[02:51] _hannes (fdraufdy@dsl-62-220-10-253.berlikomm.net) left irc: "utz utz utz"
----- silence for 1 hr and 16 minutes ----- [04:07] fredrik (fredrik@c83-248-135-181.bredband.comhem.se) left irc: "http://fredrikj.net";
----- silence for 2 hr and 49 minutes ----- [06:56] jacob22|office (jacob@enzo.strakt.com) left irc: Connection timed out
----- silence for 40 minutes ----- [07:36] dialtone (~dialtone@host111-56.pool80117.interbusiness.it) left irc: "This computer has gone to sleep"
----- silence for 4 hr and 37 minutes ----- [12:13] dialtone (~dialtone@host76-1.pool21345.interbusiness.it) joined #pypy.
----- silence for 1 hr and 6 minutes ----- [13:19] dialtone (~dialtone@host76-1.pool21345.interbusiness.it) left irc: "Leaving"
[13:21] yuuh (bvvphff@dsl-62-220-10-104.berlikomm.net) joined #pypy.
[13:23] dialtone (~dialtone@host76-1.pool21345.interbusiness.it) joined #pypy.
----- silence for 16 minutes ----- [13:39] dialtone (~dialtone@host76-1.pool21345.interbusiness.it) left irc: Remote closed the connection
[13:39] dialtone (~dialtone@host76-1.pool21345.interbusiness.it) joined #pypy.
----- silence for 1 hr and 24 minutes ----- [15:03] arigo (~arigo@c-3a8b70d5.022-54-67626719.cust.bredbandsbolaget.se) joined #pypy.
----- silence for 1 hr and 34 minutes ----- [16:37] cfbolz (~bolz@zenon.physi.uni-heidelberg.de) joined #pypy.
[16:37] <cfbolz> hi!
[16:45] <arigo> hi!
[16:46] <cfbolz> are you in sweden now?
[16:46] <arigo> yes
[16:47] <cfbolz> and allready working :-)
[16:47] <arigo> :-)
[16:53] <cfbolz> there is a bug in pypy's nested scopes which keeps a test in test_scope from succeeding.
[16:53] <cfbolz> if you have a function like
[16:53] <arigo> you
[16:53] <arigo> uh
[16:53] <cfbolz> def f(x):
[16:53] <cfbolz> def g():
[16:53] <cfbolz> return x, locals()
[16:53] <cfbolz> return g
[16:53] <cfbolz> and do f(1)() in PyPy you get
[16:53] <cfbolz> (1, {})
[16:53] <cfbolz> and (1, {'x': 1}) in CPython
[16:53] <arigo> I see
[16:53] <cfbolz> (could be a problem of locals, of course
[16:53] <cfbolz> )
[16:54] <cfbolz> but it prevents eval("x") from working in the inner function, too
[16:55] <arigo> note that eval("x") won't work in CPython either unless you use x explicitely too
[16:55] <cfbolz> right. but it won't work in PyPy even after doing that
[16:56] <arigo> of course we have a test for that in PyPy
[16:56] <arigo> test_nestedscope.py
[16:57] <arigo> but it tests the wrong thing, i.e. that 'x' is not in locals :-(
[16:57] <arigo> sometimes it would be nice to check that our pypy tests pass on top of cpython too
[16:57] <cfbolz> right
[16:58] <cfbolz> that's what CPython's tests are good for :-)
[16:58] <arigo> yes :-)
[16:59] yuuh (bvvphff@dsl-62-220-10-104.berlikomm.net) left irc: "utz utz utz"
[17:08] <hpk> hi
[17:08] <hpk> arigo: today is the last day for ep talk registration, right?
[17:08] <cfbolz> argh! I read the sentence without the "talk" :-)
[17:09] <hpk> :-)
[17:10] <arigo> hi!
[17:10] pedronis (~Samuele_P@c-3a8b70d5.022-54-67626719.cust.bredbandsbolaget.se) joined #pypy.
[17:11] <arigo> oups.
[17:11] <hpk> unless samuele has insider knowledge
[17:11] <hpk> or you
[17:14] <arigo> hpk: we're getting messy problems with (guess) running tests
[17:14] <hpk> what specifically?
[17:14] <hpk> i mean with test process or with the actual code they are testing?
[17:15] <arigo> with the file layout, basically
[17:15] <arigo> putting modified stuff in lib and lib/test2 doesn't always work
[17:15] Action: hpk issues --verbose
[17:15] <arigo> e.g. for doctests you have the following problem:
[17:16] <arigo> doctest modules contain the following code:
[17:16] <arigo> if __name__ == '__main__':
[17:16] <arigo> from test import test_myself
[17:16] <arigo> test_support.run_doctest(test_myself)
[17:16] <hpk> right
[17:16] <arigo> so even if you put a modified version of lib/test2, it imports the original version from lib-python-2.3.4/test
[17:16] <hpk> yes
[17:16] <arigo> argh.
[17:17] <hpk> that's one of the reasons i wanted to move to the simpler runtime model of keeping mods in the same tree
[17:17] <hpk> as the original
[17:17] <hpk> i discussed it with pedronis the other night
[17:17] <arigo> yes, he told me so
[17:17] <arigo> about needing either that or a full path-based override.
[17:18] <arigo> but basically both solutions would be ok.
[17:18] <hpk> "path-based override"?
[17:18] <arigo> a lib-python-2.3.4-modified, and a test subpackage using the __path__ hack and only contains the modified files and gets the rest from lib-python-2.3.4
[17:19] <hpk> ah right
[17:19] <arigo> http://www.python.org/dev/doc/devel/lib/module-pkgutil.html
[17:20] <arigo> not sure exactly if we'll get more problems, but that seems to be an "official" hack.
[17:21] <arigo> anyway we really need these modified files out of the way of pypy/lib/
[17:21] <hpk> yes
[17:21] <arigo> I wanted to write a few basic tests for our struct.py and can't because there is already a modified test_struct.py from CPython
[17:23] <hpk> hum
[17:23] <hpk> you wanted to write tests in which way, running against CPython, PyÄPy and in regrtest-stale or in py-test style?
[17:23] <arigo> in py.test style, run against CPython, with no conftest.py at all basically
[17:24] <arigo> if we can get rid of the conftest.py of lib/test2 I'd be very happy :-)
[17:25] <hpk> me too
[17:26] Action: hpk has a double-strange problem with the apache install on codespeak, don't just do 'recoversvn' before asking me ...
[17:26] <hpk> ok, so what if we do:
[17:27] <hpk> pypy-dist/lib/2.3.4 # completely unmodified
[17:27] <hpk> pypy-dist/lib/2.3.4-modified # containing both modified modules and tests
[17:27] <hpk> pypy-dist/lib/conftest.py # doing regr-testing
[17:28] <hpk> s/modified/pypy/
[17:29] <hpk> 2.3.4-pypy (instead of modified) would also contain modules reimplemented in python
[17:29] <arigo> ok
[17:29] <arigo> no
[17:29] <arigo> I agree with everything but this last comment
[17:30] <arigo> the point is that we want to keep our own lib/ and lib/test2/ for our own stuff
[17:30] <arigo> that's the whole point of this reorganization, actually
[17:31] <arigo> to have lib/test2/ available without a conftest.py in the way
[17:31] <hpk> apart from pypy/conftest.py i guess :-)
[17:31] <arigo> yes :-)
[17:35] <hpk> ok, i think it could work
[17:36] <hpk> i could tackle that refactoring after i got the apache-problem solved, hopefully this evening
[17:37] <arigo> :-)
[17:39] <hpk> this apache problem escapes me so far ... gentoo sucks :^)
[17:40] <cfbolz> why is a property of gentoo?
[17:42] <hpk> well, in fact i enjoy gentoo a lot for server platforms
[17:42] <hpk> but there is a really strange build problem with apache which i am not enjoying at all
[17:42] <cfbolz> I never tried it (no dsl at home).
[17:44] <cfbolz> arigo: test_scope passes now.
[17:44] <arigo> cfbolz: good
[17:51] <thingie25> Are greenlets used at all in pypy? Or are they just in there as in interesting future thing to use?
[17:53] <arigo> greenlets are really part of the py lib, not pypy
[17:54] <arigo> they work nicely on CPython, and not in PyPy -- they are a C extension module
[17:54] <hpk> thingie25: out of curiousity: how did you come to the idea that they belong to pypy?
[17:56] <thingie25> I assumed since greenlets and stackless preceded pypy, then greenlets were going to be part of stackless, then stackless and psyco were said to be predecessors to pypy.
[17:56] <thingie25> So assumed greenlets were a part of pypy and therefore the py module was.
[17:56] <hpk> ah ok, in some sense that is right
[17:57] <thingie25> Can I ask a greenlet question here Armin? Or is that off-topic?
[18:05] cfbolz (~bolz@zenon.physi.uni-heidelberg.de) left irc: "Leaving"
[18:06] cfbolz (~cfbolz@edv01.physi.uni-heidelberg.de) joined #pypy.
[18:06] <arigo> thingie25: sure, go ahead
[18:10] <thingie25> What is the intent of including greenlets in the py lib? Are they used by it? Are there plans to write swanky stuff based off them in there?
[18:11] <arigo> no
[18:11] <arigo> the py lib is partly a collection of unrelated pieces
[18:12] <arigo> though maybe execnet could optionally work with greenlets, or offer a "distributed greenlet" model
[18:13] <arigo> (sorry, leaving now... see you later)
[18:13] arigo (~arigo@c-3a8b70d5.022-54-67626719.cust.bredbandsbolaget.se) left irc: "[x]chat"
[18:23] <cfbolz> hpk: how should I mark tests that are modified only for speed reasons? I allready asked you this but now modified is gone and I can't just do modified="speed" anymore.
[18:23] <hpk> just note it in a comment
[18:24] <cfbolz> ok
[18:29] pedronis (~Samuele_P@c-3a8b70d5.022-54-67626719.cust.bredbandsbolaget.se) left irc: Read error: 110 (Connection timed out)
[18:35] dialtone (~dialtone@host76-1.pool21345.interbusiness.it) left irc: "This computer has gone to sleep"
----- silence for 17 minutes ----- [18:52] fredrik (fredrik@c83-248-135-181.bredband.comhem.se) joined #pypy.
----- silence for 38 minutes ----- [19:30] <cfbolz> any ideas why I get a failing assertion in sre_compile.py when using a regexp?
[19:30] <cfbolz> assert _sre.MAGIC == MAGIC, "SRE module mismatch"
[19:31] <cfbolz> in pypy of course
[19:45] dialtone (~dialtone@host111-56.pool80117.interbusiness.it) joined #pypy.
----- silence for 40 minutes ----- [20:25] <cfbolz> going home. see you.
[20:30] cfbolz (~cfbolz@edv01.physi.uni-heidelberg.de) left irc: "Leaving"
----- silence for 1 hr and 54 minutes ----- [22:24] jacob22|office (jacob@enzo.strakt.com) joined #pypy.
----- silence for 32 minutes ----- [22:56] pedronis (~Samuele_P@c-3a8b70d5.022-54-67626719.cust.bredbandsbolaget.se) joined #pypy.
----- silence for 20 minutes ----- [23:16] pedronis (~Samuele_P@c-3a8b70d5.022-54-67626719.cust.bredbandsbolaget.se) left irc: "Chatzilla 0.9.68a [Firefox 1.0.2/20050317]"
[00:00] --- Sun May 1 2005