Comment #53 on issue 1026 by julien.r...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
Note that due to bugs in previous versions of pypy, we require pypy
2.0.0-beta at least.
--
You received this message because this project is configured to send all
Updates:
Status: Fixed
Comment #52 on issue 1026 by asmeu...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
PyPy has run SymPy just fine for a while now (since before the 0.7.2
release).
--
You received this message because this project is conf
Comment #51 on issue 1026 by asmeu...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
The "__len__ should return bool or integer" error was a PyPy bug and is
fixed in their dev branch (see https://bugs.pypy.org/issue1235).
--
You received this message b
Comment #50 on issue 1026 by ronan.l...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
https://github.com/sympy/sympy/pull/1486
(BTW, it was trivial to find the source of the problem with a
simple 'py.test --pdb sympy/integrals/' run with pypy)
--
You
Comment #49 on issue 1026 by ronan.l...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
All these match problems seem to have been created by PR #1477
[https://github.com/sympy/sympy/pull/1477].
--
You received this message because you are subscribed to
Comment #48 on issue 1026 by asmeu...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
It looks like a lot of those are related to match.
--
You received this message because you are subscribed to the Google Groups
"sympy-issues" group.
To post to this gro
Comment #47 on issue 1026 by asmeu...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
Here's the latest pypy test run results, with the latest stable version of
PyPy (1.9).
Attachments:
pypy_tests.txt 161 KB
--
You received this message becaus
Comment #46 on issue 1026 by julien.r...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
It would be worth revisiting these failures to check whether they got fixed
after the latest runs of fixes involving hash randomization, which required
sorting in q
Comment #45 on issue 1026 by vlada.pe...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
As the doctests show, most of these errors are about the ordering of the
result. Now, the easy fix would be to just modify the tests to accept both
variants, but pe
Comment #44 on issue 1026 by asmeu...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
And some doctest failures:
__
Comment #43 on issue 1026 by asmeu...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
Here are the latest test failures, with the latest official release of PyPy
(1.8).
Attachments:
pypy_test_failures.txt 3.5 KB
--
You received this message be
Updates:
Labels: -PassedReview
Comment #42 on issue 1026 by vlada.pe...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
Hi, the Piecewise failure seems to have been fixed somewhere (either the
test was changed or PyPy can manage it now; regardless
Comment #41 on issue 1026 by asmeurer: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
The problem is
p1 = Piecewise((1,Interval(0,1,False,True)),(0,True))
p2 = piecewise_fold(expand((1-x)*p1))
The issue is that Python gives
In [4]: p2
Out[4]:
⎧-x + 1 fo
Updates:
Labels: -NeedsReview PassedReview
Comment #40 on issue 1026 by matt...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
This pull request was merged. There is one more remaining test failure:
$ pypy bin/test sympy/functions/elementary/tests
Updates:
Status: Started
Labels: NeedsReview
Comment #39 on issue 1026 by matt...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
I fixed almost all test failures under PyPy in
https://github.com/sympy/sympy/pull/627.
--
You received thi
Updates:
Labels: AlternatePython
Comment #38 on issue 1026 by asmeurer: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
Added a label AlternatePython for issues related to any alternate Python
implementation, including PyPy, Jython, IronPython, etc., and add
Comment #37 on issue 1026 by vlada.pe...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
The pull request is now in, and we also reported some bugs upstream (which
got fixed btw, including the stuff Aaron noted in comment #35), so we
should wait a day o
Comment #36 on issue 1026 by asmeurer: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
There's also a handful of doctest failures. See the (upcoming) sympy-bot
review on the pull request.
--
You received this message because you are subscribed to the Google Group
Comment #35 on issue 1026 by asmeurer: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
Right.
Python:
from sympy import Integer
a = range(10)
a.pop(Integer(0))
0
PyPy:
from sympy import Integer
a = range(10)
a.pop(Integer(0))
Traceback (most recent call last):
Comment #34 on issue 1026 by vlada.pe...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
Ok, I've submitted a new pull request[1] which disables NumPy in PyPy and
also includes the three commits from Renato's pull request (which still
didn't get in). Wi
Comment #33 on issue 1026 by asmeurer: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
To me, this is a PyPy issue, not a SymPy issue. The JIT optimizations just
don't do as well for highly object oriented stuff, like with metaclasses
(or at least that's what I'm
Comment #32 on issue 1026 by vlada.pe...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
Well, we also can't really ask people to test in PyPy if we have a lot of
(known) errors; we should fix them first. I wouldn't mind doing this, as I
said I will, bu
Comment #31 on issue 1026 by renato.c...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
Yes, I noticed that too, there were a lot of new pypy incompatibilities
introduced in the last months - which is a natural consequence of not
running tests on pypy
Comment #30 on issue 1026 by vlada.pe...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
Ah yes, you are right. Well, the errors are there now and they crash the
test runner, so fixing that should be a priority. Other than that, I also
get a lot more er
Comment #29 on issue 1026 by renato.c...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
I didn't. Initial numpy implementation (quite incomplete, as you see) was
introduced in pypy 1.6, those tests were made with pypy 1.5 before 1.6 was
released.
--
Updates:
Cc: renato.c...@gmail.com
Comment #28 on issue 1026 by vlada.pe...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
Renato, how did you handle the numpy-related errors? I get a whole lot of
errors like these and they crash the test runner:
Comment #27 on issue 1026 by renato.c...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
pypy's pow() failure was fixed, let's wait for the next release.
--
You received this message because you are subscribed to the Google Groups
"sympy-issues" group.
To
Comment #26 on issue 1026 by renato.c...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
Inspecting Piecewise more closely, I think there are a lot of features that
are missing. For instance, denesting Piecewise inside another Piecewise,
simplification
Comment #25 on issue 1026 by asmeurer: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
How would solving the issue of recognizing that the inner Piecewise is
never reached solve the issue of arg order? Are you sure it wouldn't just
solve it for this one specific P
Comment #24 on issue 1026 by renato.c...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
The piecewise failure is due to the ordering of the expression args. The
two sides of the assert equality are equivalent. We can test for both
cases, or we can cano
Comment #23 on issue 1026 by renato.c...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
In my branch now, we are down to 6 failures: 2 in mpmath, 1 in
ode.constantsimp, 1 in piecewise and the others are the isprime one, that
must be solved upstream, an
Comment #22 on issue 1026 by renato.c...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
Yes, I had to add the certificate to be able to use it.
--
You received this message because you are subscribed to the Google Groups
"sympy-issues" group.
To post to
Comment #21 on issue 1026 by asmeurer: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
The PyPy bugs site is giving me an SSL error (certificate not trusted). Do
you get that too?
--
You received this message because you are subscribed to the Google Groups
"symp
Comment #20 on issue 1026 by renato.c...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
The Min() failure is actually a bug in the logic of MinMaxBase. I was able
to reproduce it in CPython too - it wasn't triggered due to the ordering of
the variables
Comment #19 on issue 1026 by renato.c...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
The test_solve_polynomial1 failure is just a matter of the ordering of the
solutions. solve() doesn't have a preferred order, so maybe we should
convert both sides
Comment #18 on issue 1026 by renato.c...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
I tracked down the isprime() failure to a bug in pypy's pow(). I reported
it here:
https://bugs.pypy.org/issue792
--
You received this message because you are sub
Comment #17 on issue 1026 by asmeurer: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
That's interesting that PyPy doesn't enforce that, because we'd like to
create an In object that lets you do In(x, Y) == x in Y (Y would be some
Set object), but we can't in CPy
Comment #16 on issue 1026 by renato.c...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
Next issue:
a, b = symbols('a,b')
s = Segment((0, a), (0, b))
assert Point(0, (a + b)/2) in s
gives False in CPython, None in pypy. In this case, pypy is correct -
Comment #15 on issue 1026 by renato.c...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
I investigated a bit the exception in test_pickling, it's triggered by
`BasicMeta("test")` and `BasicType("test")`. It looks like
class A(type): pass
A("test")
w
Comment #14 on issue 1026 by vlada.pe...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
I've applied Ronan's patch enabling pytest on top of your patch (and pushed
to [1]) and run py.test on the files you had errors in (as a rule they're
more informati
Comment #13 on issue 1026 by asmeurer: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
Well, I can tell you the logcombine thing has come up before. I thought it
was changed to use in, but I guess not. Anyway, it should be "assert
logcombine(cos(-2*log(z)+b*log(
Comment #12 on issue 1026 by renato.c...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
Failures attached (ran on my 1026 branch). With the bug in polys fixed (or
at least worked around), there's only one exception raised and 9 failures.
It took 722s,
Comment #11 on issue 1026 by asmeurer: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
Can you upload the test failures as an attachment to this issue?
How long does it take to run the tests with CPython. I remember when I
tried PyPy a while back that it was painf
Comment #10 on issue 1026 by renato.c...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
More good news: on master (9afeb7ee) and with pull request 480 I just
submitted, the whole test suite runs under pypy with only 9 failures and 11
exceptions, in und
Updates:
Owner: vlada.pe...@gmail.com
Comment #9 on issue 1026 by vlada.pe...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
This is good news (thanks for the Cc, too)! As per my timeline, I'll focus
on Python 3 support first (and the assorted is
Updates:
Cc: vlada.pe...@gmail.com
Comment #8 on issue 1026 by renato.c...@gmail.com: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
On master now (3c163b80), sympy is importable in pypy 1.5.0, and
test('core') runs with only one (apparently trivial) error.
Issue 1026: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
Comment #7 by ondrej.certik:
Deferring for later.
Issue attribute updates:
Labels: -Milestone-Release0.6.3
--
You received this message because you are listed in the owner
or CC fields of this issu
Issue 1026: pypy doesn't run sympy
http://code.google.com/p/sympy/issues/detail?id=1026
Comment #6 by ondrej.certik:
Current update here:
http://groups.google.com/group/sympy/browse_thread/thread/7e57714248b4fde6
Note: the patch in the comment #1 doesn't help at all with the latest pypy,
so t
48 matches
Mail list logo
