[pypy-svn] r27104 - pypy/extradoc/minute
hpk at codespeak.net
hpk at codespeak.net
Fri May 12 08:54:48 CEST 2006
Author: hpk
Date: Fri May 12 08:54:47 2006
New Revision: 27104
Added:
pypy/extradoc/minute/pypy-sync-2006-05-11.txt
Log:
draft of minutes from yesterday's pypy-sync meeting
please amend/correct as you see fit.
Added: pypy/extradoc/minute/pypy-sync-2006-05-11.txt
==============================================================================
--- (empty file)
+++ pypy/extradoc/minute/pypy-sync-2006-05-11.txt Fri May 12 08:54:47 2006
@@ -0,0 +1,201 @@
+pypy-sync 11th May 2006, 05:00 PM - 05:25 (UTC+2)
+=====================================================
+
+attendants: michael, anto, samuele, christian, richard,
+ eric, arre, armin, holger (mods and minutes)
+
+* activity reports: everybody posted, see log below.
+
+* thread cloning approach
+
+ As discussed earlier it is urgent to have a working
+ thread cloning approach. Armin posted his suggestion
+ to pypy-dev which builds on the Leysin and other
+ previous works. He thinks it will take at least
+ 2 weeks solid work and that it involves experimentation
+ at some levels so it's hard to predict things exactly.
+ Christian agrees to this approach. Armin will start on this
+ and he will also try to get early feedback from Aurelien and Nicolas.
+ Implementation wise Michael and a bit of Holger will try
+ to help.
+
+ Moreover Christian is to describe and finalise a different
+ approach for the larger topic of "thread pickling" at the
+ early June sprint. Holger reminds that EU-wise we promised
+ the whole of Work Package 7 to be complete and having a
+ report in June and that we should get clarity about
+ according plans.
+
+* Next PyPy sprint: 2nd-9th june, Duesseldorf
+
+ It seems that Armin, Michael, Samuele, Holger, Christian, Carl,
+ Arre, Eric (partly), Anto and possibly others can make it
+ to the next PyPy sprint from
+
+ 2nd-9th June 2006
+
+ It is almost set that we are going for Duesseldorf (germany)
+ as a location. Michael to prepare an announcement until
+ saturday with review from Holger (and others). Carl to
+ continue to fix a location.
+
+ Topics will be the 0.9 release, ext-compiler, stackless and
+ more or less all topics that have been floating lately.
+
+
+IRC logs
+-------------
+
+**** BEGIN LOGGING AT Thu May 11 17:00:48 2006
+
+May 11 17:00:48 --> You are now talking on #pypy-sync
+May 11 17:00:48 --- Topic for #pypy-sync is PyPy - the flexible snake. This channel is used for pair programming and sync meetings.
+May 11 17:00:48 --- Topic for #pypy-sync set by xorAxAx at Sun Apr 16 21:44:23
+May 11 17:01:58 --> stakkars (i=gzlikbb at i577B473C.versanet.de) has joined #pypy-sync
+May 11 17:02:25 --> arigo (n=arigo at c-1f8b70d5.022-54-67626719.cust.bredbandsbolaget.se) has joined #pypy-sync
+May 11 17:02:31 <hpk> hi all
+May 11 17:02:44 <arigo> hi!
+May 11 17:02:45 <arre> hI!
+May 11 17:02:47 <ericvrp> hi
+May 11 17:02:54 <pedronis> hi
+May 11 17:03:03 <antocuni> hi
+May 11 17:03:06 <mwh> hi
+May 11 17:03:17 <hpk> ok, let's get started, i had hoped that Aurelien appears, let's see
+May 11 17:03:18 <hpk> we have
+May 11 17:03:24 <hpk> * activities
+May 11 17:03:24 <hpk> * thread cloning
+May 11 17:03:24 <hpk> * next sprint
+May 11 17:03:30 <hpk> activities then:
+May 11 17:03:37 <ericvrp> LAST: dictiterator, sequenceiterator, SoC mentor application
+May 11 17:03:37 <ericvrp> NEXT: more iterator stuff, trying to get llvm profile information
+May 11 17:03:37 <ericvrp> BLOCKERS: variety of other pypy things I want to have fun with
+May 11 17:03:39 <hpk> LAST: iceland sprint, tb meetings, planning for 0.9, non-pypy stuff
+May 11 17:03:39 <hpk> NEXT: some py lib stuff, non-pypy
+May 11 17:03:39 <hpk> BLOCKERS: -
+May 11 17:03:48 <stakkars> DONE: work on Stackless topics, scheduling implementation tasks, designing, lots of sprint troubles
+May 11 17:03:48 <stakkars> NEXT: more work on that
+May 11 17:03:48 <stakkars> BLOCK: None
+May 11 17:03:50 <arre> PREV: Recover after Tokyo sprint, translation of rctypes.
+May 11 17:03:50 <arre> NEXT: More rctypes, maybe some JIT
+May 11 17:03:50 <arre> BLOCKERS: -
+May 11 17:03:52 <antocuni> LAST: rstr refactoring, lot of thesis-related work
+May 11 17:03:54 <antocuni> NEXT: more rstr recatoring
+May 11 17:03:55 <antocuni> BLOCKERS: none
+May 11 17:04:11 <arigo> LAST: multimethod table compression (pypy-c now smaller); prepare
+May 11 17:04:11 <arigo> thread cloning
+May 11 17:04:11 <arigo> NEXT: thread cloning
+May 11 17:04:11 <arigo> BLOCKERS: time
+May 11 17:04:30 <mwh> LAST: holiday, catching up, new toy, discussion
+May 11 17:04:30 <mwh> NEXT: thread cloning?
+May 11 17:04:30 <mwh> BLOCKERS: -
+May 11 17:05:04 <hpk> ok, pedronis?
+May 11 17:05:21 <pedronis> LAST: worked on gc issues and stackless transform with Armin, worked on rctypes stuff with arre, various non-coding stuff
+May 11 17:05:23 <pedronis> NEXT: will see, we need to work on the OOPSLA paper too
+May 11 17:05:24 <pedronis> BLOCKERS: too much stuff
+May 11 17:05:43 <hpk> ok, thanks
+May 11 17:05:47 <hpk> then let's talk about thread cloning
+May 11 17:05:54 <hpk> i guess everybody is aware of armin's posting to pypy-dev
+May 11 17:06:12 <stakkars> yes, and I'd like to sign it
+May 11 17:06:58 <hpk> you mean that you agree to armin tackling the tasks regarding thread cloning as described?
+May 11 17:07:34 <stakkars> it is pretty much what we descussed, yes
+May 11 17:07:38 <hpk> good
+May 11 17:07:47 <stakkars> but I've read the logic design paper, and I'm no longer convinced that this is what they need.
+May 11 17:07:47 <hpk> then there is the still somewhat open part regarding thread pickling
+May 11 17:07:50 --> stedi67 (n=stedi67 at p54BCF4CA.dip.t-dialin.net) has joined #pypy-sync
+May 11 17:07:54 stakkars stedi67 May 11 17:07:57 <hpk> stedi67: hi stephan
+May 11 17:08:04 <stedi67> hi, sorry, I'm a bit late...
+May 11 17:08:15 <hpk> stedi67: we are in thread cloning, but post your activity lines if you like
+May 11 17:08:22 stakkars stedi67 May 11 17:08:24 <stakkars> thread pickling is in the works. I need to talk to Samuele and carry on
+May 11 17:08:33 <hpk> stakkars: can you commit to some time frame?
+May 11 17:09:18 <hpk> arigo: could you grab 'auc' one day and see if the thread cloning approach and it's APIs would really help them? (i think we should make sure)
+May 11 17:09:32 <arigo> yes
+May 11 17:09:57 <stedi67> LAST: set type
+May 11 17:09:57 <stedi67> NEXT: stackless, greenlet python module (app level)
+May 11 17:09:57 <stedi67> BLOCKERS: None
+May 11 17:10:09 <hpk> arigo: or maybe forward the pypy-dev mail with some extra explanation to Aurelien and Nicolas to be sure
+May 11 17:10:21 <stakkars> I can't speak for eric but I think the problematic part is on my side
+May 11 17:10:28 <arigo> hpk: ok
+May 11 17:10:55 stakkars stedi67 May 11 17:11:17 <stakkars> I would love to get basic things working next week, but it is safer to envision "during the sprint" in June
+May 11 17:11:33 <ericvrp> stakkars, I'm sure your part is harder but I must admit I'm struggling too because it involves a lot of PyPy that I am unfamiliar with. I would like a piece of Samuele too :) (preferably his brain)
+May 11 17:12:21 <stakkars> ericvrp: we can talk about this, and I've been through very miuch during my stay, probably I can play a bit of Samuele
+May 11 17:12:47 <stakkars> (saying that I'm very much available now)
+May 11 17:12:49 <pedronis> FYI: I'm thinking of takign a vacation the week of the iceland sprint
+May 11 17:13:12 <ericvrp> stakkars, okay , good to know. Be prepared for a load of questions tomorrow!
+May 11 17:13:13 <stakkars> good idea. in Iceland? :-D
+May 11 17:13:24 <pedronis> no, I said vacation
+May 11 17:13:28 <mwh> :)
+May 11 17:13:33 <arigo> well, wild landscapes etc.
+May 11 17:14:00 <hpk> stakkars: but it means that you will work on having it finalized during the DDorf sprint then, right?
+May 11 17:14:07 <hpk> thread pickling i mean
+May 11 17:14:09 <pedronis> kind of sort more at home likely
+May 11 17:14:22 <arigo> pedronis: sure, I know (just kidding)
+May 11 17:14:50 <stakkars> this is the plan, but I don't know how realistic it is
+May 11 17:15:10 --> rxe (n=rxe at 66.151.59.5) has joined #pypy-sync
+May 11 17:15:17 <rxe> Hi!
+May 11 17:15:25 <arigo> hi!
+May 11 17:15:32 <hpk> stakkars: sidenote: we need to determine the status of WP07 and whether you can majorly care for the an EU report finalization in June (which is promised to the EU currently)
+May 11 17:15:33 <stakkars> wahaaa
+May 11 17:15:38 <rxe> Thought I show face for a change! ;-)
+May 11 17:15:42 <hpk> rxe: hi richard :)
+May 11 17:15:46 <hpk> rxe: hehe
+May 11 17:15:52 stakkars stedi67 May 11 17:15:53 <mwh> hi rxe
+May 11 17:16:03 <ericvrp> hee richard! welcome
+May 11 17:16:10 <stakkars> hpk: understood
+May 11 17:16:13 <hpk> stakkars: it's not impossible to consider shifting it, but it also relates to the overall project situation and we need to get clarity
+May 11 17:16:17 <arigo> rxe: any three liners?...
+May 11 17:16:51 <rxe> BLOCKERS: no time ;-)
+May 11 17:16:53 <hpk> ok, then for now, i think we have enough clarity mostly
+May 11 17:17:03 <hpk> and i'd like to move on to the "next sprint" topic
+May 11 17:17:09 <arigo> rxe: ah, no, we all posted that already :-)
+May 11 17:17:16 <rxe> :-)
+May 11 17:17:35 <hpk> as you read there is the suggestion for 2nd-9th june in DDorf
+May 11 17:17:42 * arigo agrees
+May 11 17:18:08 <hpk> antocuni, ericvrp, arre, rxe, stedi67: i think you are the ones who didn't participate in this suggestion so far
+May 11 17:18:16 <hpk> do you know already if you can likely make it?
+May 11 17:18:32 <antocuni> I don't know yet
+May 11 17:18:38 <arre> It will work for me.
+May 11 17:18:39 <rxe> rxe: very unlikely:-(
+May 11 17:18:46 <antocuni> but I hope yes
+May 11 17:18:56 <ericvrp> Maybe for a very short while, (it's ony 1.5 hours from my hometown by train)
+May 11 17:18:56 <hpk> i also know from Maciej (the JS SOC guy) that he will likely be able to come
+May 11 17:19:13 <ericvrp> hpk, that's why I would at least like to meet hime
+May 11 17:19:29 <hpk> antocuni: i really hope so as well! if you need support for travel costs or so, mention it to us (e.g. me) or so
+May 11 17:19:53 <hpk> ericvrp: yes, you should try to make sure to both appear at the same time if you do
+May 11 17:20:08 <ericvrp> what a good idea :)
+May 11 17:20:08 <hpk> but also, we will have a number of release and some other critical tasks
+May 11 17:20:11 <stakkars> obvious
+May 11 17:20:20 <antocuni> I was wondering how much does it costs by train
+May 11 17:20:23 <hpk> ericvrp: yes, brilliant idea, isn't it? :)
+May 11 17:20:49 <mwh> antocuni: it's a pretty long way, i think...
+May 11 17:20:52 <hpk> antocuni: no clue, but i'd think that catching a plane from malpensa to ddorf should be feasible
+May 11 17:20:54 * stakkars grabbing coffee
+May 11 17:21:05 <hpk> antocuni: unless you are like carl a non-flier
+May 11 17:21:25 <antocuni> no, flying is fine for me. It's all about money :-)
+May 11 17:21:27 <hpk> ok, so i think Carl will check out opportunities
+May 11 17:21:40 <antocuni> maybe I can try to get fundings from my university
+May 11 17:21:48 <hpk> i think we should fix the sprint time and almost-fix the place but maybe only fix our travels when that is really known
+May 11 17:22:15 <mwh> the world cup isn't going to make traveling around germany a pain is it?
+May 11 17:22:22 <stakkars> re
+May 11 17:22:28 <hpk> world cup, hum
+May 11 17:22:33 <hpk> when was that agian? :)
+May 11 17:22:42 <mwh> i can't remember
+May 11 17:22:42 <hpk> it isn't an EU deliverable is it?
+May 11 17:22:43 <mwh> :)
+May 11 17:22:51 <mwh> i think it starts during the sprint
+May 11 17:23:01 <hpk> ah ok. i ask lene, she definitely will know this.
+May 11 17:23:10 <hpk> (soccer fan, you wouldn't believe it)
+May 11 17:23:21 <hpk> anyway, we can close up soon i think
+May 11 17:23:27 <ericvrp> I think the WC is end of june
+May 11 17:23:28 <stakkars> that can become a real blocker :-)
+May 11 17:23:31 <mwh> first match on the 9th
+May 11 17:23:37 <hpk> in DDorf?
+May 11 17:23:40 <mwh> no
+May 11 17:23:44 <hpk> puh
+May 11 17:23:46 <hpk> ok
+May 11 17:23:50 <hpk> announcement
+May 11 17:23:53 <hpk> any takers?
+May 11 17:24:04 <hpk> should go out latest saturday i think
+May 11 17:24:16 <mwh> i can do it, i guess
+May 11 17:24:21 <hpk> i can review
+May 11 17:24:28 <mwh> cool
+May 11 17:24:53 <hpk> ok then, i think we are done then. thanks all for coming!
More information about the Pypy-commit
mailing list