Source

extradoc / minute / pypy-sync-2006-03-23.txt

Full commit
=============================================
pypy-sync developer meeting 23rd March 2006
=============================================


Attendees:
  - Anders C.
  - Anders L.
  - Armin
  - Aurelien Campeas
  - Carl Friedrich
  - Christian
  - Holger
  - Michael (minutes)
  - Nik
  - Richard
  - Samuele

Summary
=======

 - activity reports

   See `the log`_ below.

 - status of tokyo sprint

   Samuele reported that the sprint announcement is nearly ready and
   should go out soon.  Planning is hampered by the fact that we don't
   really know how many Japanese participants there will be there, and
   it's not even certain how many PyPy-ers will be travelling over.
   This last bit we can actually do something about, so:

       If you're in `extradoc/sprintinfo/tokyo/people.txt`_, could you
       please update that file with your real intentions.

   .. _`extradoc/sprintinfo/tokyo/people.txt`: 
           http://codespeak.net/svn/pypy/extradoc/sprintinfo/tokyo/people.txt

 - goals of the 0.9 release

   We agreed to relase 0.9 fairly soon, probably in the second week of
   June.  The main focus of the release will be "WP07 stuff" in EU
   jargon, which is to say stackless features (including tasklet
   pickling), but the release should also include previews of the
   optimization/JIT/logic work.  Sorting out the details of this was
   delegated to the Leysin sprinters.

 - status and work planning for the next few weeks

   Basically we all know what we're working on at the moment
   (logic/GC/JIT) and we'll continue to work on these things as much
   as we're able to (holidays and bureaucracy permitting) until and
   during Leysin.

 - targetting non-Python conferences

   Fairly recently, we've missed the paper submission deadlines for
   OOPSLA and OSCON.  We need to get a handle on which conferences
   we're targetting sufficiently far in advance that we actually make
   these deadlines.  The poster deadline for OOPSLA is 30th June, we
   should try to make that...

 - moderation of the next #pypy-sync meeting

   As I am going to be on holiday, Anders Lehmann will moderate the
   pypy-sync meeting next week.

.. _`the log`:

Complete Log
============

This is the full log of the meeting::

    [16:02] stakkars: hi
    [16:03] hpk: hi christian!
    [16:03] hpk: hi all
    [16:03] aleale: hi
    [16:04] rxe joined the chat room.
    [16:04] cfbolz: should we maybe start with activivty lines?
    [16:04] cfbolz: hi richard!
    [16:04] rxe: Hi all!
    [16:05] stakkars: ok I'll start
    [16:06] hpk: ok, i guess i can take over
    [16:06] stakkars: DONE: stackless contexts (little), extension building
    [16:06] hpk: i mailed about the topcis earlier anyway
    [16:06] stakkars: NEXT: the raymond module :-)
    [16:06] stakkars: BLOCK: five lines of docs of current state and scope of gctransform would have saved 2 days
    [16:07] auc: LAST : work on logic objspace
    [16:07] auc: NEXT : idem
    [16:07] auc: BLOCKER : possibly greenlet stuff
    [16:07] nikh: LAST: gensqueak
    [16:07] nikh: NEXT: gensqueak: tuples, maybe other data structures
    [16:07] nikh: BLOCKERS: none
    [16:07] mwh: hello
    [16:07] cfbolz: LAST: GC work
    [16:07] cfbolz: NEXT: more GC work, personal stuff
    [16:07] cfbolz: BLOCKERS: personal issues
    [16:07] hpk: LAST: US trip (hp/py.test consulting), amendment-4 and (TB) communications, codespeak problems
    [16:07] hpk: NEXT: non-pypy, still a bit of amendment work
    [16:07] hpk: BLOCKERS: exhaustion or something
    [16:07] mwh: sorry for being late...
    [16:07] aleale: PREV: working on making the official OWL test pass
    [16:07] aleale: NEXT: more of the above
    [16:07] aleale: BLOCKERS:  -
    [16:07] pedronis: LAST: make ootype more static, some jit work, helper GC stuff
    [16:07] pedronis: NEXT: more helping GC stuff, jit work, leysin sprint,
    [16:07] pedronis: BLOCKERS: -
    [16:08] mwh: i have an activity report from eric:
    [16:08] arre: PREV: Non-PyPy-work, some JIT-work, travel-planing
    [16:08] arre: NEXT: More JIT and Leysin-sprint
    [16:08] arre: BLOCKERS: -
    [16:08] mwh: LAST: got LLVM JIT codegenerator working with rgenop.py graphs
    [16:08] mwh: NEXT: extend pyllvm, maybe convert it to using ctypes
    [16:08] mwh: BLOCKERS: ctypes misses C API for LLVM
    [16:08] mwh: for myself:
    [16:08] mwh: LAST: c/exceptions/gc hacking
    [16:08] mwh: NEXT: holiday
    [16:08] mwh: BLOCKERS: -
    [16:09] hpk: goden, xorAxAx: anything?
    [16:09] cfbolz: rxe: you?
    [16:09] hpk: all others except the pybot delivered i think
    [16:09] rxe: no :-(
    [16:10] hpk: then next topic?
    [16:10] arigo joined the chat room.
    [16:10] mwh: so i don't see any blockers that can be resolved here
    [16:11] cfbolz: yes
    [16:11] cfbolz: hi armin!
    [16:11] arigo: joined "pypy-snc" by mistake
    [16:11] mwh: heh
    [16:11] stakkars: moin
    [16:11] mwh: first topic is: "status of tokyo sprint"
    [16:12] hpk: i haven't dealt with this since a number of weeks - do you have news, samuele?
    [16:12] pedronis: we got the logistic information, and an announcement has been drafted
    [16:12] mwh: my impression is that things are under control but as i'm not going i haven't really been paying attention
    [16:12] pedronis: we need to decide what are the technical goals/topics for it
    [16:12] pedronis: to finish it and send it out
    [16:12] pedronis: should happen before the weekend
    [16:13] pedronis: it's of course a bit unclear what to propose
    [16:13] hpk: any info about the participants (number/likely interests)?
    [16:13] pedronis: because things are so in flux in general right now
    [16:13] pedronis: there may be 20 japanese
    [16:13] hpk: i remember that bea/me tried to extract some information and one feedback was "might be up to 20 people" but it didn't get more concrete i think
    [16:13] pedronis: form our side it would nice
    [16:13] pedronis: if people updated people.txt
    [16:14] pedronis: with their real intententions
    [16:14] hpk: yip, currently there are 9 people
    [16:14] pedronis: do know how many of "us" are really going to be there
    [16:15] arre: Jacob and Laura are not certain they can make it.
    [16:15] hpk: nikh, aleale, stakkars: you plan to go as listed in the people.txt?
    [16:15] nikh: yep, i will come
    [16:15] aleale: yep
    [16:15] stakkars: It is not clear because something else is going on.  I need to add another subtopic, just to increase confusion:
    [16:16] stakkars: EWT is considering to do a sprint in Iceland, late May. They will sponsor it.!!
    [16:16] auc: EWT ?
    [16:16] hpk: stakkars: i think we have lots of topic already, can it wait until next week?
    [16:16] stakkars: a company which I'm consulting for, interested in PyPy
    [16:16] CyDefect joined the chat room.
    [16:17] mwh: yes, this is a very packed meeting
    [16:17] auc: stakkars: but what EWT does stand for ? (if not indiscrete)
    [16:17] mwh: stakkars: is this supposed to be public information yet?
    [16:17] stakkars: hpk: I expected this to come. I only wanted to inject the information now, so people get a picture about the schedule
    [16:17] CyDefect is now known as Gromit.
    [16:18] hpk: pedronis: can you trigger bea to re-check with the japanase people and try to get the announcement finished by the weekend?
    [16:18] mwh: i'd like to move along pretty sharpish, so any concrete things to be done about tokyo ?
    [16:18] hpk: stakkars: sure, it's great - you can also post it to pypy-dev and we can discuss it there e.g.
    [16:18] mwh: like hpk just said
    [16:18] aleale: auc : see http://ewtcareers.com/
    [16:18] pedronis: hpk: that's the plan, although I think we will get more feedback form japanese side when the announcement is out
    [16:19] auc: aleale: thanks
    [16:19] stakkars: mwh: not exactly, they are discussing, it will be organized by STeve Holden, I will talk to him today. I just needed to say this now, just in case it is a bad idea
    [16:19] mwh: stakkars: ok
    [16:19] mwh: i won't make a song and dance about it in the minutes then
    [16:19] mwh: ---------
    [16:19] mwh: - goals of the 0.9 release
    [16:19] mwh: i'm guessing this is mostly going to be a stackless/gc/wp07 release
    [16:20] mwh: maybe logicobjectspace too?
    [16:20] cfbolz: gc/stackless/wp07: definitively
    [16:20] auc: mwh: depends on the timeframe
    [16:20] hpk: also it should include early or progressed optimization works
    [16:20] mwh: auc: to be released late may/early june
    [16:20] hpk: like JIT and things on object level (WP06 in EU jargon)
    [16:20] auc: mwh: that could be ok
    [16:21] auc: mwh: but not sure about *all* the constraint solving stuff
    [16:21] cfbolz: of course not
    [16:21] mwh: auc: sure
    [16:21] mwh: but some kind of meaningful preview would be cool
    [16:21] hpk: timeframe is June ... mid June like around 15th?
    [16:21] mwh: hpk: don't know if the JIT is going to be releasable for this release
    [16:22] mwh: would be cool if it was, of course
    [16:22] hpk: mwh: "include early or progressed optimization works" leaves this open
    [16:22] mwh: hpk: well, yes
    [16:22] hpk: having some results on the toy language should be feasible
    [16:23] cfbolz: we should, at least
    [16:23] mwh: i guess
    [16:23] cfbolz: next topic?
    [16:23] mwh: anyway, we don't really have time for that discussion
    [16:23] mwh: -------------
    [16:23] mwh: - status and work planning for the next few weeks
    [16:24] hpk: mwh: we can also delegate
    [16:24] mwh: hpk: true
    [16:24] hpk: in this case i'd say we agree on the June timeframe and mabye set the 15th June
    [16:24] mwh: hpk: do we want to do that?
    [16:24] hpk: and delegate to the Leysin group to define the scope and contents of the release
    [16:24] hpk: if everybody is fine with that
    [16:24] mwh: maybe a little earlier?
    [16:24] arre: Ok with me.
    [16:24] mwh: (no real reason)
    [16:24] pedronis: Ok
    [16:24] mwh: but fine with the delegation
    [16:24] stakkars: yup
    [16:25] hpk: ok, we can still adapt (10th june might be sensible as well)
    [16:25] mwh: for my part i can pretty accurately estimate how much work i'm going to do between now and leysin: none, as i go on holiday tomorrow
    [16:25] mwh: (late tomorrow, but...)
    [16:25] hpk: until leysin i will not be invovled much either except for amendment-4 and some non-core stuff, i am afraid
    [16:26] mwh: i'm guessing leysin will be jit/gc stuff
    [16:26] mwh: tasklet pickling?
    [16:26] cfbolz: yes!
    [16:26] hpk: mwh: and whatever we determine worthwhile
    [16:26] mwh: more wp09 stuff
    [16:26] stakkars: I hope I can make it to Leysin. That's a tough one
    [16:26] mwh: hpk: well, yes...
    [16:26] mwh: hpk: it was you that suggested this as a #pypy-sync topic :)
    [16:27] hpk: yes, what do you imply?
    [16:28] mwh: just that if you wanted to talk about it here, that you had some thing more to say than "and whatever we determine worthwhile"
    [16:28] stedi67 joined the chat room.
    [16:28] mwh: that hardly requires discussion...
    [16:28] hpk: mwh: that was refering to Leysin
    [16:28] mwh: no biggie
    [16:28] mwh: ok
    [16:28] hpk: and to the JIG/GC mentioning
    [16:28] mwh: so, last-but-one topic:
    [16:28] mwh: - targetting non-Python conferences
    [16:28] mwh: the oopsla deadline just sailed past us
    [16:29] mwh: how do we stop this happening again?
    [16:29] hpk: and what non-python conferences are left?
    [16:29] auc: what about ECOOP ?
    [16:29] pedronis: well, the idea is still to target the dynamic lang workshop at OOPSLA
    [16:29] mwh: a good start would be to get conference and deadlines on to the pypy calendar
    [16:29] pedronis: that has a later and different deadline than the main track
    [16:29] mwh: but i don't really know when they are
    [16:30] mwh: pedronis: ah hah
    [16:30] hpk: arigo, pedronis: i think we once said that you two would keep an eye on those conferences - can you try to list deadlines and identify worthwhile conferences from your perspective?
    [16:30] cfbolz: pedronis: ah, cool!
    [16:30] mwh: maybe i should mail pypy-dev and ask for some dates?
    [16:30] pedronis: most of the conferences had deadlines
    [16:30] pedronis: at the end of last year
    [16:30] pedronis: (nov dec)
    [16:31] mwh: argh
    [16:31] hpk: ECOOP has passed as well
    [16:32] hpk: pedronis: should we mail Roel to ask him about worthwhile conferences (explaining that we missed the deadlines nov/dec 2005)?
    [16:32] mwh: points at the clock
    [16:33] mwh: who's going to moderate next week?
    [16:33] hpk: mwh: we started 17:05 or 07 or so
    [16:33] mwh: hpk: i know
    [16:33] hpk: will probably (but not certainly) be on travel during that time next week
    [16:33] pedronis: bah, it seems that for this year dyn lang workshop
    [16:33] pedronis: they have only invited talks
    [16:34] cfbolz: pedronis: wasn't that what roel said?
    [16:34] pedronis: I forgot then
    [16:34] pedronis: too many things
    [16:34] cfbolz: yes :-(
    [16:34] cfbolz: pedronis: I guess we could still talk to him again
    [16:35] hpk: cfbolz: do you feel like asking him? I can review any mails
    [16:35] auc: maybe i'm just dense ... i can see submission deadlines for ecoop 2006 workshops around 1st April ... (2006)
    [16:35] hpk: didn't look at the workshops
    [16:35] cfbolz: hpk: sorry, not really :-(
    [16:36] pedronis: for oopsla what's left are posters
    [16:36] mwh: points at the clock, again
    [16:36] mwh: we can continue this discussion in #pypy
    [16:37] hpk: ok
    [16:37] mwh: who is going to moderate next week?
    [16:37] auc: http://icooolps.loria.fr/
    [16:37] mwh: it's not hpk (probably) and it's not me
    [16:37] auc: http://prog.vub.ac.be/~wdmeuter/RDL06/
    [16:37] auc: (roel is part of this one)
    [16:37] aleale: I can do the moderation
    [16:38] cfbolz: ok, see you all next week then
    [16:38] hpk: see you
    [16:38] rxe: Bye!
    [16:38] mwh: aleale: thank you
    [16:38] nikh: bye bye
    [16:38] mwh: meeting over
    [16:38] mwh: thanks for coming, sorry for the late start (again...)