)</div><div>(1:22:26 PM) fray: as Khem mentioned before, the contrib repo is now visible.. I don't know if anyone is using it yet though..</div>
<div>(1:22:36 PM) fray: pull model seems to be working as well..</div><div>(1:22:36 PM) khem: I am using</div><div>(1:22:45 PM) koen: fray: people are using it already</div><div>(1:22:47 PM) fray: perfect.. then we've at least got one user...</div>
<div>(1:22:54 PM) fray: even better.</div><div>(1:22:56 PM) khem: I saw Martin use it</div><div>(1:23:08 PM) khem: but devs are free to use their own public repos</div><div>(1:23:18 PM) fray: yup.. we need to make that clear as well</div>
<div>(1:23:18 PM) khem: its just a convenience for devs</div><div>(1:23:18 PM) koen: As meta-oe leader I would like to copy the oe-core guidelines as much as possible</div><div>(1:23:29 PM) Tartarus: esp if someone makes the pull request script figure out its url dynamically :)</div>
<div>(1:23:40 PM) khem: correct</div><div>(1:23:45 PM) fray: would be nice to add that to the "list of things to do" ;)</div><div>(1:24:34 PM) fray: on to the next?</div><div>(1:24:45 PM) khem: Yes</div><div>(1:25:01 PM) fray: 04) Status report on board support layer guidelines</div>
<div>(1:25:17 PM) fray: I haven't seen anything back from the board --
or RP on the Yocto side yet.. so I don't think there is anything to
report yet..</div><div>(1:25:24 PM) koen: I think we agreed to use the yocto one, minus the scary linux-yocto stuff</div>
<div>(1:25:34 PM) Tartarus: Shall we ping the board about an anwser?</div><div>(1:25:39 PM) Tartarus: Just to keep things rolling</div><div>(1:25:55 PM) fray: we need to find out the LF position on BSP postings.. then feed that to the OE board</div>
<div>(1:26:08 PM) fray: (that non-LF member positings of BSPs)</div><div>(1:26:40 PM) fray: so I'd keep the action open for RP to report back.. then we send it to the board</div><div>(1:26:51 PM) khem: right pinging board would be nice</div>
<div>(1:27:45 PM) fray: ok.. on to 05?</div><div>(1:27:50 PM) Jefro: 1 quick...</div><div>(1:27:57 PM) koen: Jefro: yes?</div><div>(1:28:52 PM) koen: 05) Status report on version retention policy and interaction with oe-core / meta-oe / $distro layers</div>
<div>(1:28:57 PM) Tartarus: OK</div><div>(1:29:03 PM) Tartarus: I posted the TSC bit to the ML</div><div>(1:29:06 PM) Tartarus: got some feedback/questions</div><div>(1:29:08 PM) koen: I like the document fray and Tartarus came up with</div>
<div>(1:29:11 PM) Tartarus: and agenda items</div><div>(1:29:13 PM) khem: I think we need a strategy to retire versions</div><div>(1:29:16 PM) khem: from oe-core</div><div>(1:29:27 PM) Tartarus: I think we're +/- ready to wiki it</div>
<div>(1:29:34 PM) fray: ya, I think that was what I got out of some of the oe-devel questions</div><div>(1:29:35 PM) koen: I would like to say that I'm not fond of micro-managing and documenting all the usecases</div>
<div>(1:29:47 PM) Tartarus: It's boiling down to "need the TSC to define
what is in what catagory" and "guess we need to see how it goes in
practice"</div><div>(1:29:52 PM) fray: and yes, I agree I think we're close enough to wiki it.. and continue to revise based on feedback</div>
<div>(1:30:00 PM) fray: Tartarus agreed</div><div>(1:30:25 PM) Tartarus: So next AI, I'll take it</div><div>(1:30:26 PM) Tartarus: wiki it</div><div>(1:30:30 PM) koen: right</div><div>(1:30:37 PM) fray: ok</div><div>
(1:30:43 PM) koen: 6) Status on layer splitting of metadata ?</div>
<div>(1:30:59 PM) khem: I have been using meta-oe layer happily</div><div>(1:31:02 PM) RP__: Sorry, lost track of time :(</div><div>(1:31:06 PM) khem: should we keep populating it</div><div>(1:31:12 PM) koen: khem: yes</div>
<div>(1:31:17 PM) khem: RP__: welcome</div><div>(1:31:28 PM) fray: no
problem.. we've got one question for you on 04 -- did you get info from
the LF on their position on non-LF members posting BSPs?</div><div>(1:31:39 PM) khem: koen: should we aim to make it the rest-of-oe-layer</div>
<div>(1:31:53 PM) RP__: I have the action to write a proposal for the steering group who will then discuss it at their first meeting</div><div>(1:32:03 PM) fray: ok</div><div>(1:32:10 PM) RP__: So we probably won't get an answer until after they meet</div>
<div>(1:32:12 PM) fray: back to 06 -- yes I think we need to keep working on meta-oe..</div><div>(1:32:17 PM) koen: khem: my original plan is to populate it further and continially evaluate it</div><div>(1:32:30 PM) khem: koen: ok</div>
<div>(1:32:31 PM) fray: it seems to be going along well from my outsider view (I've only been using oe-core so far)</div><div>(1:32:57 PM) khem: koen: ok</div><div>(1:33:10 PM) koen: I think we need to make it the pull strategy official by copying the oe-core guidelines</div>
<div>(1:33:20 PM) khem: koen: yes I agree to that</div><div>(1:33:26 PM) RP__: I must admit I'd thought we'd already dicussed a lot about meta-oe</div><div>(1:33:27 PM) koen: I volunteer to be pullmaster or whatever it's called</div>
<div>(1:33:48 PM) koen: RP__: yes, we did, but some people only read what they want</div><div>(1:33:48 PM) RP__: I think for meta-oe, it is probably reasonable to have areas of ownership</div><div>(1:33:56
PM) fray: I personally like the pull model, I'm just worried the churn
may be too high in meta-oe.. So 1) who will handle the pulls? and 2)
are there too many for one (or a small group) to handle?</div>
<div>(1:34:07 PM) fray: RP__ ya, that was going to be my next comment -- so I agree</div><div>(1:34:21 PM) RP__: Those areas of ownership need to be clear</div><div>(1:34:28 PM) RP__: documented in meta-oe actually</div>
<div>
(1:34:31 PM) fray: the question then in areas of ownership, will there
be a breakdown recipes-* like oe-core? or is it still fairly flat?</div><div>(1:34:53 PM) koen: fray: <a href="http://cgit.openembedded.org/cgit.cgi/meta-openembedded/tree/" target="_blank">http://cgit.openembedded.org/cgit.cgi/meta-openembedded/tree/</a></div>
<div>(1:34:54 PM) RP__: I'm imagining multiple layers in meta-oe along themes</div><div>(1:35:07 PM) RP__: e.g. one kind of like the graveyard we discussed</div><div>(1:35:22 PM) fray: ahh, I didn't realize you'd already broken it up.....</div>
<div>(1:35:23 PM) khem: RP__: we can have a meta-graveyard</div><div>(1:35:32 PM) fray: so then it should be easier to assign "ownership"..</div><div>(1:35:32 PM) khem: but meta-oe should be a better one</div><div>
(1:35:56 PM) koen: right now both the number or recipes and patches are small</div><div>(1:36:03 PM) RP__: koen: so are you saying no layers in meta-oe, just a split of recipes?</div><div>(1:36:22 PM) koen: RP__: at the moment no split is needed yet</div>
<div>(1:36:33 PM) RP__: koen: agreed, I'm just thinking ahead</div><div>(1:36:40 PM) koen: bit I certainly for see e.g. gnome getting its own layer</div><div>(1:36:43
PM) fray: I think as the size grows, adding more layers may become
necessary.. but we should finish the split and go ahead with the
existing directior for now..</div>
<div>(1:36:52 PM) khem: I think eventually a split might be needed though</div><div>(1:36:53 PM) fray: koen, yup matches my thinking currently</div><div>(1:37:02 PM) khem: what should be the splitting criteria</div><div>
(1:37:14 PM) koen: I think willing maintainers would be one</div>
<div>(1:37:30 PM) fray: "systems" or related functionality.. that
enables a maintainer and/or user to more logically/easily use the
component..</div><div>(1:37:32 PM) Tartarus: Do we really need more than that?</div>
<div>(1:37:37 PM) koen: and layer interdependance another</div><div>(1:37:47 PM) RP__: Do we want to have the meta-oe layout reflect this now?</div><div>(1:37:49 PM) koen: or rather non-interdependance</div><div>(1:37:53 PM) Tartarus: willing maintainers (following the normal/established elsewhere quality guildelines)</div>
<div>(1:38:13 PM) koen: RP__: I really don't know</div><div>(1:38:21 PM) RP__: That is my main concern</div><div>(1:38:38 PM) koen: there's too little experience with layers in the OE community</div><div>(1:38:46 PM) koen: and way too little maintainer experience</div>
<div>(1:39:00 PM) RP__: I think whats worrying me is its hard for someone to say "I want to maintain a TV UI layer"</div><div>(1:39:03 PM) fray: I think thats something that will come with time and a desire to control specific areas..</div>
<div>(1:39:17 PM) RP__: if the structure could accomodate that now, I think we'd help things in future</div><div>(1:39:22 PM) RP__: opie and gpe spring to mind</div><div>(1:39:27 PM) fray: ya</div><div>(1:39:41 PM) khem: so I think making a meta and meta-xxx structure would be better</div>
<div>(1:39:48 PM) khem: iow arrange it like oe-core ?</div><div>(1:40:06 PM) koen: shall we call it meta-oe to avoid confusion?</div><div>(1:40:09 PM) fray: I think the arrangement of recipes-* is a good one.....</div><div>
(1:40:12 PM) RP__: I think we need to move the data in the repo down a level into a meta-oe now, yes</div><div>(1:40:25 PM) RP__: Better to take the pain now than later</div><div>(1:40:29 PM) koen: I'll take that AI if people agree</div>
<div>(1:40:31 PM) khem: I agree</div><div>(1:40:40 PM) Tartarus: agree</div><div>(1:41:30 PM) ***RP__ is obviously in favour :)</div><div>(1:41:35 PM) fray: agree</div><div>(1:41:54 PM) koen: ok</div><div>(1:41:56 PM) fray: 07 then?</div>
<div>(1:42:07 PM) RP__: We can then put something in place saying the meta-oe is maintained by X model and by Y people</div><div>(1:42:11 PM) khem: Do we have a mechanism to track ABI between layers ?</div><div>(1:42:25 PM) RP__: khem: It needs to be written really</div>
<div>(1:42:41 PM) khem: ok</div><div>(1:42:43 PM) fray: I don't think we do.. and yes, I suggest document it for now</div><div>(1:43:07
PM) RP__: FWIW, in the post 1.0 yocto time frame I'm boxing some time
to look at this. I'd also like to talk about it in person, kind of
brainstorning maybe at the SF meeting</div>
<div>(1:43:17 PM) khem: good idea</div><div>(1:43:40 PM) Jefro: (is this still #6, layer splitting? I'll note for SF agenda)</div><div>(1:43:43
PM) RP__: (makes two things for the meeting agenda - renaming things in
oecore, an interactive session and also brainstorming layers)</div>
<div>(1:43:44 PM) khem: may be we should keep preparing agenda for F2F along the way</div><div>(1:43:51 PM) fray: (still 6)</div><div>(1:45:27 PM) fray: ok.. 07 now?</div><div>(1:45:39 PM) ***fray wants to get to the flags discussion.. :)</div>
<div>(1:45:43 PM) ***Jefro notices there are two 07s on the agenda - first one is DISTRO_FEATURE</div><div>(1:45:46 PM) koen: 07) DISTRO_FEATURE/MACHINE_FEATURE/libc features and other "flags" discussion</div><div>
(1:45:48 PM) Tartarus: So, for 07 I think I'd like to agree on the problem scope here and then move it to the ML.</div><div>(1:45:48
PM) Tartarus: What I'm trying to say is that we need a way to get OE to
meet both the needs of distributions that maintain a feed (and have end
users that create feeds) and users that care much more about initial
build time/space than feeds. And it's important that the feed use case
people not have to waste time debugging problems introduced by a user
that put up a feed and turned off things like someone in the second case
would do. Can we all agree that we need to modi</div>
<div>(1:46:10 PM) Tartarus: Historically there's been some talks about
this and an evolving opinion about what we'll allow and not allow</div><div>(1:46:12 PM) koen: you got cut off at 'modi'</div><div>(1:46:25 PM) Tartarus: Can we all agree that we need to modify things to allow for both use cases?</div>
<div>(1:46:58 PM) fray: Documenting the expected "flags" as part of the feed I think is needed..</div><div>(1:47:05
PM) Tartarus: And, I think with the signature code we've got in
oe-core, we've got the starting point for being able to fit both cases</div>
<div>(1:47:14 PM) Tartarus: fray, right, but that's not enough</div><div>(1:47:17
PM) fray: I have another concern about this and that is namespace..
I'm really worried about namespace pollution between the different
types of flags we have</div>
<div>(1:47:22 PM) RP__: Tartarus: I agree we do need to handle this, I
think we just need to do it carefully and it needs a thought out plan</div><div>(1:47:29
PM) Tartarus: If someone ignores that, changes the flags, we can't have
the distro people waste time debugging the problem</div>
<div>(1:47:46 PM) Tartarus: RP__, agree, it won't be easy</div><div>(1:47:55 PM) RP__: Tartarus: checksums might help in some ways</div><div>(1:47:57 PM) Tartarus: But I want agreement from everyone that it's something we need to solve</div>
<div>(1:48:06 PM) RP__: "send me the siginfo file for what you built"</div><div>(1:48:07
PM) fray: one possible way to do this is be sure to include the
configuration elements in the checksum, and then embed the checksum into
the package "somehow" (RPM packages have a method to do this already)..
the only concern is we need tools that can verify these things based on
the policy the user cares about</div>
<div>(1:48:14 PM) Tartarus: wait a minute fray :)</div><div>(1:48:19 PM) Tartarus: I want the real discussion on the ML</div><div>(1:48:27 PM) Tartarus: I just want agreement to tackle the problem now</div><div>(1:48:29 PM) Tartarus: brainstorm on the ML</div>
<div>(1:48:32 PM) Tartarus: and maybe more in SF</div><div>(1:48:37 PM)
RP__: Tartarus: No argument from me, we need a solution. FWIW its on the
post 1.0 list of things yocto wants to look at</div><div>(1:48:47 PM) koen: I agree that we need to tackle it</div>
<div>(1:48:48 PM) fray: yes, we need to tackle this.. :) I'm happy with
the mailing list for initial and move the discussion to in-person is SF
as well</div><div>(1:48:56 PM) Tartarus: OK</div><div>(1:48:59 PM) koen: now is the time to come up with something consistent between layers</div>
<div>(1:49:00 PM) Tartarus: khem, agree?</div><div>(1:49:01 PM) RP__: My only requirement is that we need to do it with some kind of discussed plan</div><div>(1:49:08 PM) Tartarus: RP__, agreed</div><div>(1:49:09 PM) ***khem is reading</div>
<div>(1:49:20 PM) Jefro: Tartarus has action item to move this to the mailing list?</div><div>(1:49:24 PM) khem: Yes agree</div><div>(1:49:28 PM) Tartarus: Jefro: Agreed :)</div><div>(1:49:29 PM) koen: read for 07) Continue discussion on posting of policies and guidelines ?</div>
<div>(1:49:42 PM) Tartarus: ready for 07b :)</div><div>(1:49:48 PM) fray: purpose was where do we post them, and when...</div><div>(1:50:02 PM) fray: I think we may already have the when figured out.. the where may still need some discussion or not?</div>
<div>(1:50:28 PM) RP__: fray: the oe-core list is probably where this kind of architecture discussion needs to happen</div><div>(1:50:30 PM) fray: ohh ya, and when it's posted what makes it a draft vs actual policy.. those kinds of things..</div>
<div>(1:50:32 PM) koen: I think the ml(s) are good for further discussing</div><div>(1:50:52 PM) fray: yup</div><div>(1:50:52
PM) RP__: Can someone volunteer to highlight the oe-core mailing list
on oe-dev and tell people discussion is going to happen there?</div>
<div>(1:50:56 PM) Jefro: I suggest mailing lists for discussion, but
wiki for actual text, with notations to describe draft vs. official</div><div>(1:51:18 PM) khem: RP__: I can do that</div><div>(1:51:40
PM) fray: Jefro thats what I was thinking as well.. just need a where
for things.. (and maybe some of this is already decided, but last time I
looked it wasnt completely clear to me)</div>
<div>(1:51:46 PM) RP__: The workflow I can envisage is we discuss, on
the ml and f2f, then we publish a draft. That then either becomes final
or gets ammeded until it becomes final</div><div>(1:51:51 PM) RP__: khem: thanks</div>
<div>(1:52:00 PM) koen: RP__: sounds like a plan</div><div>(1:52:14 PM)
***Jefro can volunteer to take a look at the existing wiki and suggest
structural changes, but not until after Yocto 1.0 madness</div><div>(1:52:15 PM) ***RP__ has tried to do this in the past for things that are considered "architecture" </div>
<div>(1:52:16 PM) koen: anything else, or should we go to 08) Continue discussions on infrastructure items ?</div><div>(1:53:16
PM) fray: Jefro, thats fine.. my specific issues isn't with the oe wiki
in general, just the policies and how they relate to the oe-core,
meta-oe and related going forward..</div>
<div>(1:53:29 PM) fray: koen, ya 08 is fine</div><div>(1:53:55 PM) koen: OK, 08) Continue discussions on infrastructure items</div><div>(1:54:04
PM) fray: I'm not sure what else is needed in 08 at this point.. but we
left last weeks meeting saying we'd raise some issues on the ML.. but
I'm not sure anyone did</div>
<div>(1:54:05 PM) koen: let's tackle actual hw first</div><div>(1:54:54 PM) koen: the admins send out a doc requesting hw and/or funding</div><div>(1:54:59 PM) koen: kick that to the board and LF?</div><div>(1:55:11 PM) RP__: koen: has the board responded to that yet?</div>
<div>(1:55:17 PM) RP__: I did pass it on the the LF</div><div>(1:55:25 PM) koen: RP__: it only went out to philip and me on the OE side</div><div>(1:55:35 PM) RP__: The trouble is Yocto is also hw constrained at the moment, hard as that might be to believe :/</div>
<div>(1:55:45 PM) Tartarus: Er</div><div>(1:55:47 PM) Tartarus: Where did this go?</div><div>(1:55:53 PM) Tartarus: Did I miss a members email?</div><div>(1:55:56 PM) RP__: I am discussing hardware options with people</div>
<div>(1:55:56 PM) koen: no</div><div>(1:56:24 PM) Tartarus: Any reason it didn't go to the members list? Or, why is this TSC related?</div><div>(1:56:25 PM) koen: Tartarus: Tom King made a draft on what is needed for the future</div>
<div>(1:56:38 PM) khem: I think autobuilder or some sort of tinderbox like functionality is what we are missing</div><div>(1:56:40 PM) Tartarus: So we aren't yet at the point of soliciting funds?</div><div>(1:56:40 PM) koen: it kinda is TSC related since it involved autobuilders and stuff</div>
<div>(1:57:22 PM) Tartarus: I know Tom King has been working on something to replace the old oestats thing I helped overwhelm, heh</div><div>(1:57:45 PM) RP__: The trouble is the OE infrastructure is on machines shared with four other projects</div>
<div>(1:57:55 PM) khem: yes</div><div>(1:57:57 PM) khem: indeed</div><div>(1:58:03 PM) RP__: So Tom is requesting all five projects find some funds</div><div>(1:58:17 PM) ***koen suggests killing bugzilla to free up some space</div>
<div>(1:58:28 PM) RP__: koen: ;-)</div><div>(1:58:33 PM) Tartarus: If we can get an "official" request for Funds I'll of course see if I can get any, no promises of course</div><div>(1:58:37 PM) khem: koen: space is not that big a problem</div>
<div>(1:59:22 PM) ***Jefro has another meeting in 1 min</div><div>(1:59:26 PM) khem: We are out of time I guess</div><div>(1:59:28 PM) koen: khem: I was only being half serious</div><div>(1:59:42 PM) Tartarus: I think we can call it then</div>
<div>(1:59:46 PM) fray: lets plan to move the last item nearer the beginning next time... and follow up on the status items as well..</div><div>(1:59:47 PM) Tartarus: got almost everything talked about</div><div>(1:59:48 PM) Jefro: I suggest that 09 was discussed earlier</div>
<div>(1:59:49 PM) khem: its hard to get tangents in write only meetings :)</div><div>(1:59:58 PM) fray: :)</div><div>(2:00:02 PM) koen: yes, 09 was discussed in earlier meetings</div><div>(2:00:12 PM) koen: and put in minutes</div>
<div>(2:00:14 PM) koen: and in mails</div><div>(2:00:15 PM) Jefro: earlier in this meeting as well, unless I misunderstood what it meant</div><div>(2:00:29 PM) fray: other then that, I'm good to call it as well..</div>
<div>(2:00:40 PM) khem: I would like to see OE-core support more arches and more libs is that viable goal</div><div>(2:00:56 PM) ***khem always have closing question</div><div>(2:01:03 PM) fray: I'd like the arches, (multi)libs documented.. :) So I know whats all there.. ;)</div>
<div>(2:01:22 PM) khem: I can document them</div><div>(2:01:23 PM) fray: koen -- just don't want to leave it hanging, even if there are still open issues.. thats my only concern</div><div>(2:01:39 PM) koen: true</div>
<div>(2:01:43 PM) Tartarus: *ding*</div><div>(2:01:49 PM) Tartarus: time for MLs and so on, and normal channels</div><div>(2:01:50 PM) Tartarus left the room.</div><div>(2:01:51 PM) fray: khem, that should be part of the on-going oe-core guidelines/policies stuff..</div>
<div>(2:01:55 PM) fray: yup.. meeting over ..</div><div><br></div></div><font color="#888888"><br clear="all"><br>-- <br>Jeff Osier-Mixon<div>Yocto Community Manager @Intel</div></font>