[00:00] <stakkars> bye, good nite
[00:17] cfbolz (~cfbolz@hdlb-d9b9469b.pool.mediaWays.net) left irc: Read error: 113 (No route to host)
----- silence for 23 minutes ----- [00:40] aleale (~redorlik@cpe.atm0-0-0-129140.0x3ef2fa3a.bynxx3.customer.tele.dk) left irc: "*monkey grin*"
[00:49] arigo (~arigo@c-3a8b70d5.022-54-67626719.cust.bredbandsbolaget.se) left irc: Read error: 60 (Operation timed out)
----- silence for 1 hr and 6 minutes ----- [01:55] pedronis (~Samuele_P@c-3a8b70d5.022-54-67626719.cust.bredbandsbolaget.se) left irc: "Chatzilla 0.9.68a [Firefox 1.0.2/20050317]"
----- silence for 4 hr and 34 minutes ----- [06:29] yuuh (zkzplpel@i3ED6B62D.versanet.de) left irc: "utz utz utz"
----- silence for 4 hr and 12 minutes ----- [10:41] arigo (~arigo@c-3a8b70d5.022-54-67626719.cust.bredbandsbolaget.se) joined #pypy.
----- silence for 1 hr and 33 minutes ----- [12:14] Action: hpk_ starts playing with refactoring the website a bit (to move towards showing the documentation/website contents)
[12:15] Nick change: hpk_ -> hpk
[12:15] <hpk> arigo: morning
[12:16] <arigo> morning
[12:16] <hpk> do you have any comments/suggestions to offer regarding issue1 (download/website)
[12:16] <hpk> ?
[12:17] idnar (mithrandi@idnar.user) left irc: Nick collision from services.
[12:17] idnar_ (mithrandi@idnar.user) joined #pypy.
[12:19] hpk (~hpk@merlinux.de) left irc: Read error: 104 (Connection reset by peer)
[12:21] hpk (~hpk@merlinux.de) joined #pypy.
[12:22] <arigo> "communication" isn't a very common name to look for e-mail and mailing lists
[12:22] <hpk> yes
[12:22] <hpk> do you have a suggestion?
[12:22] <arigo> not really
[12:22] <arigo> "contact"?
[12:23] <hpk> maybe
[12:23] <hpk> i thought about this as well
[12:24] <hpk> let's go with that for the time being (it's at least better than 'list')
[12:25] <hpk> but otherwise it makes sense to basically let more website content come from from pypy/documentation?
[12:28] <arigo> definitely
[12:34] <hpk> hum, i guess i'd really like to clean things up regarding the website ...
[12:34] <hpk> does it make sense right now if i spend the day on that?
[12:46] pedronis (pedronis@ratthing-b246.strakt.com) joined #pypy.
----- silence for 25 minutes ----- [13:11] lac (~lac@c-51c6e055.1321-1-64736c11.cust.bredbandsbolaget.se) joined #pypy.
[13:23] rxe (~rxe@host-212-158-203-122.bulldogdsl.com) joined #pypy.
[13:23] <rxe> Hi
[13:24] <hpk> rxe: hi richard!
[13:24] <hpk> how are things?
[13:24] <rxe> Hi! things are good
[13:24] <rxe> you?
[13:24] <rxe> Did you get a chance to look at the bio?
[13:24] <hpk> fine, especially since some sun showed up
[13:24] <rxe> :-)
[13:25] <hpk> rxe: sorry, not yet, we started the IRC session yesterday
[13:25] <hpk> did you see https://codespeak.net/issue/pypy-dev/ ?
[13:25] <rxe> oh yeah - you are doing a IRC sprint right?
[13:25] <hpk> right, although we are probably/hopefully more working from the tracker
[13:25] <hpk> if we discussed all the issues here in the IRC constantly we wouldn't get any work done, i guess :-)
[13:26] <rxe> cool - will check it out
[13:26] <rxe> :-) Can be a distraction!
[13:26] <hpk> so you are back in fine old scotland?
[13:26] <rxe> sure am. Nice to know I have a bed to go to each night!
[13:26] <hpk> :-)
[13:27] <arigo> hi :-)
[13:27] <hpk> but i guess you also enjoyed your US travels
[13:27] <rxe> hi armin
[13:27] <arigo> nice to "see" you again
[13:27] <rxe> hpk: yeah it was great fun
[13:28] <hpk> after we get the impara/sqeuak people on board and integrated the stuff with PyPy we can "visually" meet :-)
[13:28] <rxe> arigo: :-) it has been a while... so much has happened with pypy since I last checked
[13:29] <arigo> hpk: yes I think cleaning up things related to the website make sense (sorry for the delay)
[13:29] <arigo> can't stay much longer now, leaving for the office via lunch...
[13:29] <rxe> I am going to have look at making the flow tracer working again. Exceptions in the flow graph have changed recently?
[13:29] <hpk> arigo: ok, i am first trying to just hack it and do the real refactoring after M0.5
[13:30] <hpk> arigo: see you
[13:30] <arigo> rxe: oups right yes.
[13:30] Action: arigo adds an issue "should document the change to the flow model's exceptions"
[13:30] <hpk> arigo: try to take care not to make too long titles btw
[13:30] <arigo> :-)
[13:31] <hpk> :-)
[13:31] <hpk> rxe: side note: you can simply use your rxe login with the tracker
[13:31] <hpk> (i mean the codespeak/subversion one)
[13:32] arigo (~arigo@c-3a8b70d5.022-54-67626719.cust.bredbandsbolaget.se) left irc: Remote closed the connection
[13:32] <rxe> hpk: cool that worked.
[13:33] <rxe> i found a few bugs over the weekend after running some of my old code via py.py - guess I can add them there?
[13:35] <hpk> yes
[13:36] <hpk> you can set 'notify on new issues' in 'your details' to receive all top level new issues irrespective of nosy settings, btw
[13:39] <rxe> that's handy. thanks :-)
[13:39] <hpk> and you can generally reply by email
[13:40] <hpk> plus in svn-commits you can reference issues, see https://codespeak.net/issue/pypy-dev/issue1
[13:40] <hpk> the key line is the 'issue1 resolved' which connects the checkin message to the issue
[13:40] <hpk> um, 'issue1 in-progress' actually
[13:42] <rxe> hpk: traceinteractive.py was merged with py.py at last sprint IIRC
[13:43] <hpk> rxe: ups, missed that
[13:44] <hpk> but how can you use it?
[13:45] <rxe> Sorry - I need to write some docs. Use __pytrace__ = 1 on the command line, to turn off set to 0.
[13:46] <hpk> cool!
[13:46] <hpk> i still think it's showing to many space-ops by default :-)
[13:47] <rxe> :-) it is configurable... although not obvious. I added a feature last night to turn off seeing bytecode
[13:49] <hpk> the default should not show any method_w's IMO
[13:51] <rxe> yeah - I agree.
----- silence for 29 minutes ----- [14:20] <rxe> ive added some issues. It is very cool.
[14:20] <rxe> Coffee is finished - see you.
[14:20] rxe (rxe@host-212-158-203-122.bulldogdsl.com) left #pypy.
[14:24] Nick change: idnar_ -> idnar
[14:24] <pedronis> hpk: I'm not sure having news (like upcoming sprints etc) one click away from the homepage is a good idea
[14:25] <hpk> you mean it shouldn't be a menu item?
[14:26] <pedronis> no the news should be on the homepage
[14:27] <pedronis> at least the most recent ones
[14:31] <hpk> so you liked the old situation better?
[14:31] <hpk> i thought that we want to have a really high level home page
[14:31] <hpk> explaining the EU context and some higher level goals
[14:31] <hpk> (e.g. mentioning psyco and stackless is not really helpful to squeak-devers)
[14:32] <pedronis> hpk: yes, but sites really tend to put news to on their homepages
[14:32] <pedronis> otherwise it gets very static
[14:33] <pedronis> at a minimum in some sidebar
[14:37] arre (ac@ratthing-b3fa.strakt.com) joined #pypy.
[14:43] <hpk> pedronis: like this maybe? http://codespeak.net/pypy/index.cgi?home
[14:44] <hpk> (the news menu item would go then or maybe 'home' renamed to 'news' and become the news page)
[14:45] <pedronis> yes, something like that. May point is simply that we want both a short intro and recent news on the homepage, and more older news and a possibly longer intro somewhere else
[14:45] <hpk> ok
[14:45] <hpk> i'd like to rename 'home' to 'news' though
[14:46] <hpk> because i'd like to make it more obvious that it makes sense to check back on that page every now and then
[14:46] <pedronis> ok
[14:57] arigo (~arigo@ratthing-b407.strakt.com) joined #pypy.
----- silence for 27 minutes ----- [15:24] Action: arigo reviews listobject.py and dictobject.py
----- silence for 53 minutes ----- [16:17] yuuh (aaxqzrjv@i3ED6B62D.versanet.de) joined #pypy.
----- silence for 28 minutes ----- [16:45] <lac> hpk: I am not sure that it makes sense to have the high-level page as the default 'start-here' page.
[16:45] <lac> for certain kinds of readers,. this is exactly what they want.
[16:45] <lac> for other kinds of readers, the whole EU context is precisely what they want filtered out.
[16:46] <lac> Do we want everybody to start at the same place?
[16:47] <lac> oops, I am now getting list out of ranger render errrors for http://codespeak.net/pypy/
[16:48] <lac> File "/projects/pypy/www/pypywww/render/page.py", line 62, in render_contentspace
[16:48] <lac> self.content()
[16:48] <lac> File "/projects/pypy/www/pypywww/render/main.py", line 25, in content
[16:48] <lac> self.set_redirect(self.newurl)
[16:48] <lac> File "/projects/pypy/www/pypywww/render/page.py", line 22, in set_redirect
[16:48] <lac> head = self.gethead()
[16:48] <lac> File "/projects/pypy/www/pypywww/render/page.py", line 19, in gethead
[16:48] <lac> head = root[1]
[16:48] <hpk> lac: just retry
[16:49] <lac> error None type has no method append
[16:49] <hpk> wait a moment
[16:50] <lac> ok
[16:54] <hpk> lac: should work again
[16:55] <hpk> next time please just tell me there is a problem and don't spam with tracebacks :-)
[16:55] <lac> ok
[16:57] <lac> Holger: there is no 'h' in Göteborg.
[16:57] <lac> :-)
[16:57] <hpk> you can fix that now in pypy/documentation/website :-)
[16:57] <lac> cool.
[16:58] <hpk> don't forget to run py.test or 'pypy/test_all.py' in the documentation directory
[16:58] <hpk> it will perform a rough integrity check
[16:58] Action: hpk is out into the sun ...
[17:09] stakkars_ (~tismer@i3ED6B62D.versanet.de) joined #pypy.
[17:16] stakkars (fgmwlnvi@i3ED6B62D.versanet.de) left irc: Read error: 110 (Connection timed out)
[17:16] stakkars_ (~tismer@i3ED6B62D.versanet.de) left irc: Read error: 104 (Connection reset by peer)
[17:31] yuuh (aaxqzrjv@i3ED6B62D.versanet.de) left irc: Read error: 131 (Connection reset by peer)
[17:46] arre (ac@ratthing-b3fa.strakt.com) left irc: "using sirc version 2.211+KSIRC/1.3.11"
----- silence for 16 minutes ----- [18:02] stakkars (~tismer@i3ED6B62D.versanet.de) joined #pypy.
[18:02] <stakkars> arigo: hi!
[18:03] <stakkars> there is a bug in your check-in from yesterday night: typedef has a bug around line 300
[18:03] <arigo> hi
[18:03] <arigo> has it?
[18:03] <stakkars> def descr_get_dict(space, obj):
[18:03] <stakkars> w_dict = obj.getdict()
[18:03] <stakkars> if w_dict is None:
[18:03] <stakkars> typename = space.type(w_obj).getname(space, '?')
[18:04] <stakkars> w_obj is not found as a global,saysflowgraphing :-)
[18:04] <arigo> thanks :-)
[18:04] <stakkars> did you means space.type(obj) ?
[18:04] <arigo> sure
[18:05] <stakkars> so the obj should be renamed to w_obj?
[18:05] <arigo> yes, or the other way around
[18:05] <stakkars> is it from space view, the with w I guess
[18:05] <arigo> yes.
[18:05] <stakkars> did it.
[18:06] <arigo> thanks
[18:06] <arigo> change descr_set_dict too, while you're at it
[18:06] <stakkars> already done.
[18:06] <arigo> :-)
[18:07] <stakkars> I seem to be the one who runs targetpypymain all the time :-)
[18:08] <arigo> :-)
[18:14] <stakkars> I get another bug:
[18:15] <stakkars> AssertionError: [(pypy.objspace.std.fake:81) fake__new__ block@132 op=0] expecte
[18:15] <stakkars> d function, got <type 'dict'>
[18:15] etrepum (bob@ayunami.redivi.com) left #pypy.
[18:16] <stakkars> this is raised in bookkeeper line 325
[18:22] <arigo> stakkars: fixed...
[18:22] <stakkars> the caseof a builtin type wasnot caucht. But I didn't find the origin. The casemust be new
[18:25] <stakkars> you needed to avoid the dict going through. why?
[18:26] <arigo> calling dict is not explicitely supported.
[18:26] <stakkars> calling any type, actually?
[18:27] <arigo> this are supported case-by-case in builtin.py
[18:27] <stakkars> ok
[18:27] <arigo> e.g. int() and float() and str() are there.
[18:31] Action: arigo reassigning issues between "M0.5" and "current" (the latest meaning "not for the release")
[18:35] <stakkars> you think exceptionsautogeninterping is urgent? OK
[18:35] <arigo> stakkars: well, it would be nice to have it in the release, instead of having to document the strange command-line needed to regenerate
[18:36] <stakkars> :-)) this is a very good point
[18:37] <stakkars> I hope flowing will work in this early state. Well, why not, it is a different space.
[18:39] <stakkars> someobjectness: 36 percent
[18:39] <stakkars> with -no-t -no-c
[18:39] <stakkars> is that normal?
[18:40] <arigo> no no no
[18:41] <arigo> Samuele is looking forward to tracking the type pollution :-)
[18:43] <stakkars> I have folded COnstants verymuch (not checked in), so I'm interested toknow whether you see the same.
[18:45] <stakkars> when did that change occour? I'd like to revert to the prior version if possible.
[18:53] <hpk> pedronis: do you want auto-nosying too?
[19:00] <arigo> stakkars: I don't really have a clue
[19:01] <arigo> if you suspect your changes might be the cause, we need to try on another machine
[19:01] <arigo> (Samuele's doing it, there is indeed a problem)
[19:01] <stakkars> no, my changes were not checked in, and I also checked against the original version.
[19:02] <stakkars> not sure yet, it might be the dictproxy thing. Still testing...
[19:11] <stakkars> no it is not.
[19:17] <stakkars> going to find this by binary seach :-)
[19:23] yuuh (wqrffxil@i3ED6B62D.versanet.de) joined #pypy.
[19:29] <stakkars> just a few runs away from it...
[19:32] <arigo> Samuele just isolated the problem to my latest check-in
[19:32] <arigo> about getexecutioncontext() & friends
[19:33] <stakkars> well, the problem was created between revision 12288 and 12302
[19:33] <arigo> more precisely, it's not broken at rev12364
[19:34] <arigo> hum.
[19:34] <stakkars> concerning someobjectness, it was caused by a change yesterday between 15:00 and 18:00
[19:34] <stakkars> 12288 has 15 someobjects, 12302 has 1000
[19:35] <arigo> maybe we fixed and broked it again
[19:35] <stakkars> 15 minutes and I know it :-)
[19:35] <arigo> Samuele seems sure that r12364 is fine again
[19:35] <stakkars> if it doesn't help to know what it was, then I stop wasting CPU time?
[19:38] <arigo> guess so
[19:39] <arigo> yes, r12365 definitely introduced the big explosion we're seeing now
[19:40] <stakkars> definatelynot a trivial checkin.
[19:41] <arigo> indeed
[19:41] <stakkars> just for my current tests, I use one of those
----- silence for 27 minutes ----- [20:08] <stakkars> forget my results, it was at a later time, and it is sitting in the cache
[20:11] <arigo> argh argh argh. found it...
[20:11] <stakkars> ????
[20:12] <arigo> well, the problem is caused by the new way getexecutioncontext() is written
[20:13] <stakkars> can I help?
[20:14] <arigo> we should find a way to rewrite it that avoid the following trap:
[20:14] <arigo> self.__current_ec has a predefined value, so we get a prebuilt ExecutionContext instance
[20:14] <arigo> but the "except AttributeError" is also followed,
[20:15] <arigo> so we return the union of either the prebuilt ExecutionContext, or a newly created one
[20:15] <arigo> which is fine so far, apart from a detail:
[20:15] <arigo> ExecutionContext.__init__ creates a new Stack, for the framestack
[20:15] <arigo> Stacks are specialized by creation point
[20:16] <arigo> so there is no way the *prebuilt* ExecutionContext can figure out it needs the same specialized stack as the dynamically built ExecutionContext
[20:16] <arigo> as a result, the class ExecutionContext gets an *unspecialized* framestack
[20:17] <stakkars> oh
[20:17] <arigo> which is very bad, because manipulation them then influences *all the other* stacks around
[20:17] <arigo> including the valuestack of PyFrame
[20:17] <arigo> result: we get SomeObjects out of the valuestack... and they propagate everywhere.
[20:18] <arigo> at the moment, the "real bug" is detected by a warning: setting "self.__current_ec" is not allowed, because "self" is a constant.
[20:18] <arigo> so we need to rewrite getexecutioncontext() in some way...
[20:25] stakkars_ (yolexnp@dsl-62-220-13-97.berlikomm.net) joined #pypy.
[20:29] yuuh (wqrffxil@i3ED6B62D.versanet.de) left irc: Read error: 60 (Operation timed out)
[20:29] <arigo> we'll reintroduce a threadlocals object, this time stored on the space
[20:30] <arigo> the current ec will be stored there
[20:30] <arigo> then space._freeze_ will make sure it gets removed
[20:30] <arigo> so that the annotator never sees any prebuilt execution context
[20:30] <arigo> this kind of change makes sense anyway if we think about threads in the future.
[20:31] <stakkars_> yes, I agree
[20:33] stakkars (~tismer@i3ED6B62D.versanet.de) left irc: Read error: 145 (Connection timed out)
----- silence for 48 minutes ----- [21:21] arigo (~arigo@ratthing-b407.strakt.com) left irc: "dinner time"
[21:21] pedronis (pedronis@ratthing-b246.strakt.com) left irc: "Chatzilla 0.9.67 [Firefox 1.0.2/20050325]"
----- silence for 32 minutes ----- [21:53] fredrik (fredrik@c83-248-135-181.bredband.comhem.se) joined #pypy.
----- silence for 28 minutes ----- [22:21] yuuh (ndxrgghc@dsl-62-220-13-97.berlikomm.net) joined #pypy.
----- silence for 1 hr and 32 minutes ----- [23:53] fredrik (fredrik@c83-248-135-181.bredband.comhem.se) left irc: "http://fredrikj.net";
[23:54] yuuh (ndxrgghc@dsl-62-220-13-97.berlikomm.net) left irc: "utz utz utz"
[00:00] --- Tue May 17 2005