Mailman 3 March 2012 - pypy-dev

base class hierarchy for string objects breaks translation
by wlavrijsen＠lbl.gov March 15, 2012

March 15, 2012

Hi, the new "mixin" base classes for the new W_StringObject hierarchy are missing a "_mixin_=True" statement, leading to the error message below. I added those statements (also needed in the mixin base class), but that still was not enough to get the translation much further. Ideas, fix? Thanks, Wim [translation:ERROR] File "/home/wlav/pypy-dev/pypy/pypy/annotation/description.py", line 422, in __init__ [translation:ERROR] "with _mixin_: %r" % (cls,)) [translation:ERROR] … [View More]

3 6

GIL in cpyext extensions
by Stefan Behnel March 14, 2012

March 14, 2012

Hi, I stumbled over the following paragraph in https://bugs.pypy.org/issue1069 """ PyEval_SaveThread() always return NULL in pypy. But I don't know if there is a better value, since the GIL is always released in C code. """ Why should the GIL always be released? There are many C extensions that use it in one way or another to protect their internal state, to implement fast (optimistic) locking, or whatever. In order to be compatible with them, the GIL should be held unless explicitly … [View More]

1 0

mechanism for keeping track of the latest revision on default that is known to translate
by Ronny Pfannschmidt March 14, 2012

March 14, 2012

since sprints regularly cause situations where default doesn't translate, it seems helpful to have a known latest revision that translates the proposed mechanism would be to have a bookmark that tracks that version for default, a good name could be something like "last-known-translatable" that version would be a good bet for starting new branches, since its known to translate i'm not too sure yet how to integrate in the buildbot regards, Ronny

2 1

Re: [pypy-dev] [pypy-commit] pypy py3k: save the source of applevel direct tests in a temporary file: this way, we get nicer tracebacks
by Maciej Fijalkowski March 14, 2012

March 14, 2012

Can we have this on default as well please? On Thu, Feb 9, 2012 at 7:12 AM, antocuni <noreply(a)buildbot.pypy.org> wrote: > Author: Antonio Cuni <anto.cuni(a)gmail.com> > Branch: py3k > Changeset: r52297:5c66b92f0617 > Date: 2012-02-09 15:35 +0100 > http://bitbucket.org/pypy/pypy/changeset/5c66b92f0617/ > > Log: save the source of applevel direct tests in a temporary file: this > way, we get nicer tracebacks > > … [View More]

2 1

Re: [pypy-dev] [pypy-commit] pypy default: add a few unrolls to tuple
by Carl Friedrich Bolz March 13, 2012

March 13, 2012

Hi Maciek, I think this commit should come with corresponding pypy_c tests. Cheers, Carl Friedrich On 03/12/2012 09:46 PM, fijal wrote: > Author: Maciej Fijalkowski<fijall(a)gmail.com> > Branch: > Changeset: r53340:12f4497fa28c > Date: 2012-03-12 13:46 -0700 > http://bitbucket.org/pypy/pypy/changeset/12f4497fa28c/ > > Log: add a few unrolls to tuple > > diff --git a/pypy/objspace/std/tupleobject.py b/pypy/objspace/std/tupleobject.py > --- a/pypy/objspace/… [View More]std/tupleobject.py > +++ b/pypy/objspace/std/tupleobject.py > @@ -6,8 +6,8 @@ > from pypy.rlib.rarithmetic import intmask > from pypy.objspace.std.sliceobject import W_SliceObject, normalize_simple_slice > from pypy.objspace.std import slicetype > -from pypy.interpreter import gateway > from pypy.rlib.debug import make_sure_not_resized > +from pypy.rlib import jit > > class W_AbstractTupleObject(W_Object): > __slots__ = () > @@ -114,6 +114,9 @@ > def mul__ANY_Tuple(space, w_times, w_tuple): > return mul_tuple_times(space, w_tuple, w_times) > > +(a)jit.look_inside_iff(lambda space, w_tuple1, w_tuple2: > + jit.is_constant(len(w_tuple1.wrappeditems)) and > + jit.is_constant(len(w_tuple2.wrappeditems))) > def eq__Tuple_Tuple(space, w_tuple1, w_tuple2): > items1 = w_tuple1.wrappeditems > items2 = w_tuple2.wrappeditems > @@ -126,6 +129,9 @@ > return space.w_False > return space.w_True > > +(a)jit.look_inside_iff(lambda space, w_tuple1, w_tuple2: > + jit.is_constant(len(w_tuple1.wrappeditems)) and > + jit.is_constant(len(w_tuple2.wrappeditems))) > def lt__Tuple_Tuple(space, w_tuple1, w_tuple2): > items1 = w_tuple1.wrappeditems > items2 = w_tuple2.wrappeditems > @@ -137,6 +143,9 @@ > # No more items to compare -- compare sizes > return space.newbool(len(items1)< len(items2)) > > +(a)jit.look_inside_iff(lambda space, w_tuple1, w_tuple2: > + jit.is_constant(len(w_tuple1.wrappeditems)) and > + jit.is_constant(len(w_tuple2.wrappeditems))) > def gt__Tuple_Tuple(space, w_tuple1, w_tuple2): > items1 = w_tuple1.wrappeditems > items2 = w_tuple2.wrappeditems > @@ -161,6 +170,8 @@ > def hash__Tuple(space, w_tuple): > return space.wrap(hash_tuple(space, w_tuple.wrappeditems)) > > +(a)jit.look_inside_iff(lambda space, wrappeditems: > + jit.is_constant(len(wrappeditems))) > def hash_tuple(space, wrappeditems): > # this is the CPython 2.4 algorithm (changed from 2.3) > mult = 1000003 > _______________________________________________ > pypy-commit mailing list > pypy-commit(a)python.org > http://mail.python.org/mailman/listinfo/pypy-commit [View Less]

2 1

Re: [pypy-dev] [pypy-commit] pypy default: relaxing int vs long
by Amaury Forgeot d'Arc March 13, 2012

March 13, 2012

Hi Christian, The changeset 8abd5883fa0c looks wrong to me: http://mail.python.org/pipermail/pypy-commit/2012-March/060785.html diff --git a/pypy/interpreter/test/test_compiler.py b/pypy/interpreter/test/test_compiler.py --- a/pypy/interpreter/test/test_compiler.py +++ b/pypy/interpreter/test/test_compiler.py @@ -794,7 +794,7 @@ def test_tuple_constants(self): ns = {} exec "x = (1, 0); y = (1L, 0L)" in ns - assert isinstance(ns["x"][0], int) + assert … [View More]

2 1

Job
by Watts, Wendy March 12, 2012

March 12, 2012

Hello, my name is Wendy; I am a IT recruiter for vtrIT which is a division of Volt Workforce Technical Solutions located in San Francisco. I have an urgent Senior and Junior Python Engineer positions open for a client located in CA. I am reaching out to you in finding out your status of new opportunity, and if we can speak verbal? We will pay travel expenses. Please job description for your review? Specific's of the positions are as follows: * Job Title - one Sr. Python Engineer … [View More]

1 0

"import signal" fails when running PyPy interpreted
by Stefan Behnel March 12, 2012

March 12, 2012

Hi, I get an ImportError when trying to import the "signal" module in an interpreted PyPy (latest source checkout): """ python pypy/bin/py.py -c 'import signal' [platform:execute] gcc -c -O3 -pthread -fomit-frame-pointer -Wall -Wno-unused -march=core2 -O3 -msse4.1 -pipe -fPIC /tmp/usession-default-4/platcheck_2.c -o /tmp/usession-default-4/platcheck_2.o [platform:execute] gcc /tmp/usession-default-4/platcheck_2.o -pthread -lintl -lrt -o /tmp/usession-default-4/platcheck_2 [platform:execute] … [View More]

4 14

Hello, guys
by haael March 12, 2012

March 12, 2012

I'm new to PyPy. I'm mostly interested in ARM architecture. My ultimate goal is to run Python interpreter on a bare CPU without an operating system, but for now I just want to compile PyPy using arm-eabi-gcc. It does not have to work, much less support anything, but my question is: what is the status of ARM target? There's nothing about it in the documentation. haael

4 4

Why isn't PyPy compiling itself not the same as CPython?
by Christopher Swenson March 11, 2012

March 11, 2012

This may be a known issue, but it's a little worrying. I compiled PyPy with CPython, and I got a different binary than when I compiled it with itself: -rwxrwxr-x 1 swenson swenson 59314173 2012-03-11 17:45 pypy-c* -rwxrwxr-x 1 swenson swenson 59297884 2012-03-11 18:52 pypy-c2 I know that the final check with GCC is that it compiles itself with itself, and checks that it gets itself. Is there some randomness in translate.py? --Christopher

3 3