[00:15] fredrik (fredrik@c83-248-135-181.bredband.comhem.se) joined #pypy.
----- silence for 24 minutes ----- [00:39] _hannes (trcnnijh@dsl-62-220-11-56.berlikomm.net) left irc: "utz utz utz"
----- silence for 21 minutes ----- [01:00] <pedronis> night
[01:00] pedronis (~Samuele_P@c-0f8b70d5.022-54-67626719.cust.bredbandsbolaget.se) left irc: "Chatzilla 0.9.68a [Firefox 1.0.2/20050317]"
----- silence for 1 hr and 1 minute ----- [02:01] ludal (~ludal@lab75-1-81-57-254-81.fbx.proxad.net) left irc: "Download Gaim: http://gaim.sourceforge.net/";
----- silence for 2 hr and 18 minutes ----- [04:19] fredrik (fredrik@c83-248-135-181.bredband.comhem.se) left irc: "http://fredrikj.net";
[04:34] hpk (~hpk@merlinux.de) left irc: Read error: 110 (Connection timed out)
----- silence for 2 hr and 42 minutes ----- [07:16] idnar (mithrandi@idnar.user) left irc: Read error: 110 (Connection timed out)
----- silence for 1 hr and 23 minutes ----- [08:39] idnar (mithy@idnar.user) joined #pypy.
----- silence for 1 hr and 54 minutes ----- [10:33] sanxiyn (~tinuviel@211.104.100.115) joined #pypy.
----- silence for 26 minutes ----- [10:59] hpk (~hpk@merlinux.de) joined #pypy.
[11:05] <sanxiyn> Hello
[11:08] <hpk> hi seo
[11:12] <sanxiyn> Any recommendation for API doc generator for Python?
[11:12] <hpk> i guess no
[11:12] <hpk> although i wanted to look into that in the next weeks
[11:12] <hpk> s/although/but/
[11:13] <sanxiyn> Have a need to generate good looking and voluminous "manual" which will be given to the company and nobody will read :(
[11:13] <sanxiyn> Well, if it can be easily read, that's better.
[11:14] <sanxiyn> Now I am looking at Epydoc, but...
[11:16] <sanxiyn> Doesn't seem to handle encoding/i18n issues well. I need to fix it.
[11:23] ludal (~ludal@logilab.net2.nerim.net) joined #pypy.
----- silence for 24 minutes ----- [11:47] mwh (~user@pc150.maths.bris.ac.uk) joined #pypy.
----- silence for 20 minutes ----- [12:07] sanxiyn (tinuviel@211.104.100.115) left #pypy ("Bye").
----- silence for 36 minutes ----- [12:43] <arigo> hpk: replying to your pypy-dev reply:
[12:43] <arigo> looks like a good idea. Should I make the repository change accordingly?
[12:44] <hpk> sure, go ahead
[12:44] <arigo> ok
[12:44] <hpk> maybe write a short note on pypy-dev in reply or so
[12:44] <arigo> yes.
[12:53] <hpk> arigo: do we send diffs on pypy/testresult somewhere?
[12:54] <arigo> I guess not to pypy-svn, at least -- probably nowhere is fine
[12:54] <hpk> makes sense
[13:02] pedronis (~Samuele_P@c-0f8b70d5.022-54-67626719.cust.bredbandsbolaget.se) joined #pypy.
[13:06] <arigo> ok. feel free to check in lots of test_*.txt files all around http://codespeak.net/svn/pypy/testresult/ :-)
[13:07] <hpk> arigo: did you adapt the quickreport also?
[13:07] <arigo> yes
[13:07] <arigo> it's not clever enough to look into all subdirs for the latest version, or something like that
[13:08] <arigo> you just have to 'cd' somewhere and run quickreport from there.
[13:08] <hpk> ok
[13:11] <pedronis> hi
[13:11] fredrik (fredrik@c83-248-135-181.bredband.comhem.se) joined #pypy.
[13:11] <pedronis> we have test failing for this
[13:11] <arigo> pedronis: hi
[13:12] <pedronis> >>> str(TypeError)
[13:12] <pedronis> 'exceptions.TypeError'
[13:12] <pedronis> this is (C)Python
[13:13] <pedronis> >>>> str(TypeError)
[13:13] <pedronis> "<pypy type 'TypeError'>"
[13:13] <pedronis> this is PyPy
[13:13] <mwh> haha
[13:13] <mwh> i recommend changing the tests...
[13:13] <pedronis> exception are new style, and new-style str===repr also CPython
[13:14] <pedronis> is it what you did for your patch?
[13:14] <pedronis> of course we should probably remove the pypy prefix
[13:14] <pedronis> there's a xxx about that
[13:17] <mwh> yes
[13:23] <pedronis> I'm wondering how much user code int the wild depends on this
[13:24] <mwh> i'm willing to be almost none
[13:25] <mwh> could be wrong, of course
[13:25] <mwh> i'm hoping to change it for python 2.5, so i hope i'm not :)
[13:25] <pedronis> well, maybe there will some uglification of error messages
[13:26] <mwh> i don't think so
[13:26] <pedronis> although an error message using the exc name is pretty much developer oriented
[13:26] <mwh> i think most code prints exc_type.__name__
[13:26] <mwh> i think traceback.py does
[13:27] <pedronis> hoping so
[13:27] <pedronis> of course some tests use "%s" % exc_type
[13:36] <hpk> pedronis: side question: will PEP 340 be possible for jython?
[13:37] stakkars (~tismer@rosine80.inf.fu-berlin.de) joined #pypy.
[13:41] stakkars (~tismer@rosine80.inf.fu-berlin.de) left irc: Client Quit
[13:42] stakkars (~tismer@rosine80.inf.fu-berlin.de) joined #pypy.
[13:43] <pedronis> hpk: yes, it just a rarragement around generators
[13:50] <arigo> it looks similar to greenlet :-) without the ability to jump between nested calls
[13:50] <mwh> well, jumping between nested calls needs stackless or greenlets, right?
[13:50] <stakkars> yes
[13:51] <mwh> almost by definition
[13:51] <stakkars> stakelets or greenless
[13:53] <arigo> yes, though it's interesting to note that {stack,green}lets don't necessarily require new syntax
[13:56] <dialtone> python should invest more time in its standard library rather than in the language itself IMHO
[13:59] <mwh> unfortunately, noone's HO counts for much :/
[13:59] <hpk> arigo: indeed
[14:01] <hpk> integrating greenlets into CPython would be interesting (although i can imagine that this might indeed break jython)
[14:02] <arigo> hpk: that's out of the question, I think, for at least two reasons
[14:02] <arigo> which are the same as for Psyco (less so, though)
[14:02] <arigo> it's too platform-specific and it's too obscure a C hack
[14:03] <arigo> my HO is that Python should integrate the very limited about of software stacklessness that covers 90% of the case: Python function calling Python function
[14:05] <arigo> the trouble is that too many people can understand and "contribute" ideas to this kind of syntax discussions.
[14:06] <pedronis> well, Guido likes language design and is what he does
[14:06] <pedronis> and given that Python never went for maximal expressivity with very few powerful basic blocks
[14:07] <pedronis> but of a balancing act between rigid forms for understandability vs. expresiveness
[14:07] <pedronis> it will go on for a while
[14:07] <arigo> yes
[14:08] <arigo> it makes sense, I guess. Though it would be nice to have some basic primitives available in addition :-)
[14:09] <pedronis> although the this discussion is making clear that the balancing act is getting harder with time
[14:10] Action: arigo finished packing his life in two suitcases
[14:10] <mwh> the current proposal scares me a bit
[14:11] <stakkars> what's going on, what should I read?
[14:11] <pedronis> well we have new syntax form that can hide intent as much as reveal it, also the fact of it being a slight variation on for can be confusing
[14:12] <pedronis> I presume that in some situations with bad naming one cannot know without looking at defs whether looping will happen or not
[14:13] <hpk> i envision code sprinkled with decorators and block statements and getting no clue ...
[14:13] <hpk> also i am slightly surprised that Guido is so new-syntax happy
[14:14] <hpk> at last EuroPython he indicated that 2.5 would not contain syntax changes
[14:14] <stakkars> argh
[14:14] <mwh> hpk: maybe we can ambush him this time :)
[14:15] <hpk> now he seems to contemplate PEP340 and optional typing for 2.5, or am i mistaken?
[14:16] <hpk> IMO it would be good if syntax-changes releases happened every two minor releases (2.4 and 2.6, e.g.)
[14:17] <mwh> i don't think optional typing is in for 2.5
[14:17] <mwh> could be wrong, though
[14:20] dialtone (~dialtone@host111-56.pool80117.interbusiness.it) left irc: "Leaving"
[14:22] pedronis (~Samuele_P@c-0f8b70d5.022-54-67626719.cust.bredbandsbolaget.se) left irc: "Chatzilla 0.9.68a [Firefox 1.0.2/20050317]"
[14:28] dialtone (~dialtone@host111-56.pool80117.interbusiness.it) joined #pypy.
[14:39] <arigo> btw, here is some interesting statistics: the SF tracker for "feature requests"
[14:39] <arigo> contains 13 requests in category "Python Interpreter Core"
[14:40] <arigo> and 46 in "Python Library"
[14:41] Action: stakkars still wading through PEP 340 -- peps that long aren't nice
[14:41] <arigo> 4 in "Parser/Compiler"
[14:42] <arigo> https://sourceforge.net/tracker/reporting/?atid=355470&what=category&span=&period=lifespan&group_id=5470#b
[14:45] <mwh> i think one of the parser/compiler ones is mine :)
[14:46] <arigo> good bye
[14:46] arigo (~arigo@vicky.ecs.soton.ac.uk) left irc: "bye bye"
[14:55] pedronis (pedronis@ratthing-b246.strakt.com) joined #pypy.
----- silence for 27 minutes ----- [15:22] Action: hpk changed the testmap format considerably and hopes it's obvious now
[15:25] <hpk> s/now/obvious/
[15:25] <hpk> s/obvious obvious/obvious enough/
[15:27] _hannes (fdraufdy@dsl-62-220-10-253.berlikomm.net) joined #pypy.
[15:31] <pedronis> hpk: what happened to modifed= ?
[15:31] <hpk> automatically detected
[15:32] <hpk> (i am still thinking about how to make that more obvious, but we had the same problem before)
[15:35] Gromit (~bear@A36c4.a.pppool.de) joined #pypy.
[15:41] Gromit (~bear@A36c4.a.pppool.de) left irc: "Client Exiting"
----- silence for 19 minutes ----- [16:00] idnar (mithy@idnar.user) left irc: "home time"
[16:02] <hpk> pedronis: i have implemented the "megapystone" thingie
[16:02] <hpk> do we even need timeouts in seconds?
[16:02] <hpk> in my WC you can do: py.test -T 100MP
[16:03] <hpk> which means timeout after a time equivalent to running 100 million pystones
[16:03] <hpk> which is 15 minutes on codespeak, for example
[16:03] <pedronis> hpk: I see
[16:04] <hpk> this way we can put timings on the tests at some point
[16:04] <pedronis> I'm concetrated trying to pass test_coercion and it needs changes here and there
----- silence for 19 minutes ----- [16:23] <stakkars> pedronis: hi! DOwe have such a thing like a no-op?
[16:24] <pedronis> not so far
[16:24] <pedronis> hpk: it passes with oldstyle and after changing the test itself and fixing bugs
[16:24] <pedronis> hpk: where should I put the new result
[16:25] <hpk> pypy/lib/test2/
[16:25] <pedronis> not that
[16:25] <stakkars> should I make one, or continue to explicitly remove unneeded ops by hand?
[16:25] <pedronis> the *.txt and *.out
[16:25] <hpk> pedronis: ?
[16:25] <hpk> there is a canonical place
[16:25] <pedronis> mixed@mixed ?
[16:26] <hpk> the idea is that you put it into pedronis@whatever
[16:26] <hpk> but that should have gotten created, now?
[16:26] <hpk> s/now/no/
[16:26] <pedronis> yes
[16:26] <hpk> so you just add this directory with the text file(s)
[16:26] <pedronis> just wondering whether we update mixed@mixed too?
[16:26] <hpk> pedronis: mixed@mixed will go soon
[16:27] <pedronis> ok
[16:27] <hpk> we just didn't want to loose valuable data
[16:27] <pedronis> anyway I have put a True in conftest
[16:27] <hpk> and oldstyle=True
[16:27] <pedronis> yes
[16:27] <hpk> that's fine
----- silence for 19 minutes ----- [16:46] <pedronis> hpk: lib/test2/conftest.py needs updating, is still referring to TestDecl. You get a problem if you try to run all PyPy own tests
[16:47] <hpk> ok, i fix it
[16:47] <hpk> hoo-a
----- silence for 23 minutes ----- [17:10] <pedronis> stakkars: btw, what is the correct way to modify _exceptions.py
[17:11] <pedronis> we fail test_eof for example just because of SyntaxError textual repr
[17:17] <stakkars> didI break something?
[17:17] <pedronis> no, but the things marked as hacks in _exceptions.py
[17:17] <pedronis> are not enough to pass the tests
[17:18] <pedronis> test_eof.py is a small example of that
[17:18] <stakkars> we need to add a few correct implementations of the error reports into the
[17:19] <stakkars> pypy/tool/_enum_exceptions thing that I wrote.
[17:19] <stakkars> because it cannot guessall cases
[17:20] <stakkars> should I interrupt my annotation matching and workon that?
[17:21] <pedronis> not immediately but at some point we need to fix that
[17:21] <pedronis> to pass some of the tests
[17:21] <stakkars> test_eof is a standard test?
[17:21] <pedronis> yes, a standard CPython one
[17:22] <pedronis> and I think some other tests have similar problems
[17:22] <stakkars> I'm wondering how to rewrite the complicated exception handling that ovfcheck implies, into the simple thing we finally want.
[17:23] <stakkars> because it is a function
[17:23] <stakkars> how do I run the cpythontest right now? (I was toolong elsewhere)
[17:24] <hpk> stakkars: what do you mean?
[17:24] <hpk> i mean which directory?
[17:24] <stakkars> how do I test test_eof.py, simply py.py it?
[17:24] <hpk> yes, that's one option
[17:26] <hpk> you can also test_all.py/py.test it but that will want to generate a report
[17:26] <stakkars> I just need the error,nothing else
[17:31] <stakkars> pedronis: ah, I think the syntaxerror is one that I handle rather correctly, but the format seems to be different between versions
[17:32] <stakkars> I generated _exceptions.py with Python 2.3.3, which has slightly different format.
[17:32] <stakkars> maybe you just regenerate it and see what happens.
[17:37] <pedronis> stakkars: no
[17:38] <pedronis> I get the same kind of stuff for SyntaxError
[17:38] <stakkars> I see, it needs an implementation.Anyway, it needs the one for Py 2.3.4
[17:39] Action: stakkars urgently needsto move to a newer Python
----- silence for 19 minutes ----- [17:58] Nick change: pedronis -> pedronis_afk
----- silence for 23 minutes ----- [18:21] <stakkars> pedronis_afk: I implemented SyntaxError__str__ and added it to the generation process.
[18:22] <stakkars> See the end of _enum_exceptions.py
[18:22] <stakkars> can you please run it on your machine and generate the exceptionsinterp?
----- silence for 39 minutes ----- [19:01] idnar (mithrandi@idnar.user) joined #pypy.
[19:15] ludal (~ludal@logilab.net2.nerim.net) left irc: "Download Gaim: http://gaim.sourceforge.net/";
[19:29] Nick change: pedronis_afk -> pedronis
[19:35] <stakkars> pedronis: can you please run _enum_exceptions.py on your machine and regenerate exceptionsinterp? might work
[19:38] <pedronis> stakkars: I'm doing that right now
[19:40] <stakkars> pedronis: I'm quite far with that ovfcheck thing,after I got familiar with it.
[19:41] <stakkars> one quite hackish detail is there to find the Overflow check block in the chain of possilities after the function call,
[19:41] <stakkars> and then topatch to no longer say "failed to implement", but the error message from rarithmetic.py :-)
[19:42] <stakkars> I'm wondering if this is so nice, but well, it works and has lots of asserts
[19:45] <pedronis> stakkars: the change mostly work
[19:45] <pedronis> there was a problem with buffer not always being defined, I fixed that
[19:46] <pedronis> now I get
[19:46] <pedronis> PyPy: EOF while scanning triple-quoted string (<string>, line 1)
[19:46] <pedronis> vs.
[19:46] <stakkars> thanks, I was too deeply intoflowgraphs
[19:46] <pedronis> CPython: EOF while scanning triple-quoted string (<string>, line 1)
[19:46] <pedronis> oops
[19:47] <pedronis> CPython: EOF while scanning triple-quoted string (line 1)
[19:47] <pedronis> no <string> in this case
[19:47] <stakkars> well,myimplementation is based upon what CPython 2.3.3does
[19:47] <pedronis> I see
[19:47] <pedronis> I need to understand why the diff happens
[19:47] <stakkars> I'd say keepit this way for now, and I'll fix it next times
[19:48] <stakkars> ah, it also coud be a difference in the creation of the exception.
[19:49] <pedronis> yes
[19:49] <pedronis> it seems so
[19:49] <pedronis> in this case in CPython
[19:49] <pedronis> filename is None
[19:49] <pedronis> I mean
[19:49] <pedronis> synerr.filename
[19:50] <stakkars> so my code is probably ok
[19:51] <pedronis> yes
[19:51] <pedronis> well, start to checkin what we have so far
[19:53] <stakkars> I'm done with what mul does, and now I'm fiddling with _floordiv. Will check in after completion.
[19:54] <stakkars> tedious stuff. I'm tempted to add methods to blocks which insert orremovean operation consitently.
[19:54] <stakkars> but I can stand it and keepkacking :-)
[19:54] <stakkars> hacking (b)
[19:55] <stakkars> have to go,seeyou
[19:58] <pedronis> see you
[20:07] stakkars (~tismer@rosine80.inf.fu-berlin.de) left irc: Read error: 60 (Operation timed out)
----- silence for 46 minutes ----- [20:53] Action: pedronis leaving
[20:53] pedronis (pedronis@ratthing-b246.strakt.com) left irc: "Chatzilla 0.9.67 [Firefox 1.0.2/20050325]"
----- silence for 2 hr and 6 minutes ----- [22:59] arigo (~arigo@c-51c6e055.1321-1-64736c11.cust.bredbandsbolaget.se) joined #pypy.
----- silence for 21 minutes ----- [23:20] arigo (~arigo@c-51c6e055.1321-1-64736c11.cust.bredbandsbolaget.se) left irc: "[x]chat"
[00:00] --- Sat Apr 30 2005