[oe] MINUTES: OE TSC 28 February 2012

[Jeff Osier-Mixon]

OpenEmbedded Technical Steering Committee
28 February, 2012

Attending: Khem, Richard, Mark, Paul
Apologies: Tom
Minutes: Jefro

________________________________________________
Agenda

1. pick a chair
   RP (by default)

2. open issues
   resolved at ELC, only status is left

3. status

   a. oe-core
      lots of churn, patches coming in heavily
      RP thinks most things merged

   b. python 3 transition
      going slowly

   c. bitbake quoting fix merged, now enforce syntax, e.g. FOO = "1"

   d. khem worked on some wiki pages last week

   e. hobv2 merged, please play with it

   f. 4.7 upgrade not yet posted (khem), hold until after April release

   g. agreement on toolchain issue:
      toolchains not in oe-core will be split out into their own layers,
      including toolchains in meta-oe

   h. collab summit, bsp summit
      coming up in April

________________________________________________
Raw Transcript

(9:00:04 AM) bluelightning: hi folks
(9:02:12 AM) RP__: hi bluelightning
(9:02:16 AM) RP__: hi khem
(9:02:19 AM) RP__: fray: ping?
(9:02:35 AM) khem: hi RP__
(9:02:37 AM) RP__: I think I saw an email saying Tom couldn't make it?
(9:02:38 AM) fray: hey I'm here..
(9:02:44 AM) fray: ya, he's out
(9:02:48 AM) khem: I will grab a cup of tea
(9:02:57 AM) RP__: cool, we have everyone who will be here then
(9:02:58 AM) khem: too many police officers on road today :)
(9:03:32 AM) ***fray is really sick last couple of days.. at least the room
stopped spinning finally
(9:03:45 AM) ***fray 's temper is a bit short though.. :(
(9:04:49 AM) RP__: fray: Sorry to hear that, I hope you have a swift
recovery
(9:05:00 AM) RP__: Jefro: do we have an agenda?
(9:05:18 AM) fray: so do I.. I'm supposed to give an intro to Yocto talk
and hands-on w/ Poky class at RTECC in Dallas next tuesday.. :P
(9:06:15 AM) RP__: fray: I'm just getting stressed out by the amount of
stuff I need to do before I go away on Saturday
(9:06:35 AM) ***RP__ found water all over the garage floor before. Seems to
be coming from the Daytona which is scary
(9:06:45 AM) fray: I understand.. I was supposed to have the hard drive
prepared for imaging yesterday..
(9:06:54 AM) RP__: Does anyone have topics for discussion?
(9:06:57 AM) fray: needless to say, I'm still working on it.. (I was
completely useless on Sunday and Montday)
(9:07:00 AM) fray: I don't have anything
(9:07:15 AM) fray: did we want to mention the wiki day?  (do we want to
schedule that yet?)
(9:07:24 AM) khem: I dont have any pressing topics
(9:07:28 AM) RP__: I think it would be good to give people some notice
(9:07:37 AM) RP__: Sometime after next week though! ;-)
(9:07:41 AM) khem: 2 weeks would be good time
(9:07:43 AM) fray: yes
(9:07:57 AM) fray: we said mid march in the ELC meetings
(9:08:12 AM) RP__: so that all approximately fits together
(9:08:22 AM) RP__: who is going to send out a proposal?
(9:08:51 AM) bluelightning: I can do it
(9:09:55 AM) Jefro: sorry - fell of the net again (*)&%
(9:10:00 AM) RP__: bluelightning: thanks
(9:10:20 AM) Jefro: we actually fulfilled the major topics at the f2f,
status is all we have to discuss today
(9:10:28 AM) RP__: Jefro: thanks
(9:10:40 AM) RP__: We've had a lot of churn with the patches coming in
rather heavily
(9:10:45 AM) RP__: I think most things have merged
(9:11:00 AM) RP__: I know there are some patches from Khem I've not taken
yet and need to discuss in more detail
(9:11:39 AM) RP__: There are some cleanups I would like to make such as the
", 1" -> ", True" substituition, use appendVar in places and so on
(9:12:06 AM) khem: RP__: is there plan to go for python 3.x for bitbake
(9:12:08 AM) RP__: I don't seem to manage to find time to get to those
though and a merge window is a bad time due to the amount of patches that
would fail to apply
(9:12:18 AM) fray: yup
(9:12:30 AM) RP__: khem: we've done a lot of the ground work so far
(9:12:36 AM) RP__: khem: so its being worked on slowly
(9:12:49 AM) khem: ok.
(9:12:50 AM) fray: we should investigate bitbake w/ python 3 after the
current release cycle..
(9:12:59 AM) fray: just to get any idea how close (or not) we are..
(9:13:24 AM) ***RP__ agrees
(9:13:33 AM) RP__: I also merged the bitbake quoting fix
(9:13:44 AM) RP__: nbbody was going to fix things until I did :/
(9:14:02 AM) fray: which quoting fix?
(9:14:11 AM) ***bluelightning has already added a note on quoting to the
metadata migration wiki page, FYI
(9:14:21 AM) khem: Generally bitbake's error recovery sucks a bit imo
(9:14:38 AM) khem: I did touch some wiki page's last week
(9:14:48 AM) khem: to mention about oe-core/oe-dev
(9:14:54 AM) RP__: fray: We now enforce FOO = "1"
(9:14:56 AM) bluelightning: khem: I'd like to think we've improved error
handling quite a bit lately, are there any outstanding error situations you
can point to?
(9:14:57 AM) khem: one of the page was getting started page
(9:15:01 AM) fray: ahh ok
(9:15:02 AM) RP__: fray: FOO = 1 wont' work
(9:15:13 AM) fray: excellent
(9:15:19 AM) RP__: khem: I think bitbake's error handling has massively
improved thanks to Paul's work
(9:15:37 AM) khem: RP__: I agree on that it has improved
(9:15:39 AM) RP__: Or my favourite: FOO = "1 \
(9:15:43 AM) bluelightning: if I've missed anything by all means highlight
it so I can attack it :)
(9:15:48 AM) RP__: with a following blank line
(9:16:01 AM) RP__: khem: right, if you see issues, file bugs
(9:16:02 AM) bluelightning: RP__: yes that's a nice one... never realised
that was legal...
(9:16:12 AM) RP__: bluelightning: it wasn't meant to be legal
(9:16:49 AM) khem: multiline strings and multiline comments are tricky even
C compilers sucked in past on them
(9:17:38 AM) RP__: So anything else we need to talk about
(9:17:41 AM) RP__: ?
(9:17:53 AM) RP__: We should be going into a bit of a stability/bug fixing
drive now
(9:18:04 AM) bluelightning: I've got nothing, most of my long-pending
patches are getting merged finally :)
(9:18:08 AM) RP__: also,hobv2 merged
(9:18:14 AM) RP__: people should play with it :)
(9:18:25 AM) Jefro: I can chime in with one thing - the Yocto event at
Collab Summit. Philip mentioned that he'd like to have several OE folks
there.
(9:18:28 AM) fray: ahh I knew hobv2 was going in soon, but it's there now?
(9:18:31 AM) ***RP__ has a stack of patches outstanding :(
(9:18:44 AM) RP__: fray: its in
(9:18:54 AM) fray: excellent!
(9:18:56 AM) RP__: fray: run "hob"
(9:19:03 AM) khem: RP__: I have a working gcc 4.7 which is uncovering
interting bugs
(9:19:15 AM) khem: so I will post fixes as I see and fix them
(9:19:29 AM) RP__: khem: I've held off the gcc patch just yet until I can
have more of a look at it
(9:19:36 AM) khem: sure.
(9:19:48 AM) khem: I did not post 4.7 upgrade yet
(9:19:53 AM) RP__: khem: we should try and push some of these fixes direct
to the upstreams
(9:19:56 AM) khem: since I wanted the prework to first go in
(9:20:12 AM) khem: RP__: yes, true
(9:20:30 AM) RP__: khem: right, I'd really like to hold 4.7 on a branch
until after the April release
(9:20:36 AM) fray: is 4.7 "stable" release out yet?  or any idea when it
will be?
(9:20:40 AM) khem: nooo
(9:20:47 AM) khem: its a prework
(9:21:02 AM) fray: I'm almost say a toolchain layer until 4.7 can be
merged.. maybe a toolchain-next layer?
(9:21:09 AM) RP__: khem: I know but we are supposed to be entering a
stablisation window at this point
(9:21:14 AM) khem: so we can use it offficially the same day its released
and help gcc find and fix some bugs in return
(9:21:32 AM) khem: RP__: that noo was meant for fray
(9:21:57 AM) RP__: fray: I think a branch should be ok
(9:22:02 AM) fray: ya, I thought 4.7 was still too far off to consider it
for right now
(9:22:16 AM) khem: RP__: the reason I did not post 4.7 patches is same
(9:22:23 AM) bluelightning: I forgot to ask people; were there significant
discussions at ELC about oe-core/meta-oe reorg that are worth discussing
here?
(9:22:37 AM) khem: I wanted prework to be in there so folks can then use
the 4.7 branch to try it out
(9:22:40 AM) ***RP__ wasn't party to those
(9:22:56 AM) khem: bluelightning: I talked a bit
(9:22:57 AM) RP__: khem: I agree we should put it somewhere obvious so
people can try it
(9:22:59 AM) khem: with koen
(9:23:11 AM) khem: RP__: master-next ?
(9:23:17 AM) khem: oh well a contrib branch is ok too
(9:23:39 AM) bluelightning: khem: ah ok... I talked to him at FOSDEM about
it as well, took some notes... haven't had a chance to formulate a coherent
plan yet
(9:23:50 AM) khem: bluelightning: I think we should keep the discussion
about sprilling meta-oe going
(9:23:52 AM) RP__: khem: it might be appropriate for master-next, yes
(9:24:12 AM) bluelightning: khem: definitely... we have a tentative plan I
think from our existing discussion
(9:24:25 AM) bluelightning: (on the mailing list)
(9:24:30 AM) khem: I think, I did some work on oe-core/meta-oe duplication
but thats pretty much it
(9:24:55 AM) bluelightning: khem: yes I saw that, thanks for that work :)
(9:25:32 AM) khem: biggest issue is toolchain for many folks
(9:25:41 AM) khem: so I think that should be the first one to be spun out
(9:25:50 AM) RP__: khem: people want 4.5?
(9:26:01 AM) khem: RP__: angstrom stable uses it
(9:26:11 AM) khem: and angstrom uses for long cycles
(9:26:15 AM) RP__: khem: with no plans to update?
(9:26:25 AM) khem: so I am expecting it to be in use for couple of years
from now at least
(9:26:57 AM) khem: RP__: I think once meta-oe releases then there would be
a chance to have 2012 angstrom
(9:27:10 AM) bluelightning: is there still concern about arm in gcc > 4.5 ?
or was that never a serious issue?
(9:27:55 AM) khem: bluelightning: no idea. I think gcc 4.5 is stable and
angstrom does not like to tinker with toolchain too much
(9:28:08 AM) bluelightning: ok
(9:28:33 AM) fray: but would anyone use 4.5 with an updated oe-core?
(9:28:46 AM) fray: I'd expect that if you stay with an older toolchain,
you'd also stay with an older oe-core/angstrom
(9:28:51 AM) bluelightning: fray: well, angstrom does right now...
(9:28:51 AM) khem: so far angstrom  has been doing that yes
(9:29:12 AM) khem: usually its not a problem
(9:29:17 AM) fray: that very well could be a distro choice, so we need to
enable the ability to do it.. but do we need it in oe-core itself?
(9:29:39 AM) khem: its not in oe-core
(9:29:44 AM) khem: its in meta-oe
(9:29:49 AM) fray: ohh ok, I misunderstood then
(9:30:02 AM) khem: same is true for eglibc 2.12
(9:30:13 AM) bluelightning: fray: it's in meta-oe, but what would make it
easier is if it was in its own layer (which may well be in the
meta-openembedded repo still)
(9:30:27 AM) fray: yes, I agree..
(9:30:29 AM) khem: and I would love to see X11 from meta-oe/oe-core merge
(9:30:44 AM) fray: at Wind River we have a tendency to make our toolchains
into layers, so they are easily replacable (upgradeable)..
(9:31:12 AM) khem: fray: yes we should have done it to start with but we
were so new to layers we were afraid of them :)
(9:31:42 AM) RP__: khem: X11?
(9:31:45 AM) fray: well there is the fine line between not enough layers
and too many.. ;)
(9:31:51 AM) RP__: khem: I thought we had most X11 things sorted now?
(9:32:07 AM) fray: I didn't realize that meta-oe had anything unique for
X11 anymore..
(9:32:18 AM) RP__: I think splitting the toolchain pieces in meta-oe apart
is a fine idea and nobody seems to object to that :)
(9:32:20 AM) fray: I thought it had been merged, other then a few utilities
(9:33:37 AM) khem: I tried core-image-sato build with meta-oe enabled
(9:33:40 AM) khem: did not work
(9:34:07 AM) RP__: khem: hmm, any idea what the issue was
(9:34:11 AM) khem: so I vehemently blamed it
(9:34:14 AM) RP__: we need to find and deal with things like this
(9:34:46 AM) khem: RP__: yes, we should but then something else pops up :)
(9:34:50 AM) fray: yes..  we should probably have someone look into meta-oe
and plan for a merge if reasonable..  or at least identify issues like this
(9:35:11 AM) khem: my point is it should be easily overlayed on oe-core and
not regress oe-core
(9:36:13 AM) RP__: khem: agreed, just need someone with some time to look
at it
(9:36:17 AM) fray: yes
(9:36:30 AM) khem: bluelightning: will you send follow up to the layer
split discussions of your meeting at FOSDEM
(9:36:43 AM) bluelightning: khem: will do yes
(9:36:57 AM) khem: that should get us some  further in decision process
(9:37:21 AM) bluelightning: khem: I think we should just press ahead on the
toolchain split out though
(9:37:35 AM) RP__: As the TSC, I think we can agree that?
(9:37:39 AM) khem: bluelightning: yeah
(9:38:03 AM) fray: yes
(9:38:05 AM) ***Jefro listens for detractors
(9:38:16 AM) khem: ok.
(9:38:45 AM) khem: Jefro: when/where is Collab Summit
(9:38:54 AM) Jefro: Collab Summit is April 3-5 in San Francisco
(9:39:07 AM) Jefro: there will be a Yocto working day all day on Weds Apr 4
(9:39:21 AM) Jefro: Collab is invitation-only- I am working to secure a
certain number of invitations
(9:39:22 AM) khem: ok and thats in Fremont ?
(9:39:40 AM) Jefro: no - there is also an invitation-only Yocto BSP Summit
in Fremont, that is on April 2
(9:39:48 AM) khem: Jefro: extend to me if you have spare :)
(9:40:00 AM) Jefro: khem I will definitely try
(9:40:09 AM) khem: both yocto/collab that is
(9:40:22 AM) khem: oh brb
(9:40:55 AM) RP__: The bsp summit is to try and get everyone on the same
page around BSPs and layers, the various definitions and so on
(9:41:21 AM) bluelightning: hmm that would be interesting...
(9:41:39 AM) RP__: if the yocto people can't reach a consensus amongst
themselves it seems crazy to involve more people in the discussion
(9:41:39 AM) Jefro: bluelightning you should have gotten an invitation. any
chance you'll be in town?
(9:41:56 AM) RP__: I don't think there is that much of an issue, just need
to get people together and spend some time talking through what we have
today
(9:42:02 AM) RP__: and then see where things go
(9:43:04 AM) bluelightning: Jefro: I'm not sure... I need to check if I'm
available, and also would depend on travel approval
(9:43:17 AM) Jefro: yes, travel will be a problem
(9:44:03 AM) Jefro: I'm planning to at least record parts of the event.
Will also talk to Mentor to see if we can let people join via webex, but
interactive participation would be limited.
(9:46:00 AM) Jefro: any other items?
(9:46:21 AM) bluelightning: khem's still afk I think
(9:46:33 AM) RP__: nothing from me
(9:46:43 AM) bluelightning: nothing from me either

-- 
Jeff Osier-Mixon http://jefro.net/blog
Yocto Project Community Manager @Intel http://yoctoproject.org