extradoc / minute / pypy-sync-2005-07-21.txt

pypy-sync developer meeting 21st July 2005

         Anders Chrigstroem, Samuele Pedroni, Adrien Di Mascio, 
         Ludovic Aubry, Carl Friedrich Bolz, Holger Krekel (minutes), 
         Richard Emslie (partly)

with pre-sent info: Anders Lehmann
missing/without info: Christian Tismer, Armin Rigo (still holiday?)

Regular Topics 

- roll call. holger opens the meeting. 

- activity reports (3 prepared lines of info). 
  All Attendees submitted activity reports (see `IRC-Log`_ 
  at the end and 'LAST/NEXT/BLOCKERS' entries in particular)

- resolve conflicts/blockers
  No direct conflicts were discovered. Main "blocker" appears to
  be vacations. 

Topics of the week

Heidelberg Sprint Planning Status 

Carl reports that the sprint room is fixed for the duration
from 22nd to 29th of August at the university of Heidelberg. 
He also started a `Heidelberg sprint announcement`_.  The 
Hildesheim crew should finalize it next week and send it out. 
There is a bit of a problem with network connectivity because
the "Rechenzentrum" mandates VPN-clients.  Carl will check into
that but would enjoy help! 

Accomodation wise there is the possibility to rent a room
from a flat in Heidelberg. That would cost around 300 Euros. 
Richard is interested in that.  Others are thinking about it.
Everbody interested should send Carl a mail, he might even
be able to get two rooms like this. (Hotels are usually
quite expensive in Heidelberg). 

.. _`Heidelberg sprint announcement`:

compliance tests failures after 2.4.1 merge: how/when to tackle?

Holger notices that after the 2.4.1 merge only roughly 50% of the 
compliance `tests are passing`_.   Samuele thinks that fixing them 
requires a lot of work, approximately one week.  The attendees 
decide that the Hildesheim Sprint crew should decide about a 
strategy of how to tackle this problem.  Mails will be sent
to pypy-dev and pypy-funding (because it touches important 
EU issues). 

.. _`tests are passing`:

0.6.2 release: status / when do we think to actually do it? Do we still?

We have interpreter documentation and the parser situation is improving
(running the parser on top of 2.3 yet providing 2.4 semantics). 
The main blocker for the release is the compliance test situation. 
Therefore further discussion/decisions are postponed until we
have a strategy regarding the compliance tests (see previous topic). 

re-assigning tasks (related to NEXT activities)

Holger notes that the hildesheim sprint crew might interfere 
with the Parser (ludal) and os-level implementations (aleale). 
The hildesheim sprint crew should be left free to work on 
the issues as it sees fit.  For people outside the sprint
it is recommended to work in short-commit cycles to avoid
conflicts and to follow the commits from the sprint. 
The sprint group will try to provide info on pypy-dev or
on IRC. 


Holger closes the meeting in time at 13:30pm. 

.. _`IRC-log`: 

Here is the full IRC log:: 

 **** BEGIN LOGGING AT Thu Jul 21 12:19:28 2005
 Jul 21 13:02:38 <hpk>	ok, we'll start anyway, although there appear some people to be missing
 Jul 21 13:02:59 -->	arre ( has joined #pypy-sync
 Jul 21 13:03:18 <cfbolz>	ok
 Jul 21 13:03:35 <hpk>	everbody ready to start? 
 Jul 21 13:03:43 <cfbolz>	yes
 Jul 21 13:03:48 <arre>	I am.
 Jul 21 13:03:59 <pedronis>	yes
 Jul 21 13:04:06 <adim>	I am
 Jul 21 13:04:16 <hpk>	here is the agenda (posted already yetserday to pypy-funding):
 Jul 21 13:04:20 <hpk>	- roll call. holger opens the meeting.
 Jul 21 13:04:20 <hpk>	- activity reports (3 prepared lines of info).
 Jul 21 13:04:20 <hpk>	- resolve conflicts/blockers
 Jul 21 13:04:20 <hpk>	- Heidelberg Sprint Planning Status
 Jul 21 13:04:20 <hpk>	- compliance tests failures after 2.4.1 merge: how/when to tackle?
 Jul 21 13:04:20 <hpk>	- 0.6.2 release: status / when do we think to actually do it? Do we still?
 Jul 21 13:04:20 <hpk>	- re-assigning tasks (related to NEXT activities)
 Jul 21 13:04:46 <hpk>	activity reports: i suggest the following order: hpk,aleale,ludal,adim,pedronis,arre,cfbolz,
 Jul 21 13:05:05 <hpk>	LAST: interpreter documentation, organizing hildesheim sprint, codespeak migration planning, mentoring so
 Jul 21 13:05:05 <hpk>	me os-level commits, fixing py test bugs
 Jul 21 13:05:05 <hpk>	NEXT: Hildesheim-Sprint orga + hacking, EU issues
 Jul 21 13:05:05 <hpk>	BLOCKERS: too many things going on
 Jul 21 13:05:16 ---	You are now known as aleale
 Jul 21 13:05:23 <aleale>	LAST: working on posix (for pypy), started on math
 Jul 21 13:05:23 <aleale>	NEXT: finish posix, and math ? (will try the lib-python tests)
 Jul 21 13:05:23 <aleale>	BLOCKERS: need to understand exceptions better (interpreter level, annotation of, etc), a wedding, an immanent move to Germany
 Jul 21 13:05:28 ---	You are now known as hpk
 Jul 21 13:05:31 <ludal>	last week : parser work, pysymbol and pytoken, switching to using integers for rule names instead of strings
 Jul 21 13:05:31 <ludal>	next week : vacation + eventually resolving annotation problems + astbuilder to make annotation work further
 Jul 21 13:05:31 <ludal>	blocker : vacation ;)
 Jul 21 13:05:42 <adim>	LAST: parser's annotation / astbuilder
 Jul 21 13:05:42 <adim>	NEXT: teach a course
 Jul 21 13:05:42 <adim>	blockers: no real one
 Jul 21 13:05:57 <pedronis>	Last: bug fixing in annotator and rtyper, progress rtyping PyPy adding rtyper features, implemented hook to control rtyping order, integrated suggestion to continue rtyping even after errors
 Jul 21 13:05:59 <pedronis>	Next: sprint (we are down to 57 rtyping problems with the translation snapshot)
 Jul 21 13:06:00 <pedronis>	Issues: tried to annotate the trunk, now annotation finishes but quite some involved SomeObject problems related to parser code
 Jul 21 13:06:47 <hpk>	arre, cfbolz next
 Jul 21 13:07:15 <arre>	Last week: annotation/rtyper work and learning more about their internals.
 Jul 21 13:07:26 <arre>	Next week: Vacation until Heidelberg.
 Jul 21 13:07:35 <arre>	Blockers: None at the moment.
 Jul 21 13:07:39 <cfbolz>	LAST: finalized sprint room, trying to get cheap accomodation, looking for the cheapest publich transport tickets
 Jul 21 13:07:39 <cfbolz>	NEXT: checking network connectivity in the sprint room, Hildesheim sprint
 Jul 21 13:07:39 <cfbolz>	BLOCKERS: not much, still a bit own stuff to do
 Jul 21 13:08:03 <hpk>	ok, there appear to be no conflicts
 Jul 21 13:08:16 <hpk>	except that the hildesheim sprint could provide conflicts
 Jul 21 13:08:24 <hpk>	(e.g. os-level functions or the parser situation) 
 Jul 21 13:08:34 <hpk>	ludal: before you leave to holiday could you send a mail to pypy-dev about the status? 
 Jul 21 13:08:43 <hpk>	(or post it to the issue) 
 Jul 21 13:08:59 <hpk>	this way the hildesheim group knows where it could start to work if it wants to
 Jul 21 13:09:23 <ludal>	sure, actually I'm not leaving paris so I'll be available online from time to time and hopefully following developpements of the sprint
 Jul 21 13:09:36 <hpk>	ludal: great, ok
 Jul 21 13:09:51 <hpk>	next topic: heidelberg status, carl? 
 Jul 21 13:10:08 <cfbolz>	the room is secured from 22nd to 29th
 Jul 21 13:10:10 <cfbolz>	!
 Jul 21 13:10:22 <cfbolz>	have to check network connectivity: it is possible that we need some sort of VPN client to access the network
 Jul 21 13:10:38 <cfbolz>	I'm trying to find student rooms
 Jul 21 13:10:52 -->	rxe ( has joined #pypy-sync
 Jul 21 13:10:54 <cfbolz>	if there is enough interest
 Jul 21 13:11:08 <hpk>	network connectivity: we should have our own router/machine that makes the VPN transparent if possible
 Jul 21 13:11:28 <cfbolz>	yes, if someone helps me with that
 Jul 21 13:11:52 <cfbolz>	the problem is that the local admins from the rechenzentrum are kind of inflexible
 Jul 21 13:12:11 <cfbolz>	so I would have to do it myself
 Jul 21 13:12:22 <hpk>	cfbolz: i see, can you try to make your own laptop work? 
 Jul 21 13:12:29 <cfbolz>	yes, of course
 Jul 21 13:12:43 <hpk>	i saw you also checked in a sprint announcement
 Jul 21 13:12:49 <hpk>	we should finalize this next week and send it out
 Jul 21 13:12:54 <cfbolz>	indeed
 Jul 21 13:13:14 <hpk>	cfbolz: is the 200-300 euro room still a concrete thing? 
 Jul 21 13:13:17 <cfbolz>	yes
 Jul 21 13:13:26 <hpk>	so let's try to decide here
 Jul 21 13:13:47 <hpk>	cfbolz has a room in a living group (WG) that costs 200-300 euro and two people could sleep there for the sprint week
 Jul 21 13:14:03 <cfbolz>	I could maybe even get two rooms
 Jul 21 13:14:07 <cfbolz>	not sure yet
 Jul 21 13:14:30 <hpk>	is anyone basically interested? (hotels are kind of expensive in heidelberg)
 Jul 21 13:14:44 <rxe>	sounds good 
 Jul 21 13:15:18 <cfbolz>	another question is how we pay for that
 Jul 21 13:15:24 <rxe>	(that means i am interested)
 Jul 21 13:15:37 <hpk>	(rxe: if you want to send your activity reports, just paste your three LAST/NEXT/BLOCKERS lines in)
 Jul 21 13:15:51 <hpk>	arre, pedronis, ludal: anyone of you interested? 
 Jul 21 13:16:28 <rxe>	LAST: dummy Opaque type in llvm2
 Jul 21 13:16:28 <rxe>	NEXT: nothing planned
 Jul 21 13:16:28 <rxe>	BLOCKERS: nothing planned
 Jul 21 13:16:34 <hpk>	cfbolz: i think the people using the room should pay, maybe they can get a receipt? 
 Jul 21 13:17:10 <cfbolz>	ok, I can check that
 Jul 21 13:17:26 <rxe>	well 300 euros seems reasonable priced for one person
 Jul 21 13:17:58 <hpk>	arre, pedronis, ludal: ok, this is lasting too long, write cfbolz a mail if you are interested
 Jul 21 13:17:58 <cfbolz>	you could even stay there longer and visit heidelberg a bit :-)
 Jul 21 13:18:31 <cfbolz>	I guess the hardest HD sprint issues are solved, now that we have the room?
 Jul 21 13:18:39 <ludal>	hpk: sure, can't decide now
 Jul 21 13:18:59 <hpk>	cfbolz: yes, the sprint is fixed 22nd-29th, announcement next week
 Jul 21 13:19:23 <hpk>	next topic: compliance tests failures after 2.4.1 merge: how/when to tackle?
 Jul 21 13:19:50 <hpk>	any opinions on this? (i guess you all noticed that we have only ~50% compliance tests passing)
 Jul 21 13:20:17 <hpk> for reference
 Jul 21 13:20:31 <rxe>	cfbolz: where is nearest airport (sorry - next topic)?
 Jul 21 13:20:34 <--	ludal ( has left #pypy-sync
 Jul 21 13:21:00 <hpk>	rxe: let cfbolz add this info to the announcement
 Jul 21 13:21:01 <cfbolz>	rxe: frankfurt I guess
 Jul 21 13:21:01 -->	ludal ( has joined #pypy-sync
 Jul 21 13:21:15 <rxe>	ok - thanksx
 Jul 21 13:21:23 *	hpk notices 9 minutes left
 Jul 21 13:21:46 <pedronis>	hpk: fixing the tests is a lot of work
 Jul 21 13:22:25 <pedronis>	something around a week
 Jul 21 13:22:59 <hpk>	i suggest that we task the hildesheim sprint crew to discuss this and come up with a strategy when/how to tackle the problem
 Jul 21 13:23:24 <pedronis>	yes, I don't think we can decide something now
 Jul 21 13:23:28 <hpk>	everbody ok with this?
 Jul 21 13:23:31 <cfbolz>	yep
 Jul 21 13:23:34 <rxe>	yes
 Jul 21 13:23:52 <adim>	ok with me
 Jul 21 13:23:53 <arre>	yes
 Jul 21 13:24:02 <ludal>	ok
 Jul 21 13:24:32 <hpk>	ok, we will send according infos to pypy-dev and possibly pypy-funding (it's a critical EU issue in some ways)
 Jul 21 13:24:47 <hpk>	next topic: 0.6.2 release: status / when do we think to actually do it? Do we still?
 Jul 21 13:24:59 <hpk>	this is a follow up on the last week's topic
 Jul 21 13:25:03 <hpk>	interpreter documentation is now mostly there
 Jul 21 13:25:13 <hpk>	the parser is getting there
 Jul 21 13:25:27 <hpk>	ludal: does the parser/compiler now run on top of Python 2.3 providing 2.4 semantics? 
 Jul 21 13:25:48 <hpk>	(my limited manual tests seem to indicate that it does)
 Jul 21 13:26:01 <ludal>	mostly, you'll have to point out if you see problems
 Jul 21 13:26:30 <hpk>	ok, then the main blocker for a 0.6.2 release is the compliance/test situation
 Jul 21 13:27:00 <hpk>	so we should re-discuss it next week (after the hildesheim crew dicussed it) if nobody objects
 Jul 21 13:27:05 <pedronis>	yes
 Jul 21 13:27:16 <adim>	ok
 Jul 21 13:27:42 <cfbolz>	I don't
 Jul 21 13:27:44 <hpk>	ok, then last (somewhat optional) topic: - re-assigning tasks (related to NEXT activities)
 Jul 21 13:28:33 <hpk>	like mentioned earlier, i think that the hildesheim crew might interfer with the os-level calls (aleale) and parser (ludal) developments
 Jul 21 13:29:15 <hpk>	i think that the hildesheim crew should be allowed to work on whatever it wants and try to signal via IRC or pypy-dev/commits
 Jul 21 13:29:16 <ludal>	ok
 Jul 21 13:29:44 <hpk>	so it's especially important to have short commit cycles to avoid conflicts
 Jul 21 13:30:02 <hpk>	(i.e. don't work for three days before checking something in)
 Jul 21 13:30:28 <hpk>	ok, it's 13:30 ... closing time for the meeting
 Jul 21 13:30:34 <hpk>	thanks for coming!
 Jul 21 13:31:09 <cfbolz>	see you!
 Jul 21 13:31:21 <rxe>	bye
 Jul 21 13:31:24 <hpk>	see you!
 Jul 21 13:31:27 <adim>	Bye
 Jul 21 13:32:28 <--	rxe ( has left #pypy-sync
 **** ENDING LOGGING AT Thu Jul 21 13:32:33 2005