From issues-reply at bitbucket.org  Mon Aug  1 05:51:44 2016
From: issues-reply at bitbucket.org (Richard Plangger)
Date: Mon, 01 Aug 2016 09:51:44 -0000
Subject: [pypy-issue] Issue #2359: VMProf does not display traces correctly
 (pypy/pypy)
Message-ID: <20160801095144.70383.59073@celery-worker-102.ash1.bb-inf.net>

New issue 2359: VMProf does not display traces correctly
https://bitbucket.org/pypy/pypy/issues/2359/vmprof-does-not-display-traces-correctly

Richard Plangger:

Here is a python example:

https://paste.pound-python.org/show/RXatEmOq95xg4HcneO50/

And the corresponding cpu profile:

http://vmprof.com/#/88a155cf8a54c7ddc23374aa4fd407b1

The issue might be in trace_from_start of tmp_callback.

Needs more investigation.



From issues-reply at bitbucket.org  Mon Aug  1 21:33:47 2016
From: issues-reply at bitbucket.org (RyoYokoyama)
Date: Tue, 02 Aug 2016 01:33:47 -0000
Subject: [pypy-issue] Issue #2360: Windows ctypes.CDLL : TypeError: function
 name must be a string or integer (pypy/pypy)
Message-ID: <20160802013347.1354.17200@celery-worker-104.ash1.bb-inf.net>

New issue 2360: Windows ctypes.CDLL : TypeError: function name must be a string or integer
https://bitbucket.org/pypy/pypy/issues/2360/windows-ctypescdll-typeerror-function-name

RyoYokoyama:

On Windows, ctypes.CDLL cannot accept unicode function name.

On PyPy2:
```
#!python
>>>> import ctypes
>>>> lib = ctypes.CDLL('sqlite3')
>>>> lib[b'sqlite3_libversion']()
1642690176
>>>> lib[u'sqlite3_libversion']()
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "c:\pypy2\lib-python\2.7\ctypes\__init__.py", line 387, in __getitem__
    func = self._FuncPtr((name_or_ordinal, self))
  File "c:\pypy2\lib_pypy\_ctypes\function.py", line 254, in __init__
    ptr = self._getfuncptr([], ctypes.c_int)
  File "c:\pypy2\lib_pypy\_ctypes\function.py", line 439, in _getfuncptr
    self._ptr = cdll.getfunc(self.name, ffi_argtypes, ffi_restype)
TypeError: function name must be a string or integer
```

On CPython 2.7:
```
#!python
>>> import ctypes
>>> lib = ctypes.CDLL('sqlite3')
>>> lib[b'sqlite3_libversion']()
1787770432
>>> lib[u'sqlite3_libversion']()
1787770432
```



From issues-reply at bitbucket.org  Mon Aug  1 22:22:27 2016
From: issues-reply at bitbucket.org (Craig Rodrigues)
Date: Tue, 02 Aug 2016 02:22:27 -0000
Subject: [pypy-issue] Issue #2361: TypeError: __class__ assignment: 'A'
 object layout differs from 'B' (pypy/pypy)
Message-ID: <20160802022227.31821.17442@celery-worker-104.ash1.bb-inf.net>

New issue 2361: TypeError: __class__ assignment: 'A' object layout differs from 'B'
https://bitbucket.org/pypy/pypy/issues/2361/typeerror-__class__-assignment-a-object

Craig Rodrigues:

Alex Gaynor was able to come up with this test case which passes under CPython
but fails under Pypy:

```
class A(object):
    __slots__ = ["a"]

class B(object):
    __slots__ = ["a"]

inst = A()
inst.a = 3

inst.__class__ = B
```


I was also able to reproduce this bug in the Twisted test suite.
I can do this by doing the following:

```
git clone https://github.com/twisted/twisted twisted_test1
cd twisted_test1
pypy -m unittest twisted.test.test_rebuild
```



From issues-reply at bitbucket.org  Tue Aug  2 23:04:07 2016
From: issues-reply at bitbucket.org (RyoYokoyama)
Date: Wed, 03 Aug 2016 03:04:07 -0000
Subject: [pypy-issue] Issue #2362: Django's development server repeats
 reload infinitely (pypy/pypy)
Message-ID: <20160803030407.41090.90801@celery-worker-102.ash1.bb-inf.net>

New issue 2362: Django's development server repeats reload infinitely
https://bitbucket.org/pypy/pypy/issues/2362/djangos-development-server-repeats-reload

RyoYokoyama:

[Django's development server](https://docs.djangoproject.com/ja/1.9/ref/django-admin/#runserver) repeats reload infinitely.

```
#!python

C:\pypyreload>pypy manage.py runserver
Performing system checks...

System check identified no issues (0 silenced).

You have unapplied migrations; your app may not work properly until they are applied.
Run 'python manage.py migrate' to apply them.
August 03, 2016 - 11:44:13
Django version 1.9.8, using settings 'pypyreload.settings'
Starting development server at http://127.0.0.1:8000/
Quit the server with CTRL-BREAK.
Performing system checks...

System check identified no issues (0 silenced).

You have unapplied migrations; your app may not work properly until they are applied.
Run 'python manage.py migrate' to apply them.
August 03, 2016 - 11:44:16
Django version 1.9.8, using settings 'pypyreload.settings'
Starting development server at http://127.0.0.1:8000/
Quit the server with CTRL-BREAK.
Performing system checks...

System check identified no issues (0 silenced).

...
```

I found that in django.utils.autoreload, os.stat([path to lib-python\2.7\posixpath.py]).st_ctime returned 2 different values and this file was recognized as modified.

The attached simple code reproduces this problem.

My Environment:

* Windows 10 64bit
* pypy2.7 v5.3.1




From issues-reply at bitbucket.org  Wed Aug  3 13:54:15 2016
From: issues-reply at bitbucket.org (9001)
Date: Wed, 03 Aug 2016 17:54:15 -0000
Subject: [pypy-issue] Issue #2363: Segfault in GC: pinned object in
 'objects_to_trace' (pypy/pypy)
Message-ID: <20160803175415.33783.82824@celery-worker-101.ash1.bb-inf.net>

New issue 2363: Segfault in GC: pinned object in 'objects_to_trace'
https://bitbucket.org/pypy/pypy/issues/2363/segfault-in-gc-pinned-object-in

9001:

I encountered a reproducible segfault when running a particular script, both with release builds and with lldebug.

```
PyPy assertion failed at rpython_memory_gc.c:22908:
in pypy_g_IncrementalMiniMarkGC_visit: pinned object in 'objects_to_trace'
```

The script and further details are provided in the attachment.



From issues-reply at bitbucket.org  Tue Aug  9 21:13:30 2016
From: issues-reply at bitbucket.org (Dewald Pretorius)
Date: Wed, 10 Aug 2016 01:13:30 -0000
Subject: [pypy-issue] Issue #2364: gctransformer issue while trying to
 translate pypy to ARM (pypy/pypy)
Message-ID: <20160810011330.10329.37501@celery-worker-101.ash1.bb-inf.net>

New issue 2364: gctransformer issue while trying to translate pypy to ARM
https://bitbucket.org/pypy/pypy/issues/2364/gctransformer-issue-while-trying-to

Dewald Pretorius:

I've followed the guide from here: http://rpython.readthedocs.io/en/latest/arm.html#arm

All steps were successful, up to "pypy /home/username/pypy2-v5.3.1-src/rpython/bin/rpython -O1 --platform=arm target.py"

which fails as follows:
root at ubuntu:/srv/chroot/precise_arm# pypy /home/username/pypy2-v5.3.1-src/rpython/bin/rpython -O1 --platform=arm target.py
[translation:info] 2.7.8 (2.5.0+dfsg-2, Feb 12 2015, 05:23:41)
[PyPy 2.5.0 with GCC 4.9.2]
[platform:msg] Set platform with 'arm' cc=None, using cc='gcc', version='Unknown'
[translation:info] Translating target as defined by target
Traceback (most recent call last):
  File "app_main.py", line 75, in run_toplevel
  File "/home/username/pypy2-v5.3.1-src/rpython/bin/rpython", line 20, in <module>
    main()
  File "/home/username/pypy2-v5.3.1-src/rpython/translator/goal/translate.py", line 216, in main
    targetspec_dic, translateconfig, config, args = parse_options_and_load_target()
  File "/home/username/pypy2-v5.3.1-src/rpython/translator/goal/translate.py", line 173, in parse_options_and_load_target
    set_opt_level(config, translateconfig.opt)
  File "/home/username/pypy2-v5.3.1-src/rpython/config/translationoption.py", line 379, in set_opt_level
    config.translation.gc = config.translation.gc
  File "/home/username/pypy2-v5.3.1-src/rpython/config/config.py", line 68, in __setattr__
    self.setoption(name, value, 'user')
  File "/home/username/pypy2-v5.3.1-src/rpython/config/config.py", line 112, in setoption
    child.setoption(self, value, who)
  File "/home/username/pypy2-v5.3.1-src/rpython/config/config.py", line 271, in setoption
    homeconfig.setoption(name, reqvalue, who2)
  File "/home/username/pypy2-v5.3.1-src/rpython/config/config.py", line 111, in setoption
    'option %s' % (value, name))
ConflictConfigError: cannot override value to ref for option gctransformer

I'd really appreciate any assistance as our tests have demonstrated a seven factor speed increase in using pypy over python2.7.

I have made sure that I have all the dependencies installed that I need to build pypy from source - by building pypy from source.

I've also tried translating the full interpreter
root at ubuntu:/home/username/pypy2-v5.3.1-src/pypy/goal# /home/username/pypy2-v5.3.1-src/pypy/goal/pypy-c /home/username/pypy2-v5.3.1-src/rpython/bin/rpython -Ojit --platform=arm --gcrootfinder=shadowstack --jit-backend=arm targetpypystandalone.py
[translation:info] 2.7.10 (7e8df3df96417c16c2d55b41352ec82c9c69c978, Aug 09 2016, 00:24:37)
[PyPy 5.3.1 with GCC 4.9.2]
[platform:msg] Set platform with 'arm' cc=None, using cc='gcc', version='Unknown'
[translation:info] Translating target as defined by targetpypystandalone
[platform:execute] sb2 -t ARM gcc -c -O3 -pthread -fomit-frame-pointer -Wall -Wno-unused /tmp/usession-release-pypy2.7-v5.3.1-13/gcctest.c -o /tmp/usession-release-pypy2.7-v5.3.1-13/gcctest.o
[platform:execute] sb2 -t ARM gcc -c -O3 -pthread -fomit-frame-pointer -Wall -Wno-unused /tmp/usession-release-pypy2.7-v5.3.1-13/gcctest.c -o /tmp/usession-release-pypy2.7-v5.3.1-13/gcctest.o
[platform:Error] sb2: Error: Invalid target specified, aborting.
Traceback (most recent call last):
  File "/home/username/pypy2-v5.3.1-src/rpython/bin/rpython", line 20, in <module>
    main()
  File "/home/username/pypy2-v5.3.1-src/rpython/translator/goal/translate.py", line 216, in main
    targetspec_dic, translateconfig, config, args = parse_options_and_load_target()
  File "/home/username/pypy2-v5.3.1-src/rpython/translator/goal/translate.py", line 155, in parse_options_and_load_target
    targetspec_dic = load_target(targetspec)
  File "/home/username/pypy2-v5.3.1-src/rpython/translator/goal/translate.py", line 97, in load_target
    mod = __import__(specname)
  File "targetpypystandalone.py", line 6, in <module>
    from pypy.interpreter import gateway
  File "/home/username/pypy2-v5.3.1-src/pypy/interpreter/gateway.py", line 19, in <module>
    from pypy.interpreter.eval import Code
  File "/home/username/pypy2-v5.3.1-src/pypy/interpreter/eval.py", line 6, in <module>
    from pypy.interpreter.baseobjspace import W_Root
  File "/home/username/pypy2-v5.3.1-src/pypy/interpreter/baseobjspace.py", line 6, in <module>
    from rpython.rlib.debug import make_sure_not_resized
  File "/home/username/pypy2-v5.3.1-src/rpython/rlib/debug.py", line 9, in <module>
    from rpython.rtyper.lltypesystem import rffi
  File "/home/username/pypy2-v5.3.1-src/rpython/rtyper/lltypesystem/rffi.py", line 540, in <module>
    NUMBER_TYPES = setup()
  File "/home/username/pypy2-v5.3.1-src/rpython/rtyper/lltypesystem/rffi.py", line 529, in setup
    names = populate_inttypes()
  File "/home/username/pypy2-v5.3.1-src/rpython/rtyper/lltypesystem/rffi.py", line 523, in populate_inttypes
    platform.populate_inttypes(populatelist)
  File "/home/username/pypy2-v5.3.1-src/rpython/rtyper/tool/rfficache.py", line 95, in populate_inttypes
    sizes = sizeof_c_types(names_c, **kwds)
  File "/home/username/pypy2-v5.3.1-src/rpython/rtyper/tool/rfficache.py", line 45, in sizeof_c_types
    answer = ask_gcc(question, **kwds)
  File "/home/username/pypy2-v5.3.1-src/rpython/rtyper/tool/rfficache.py", line 35, in ask_gcc
    return build_executable_cache([c_file], eci, ignore_errors=ignore_errors)
  File "/home/username/pypy2-v5.3.1-src/rpython/tool/gcc_cache.py", line 28, in build_executable_cache
    result = platform.execute(platform.compile(c_files, eci))
  File "/home/username/pypy2-v5.3.1-src/rpython/translator/platform/__init__.py", line 54, in compile
    ofiles = self._compile_o_files(cfiles, eci, standalone)
  File "/home/username/pypy2-v5.3.1-src/rpython/translator/platform/__init__.py", line 76, in _compile_o_files
    ofiles.append(self._compile_c_file(self.cc, cfile, compile_args))
  File "/home/username/pypy2-v5.3.1-src/rpython/translator/platform/posix.py", line 40, in _compile_c_file
    cwd=str(cfile.dirpath()))
  File "/home/username/pypy2-v5.3.1-src/rpython/translator/platform/arm.py", line 47, in _execute_c_compiler
    self._handle_error(returncode, stderr, stdout, outname)
  File "/home/username/pypy2-v5.3.1-src/rpython/translator/platform/__init__.py", line 152, in _handle_error
    raise CompilationError(stdout, stderr)
CompilationError: CompilationError(err="""
        sb2: Error: Invalid target specified, aborting.
        """)

Any advise would be really appreciated.



From issues-reply at bitbucket.org  Wed Aug 10 12:31:56 2016
From: issues-reply at bitbucket.org (Brecht Machiels)
Date: Wed, 10 Aug 2016 16:31:56 -0000
Subject: [pypy-issue] Issue #2365: rinohtype much slower on PyPy3 (pypy/pypy)
Message-ID: <20160810163156.9464.71320@celery-worker-103.ash1.bb-inf.net>

New issue 2365: rinohtype much slower on PyPy3
https://bitbucket.org/pypy/pypy/issues/2365/rinohtype-much-slower-on-pypy3

Brecht Machiels:

```
#!bash

pip install rinohtype
wget http://docutils.sourceforge.net/docs/user/rst/demo.txt
rm -rf demo.rtc && time rinoh demo.txt
```

CPython 3.3.6 takes about 16 seconds on my machine. PyPy3 takes about 34 seconds.

(I installed PyPy through pyenv, which reports pypy3.3-5.2-alpha1, but `pypy -V` reports "PyPy 5.2.0-alpha0")



From issues-reply at bitbucket.org  Fri Aug 12 05:53:53 2016
From: issues-reply at bitbucket.org (Edd Barrett)
Date: Fri, 12 Aug 2016 09:53:53 -0000
Subject: [pypy-issue] Issue #2366: Make PyPy W^X Compatible (pypy/pypy)
Message-ID: <20160812095353.8946.52620@celery-worker-104.ash1.bb-inf.net>

New issue 2366: Make PyPy W^X Compatible
https://bitbucket.org/pypy/pypy/issues/2366/make-pypy-w-x-compatible

Edd Barrett:

W^X is a mitigation technique whereby allocated pages are either writeable or executable, but never both. By working this way it makes it much harder for an attacker who has an existing memory vulnerability (e.g. a buffer overflow or use-after-free) to craft malicious binary code and jump to it.

https://en.wikipedia.org/wiki/W%5EX

OpenBSD is pushing to make all package software W^X compliant, which includes PyPy. Note however that W^X benefits everyone, not just OpenBSD users. Pax/Linux users would also be able to run PyPy.

The overall model in a JIT is something like this:

```
mem = alloc_writable_pages();
write_code(mem);
mprotect(mem, PROT_EXEC);
```

Then every time you need to mutate the memory, you have to `mprotect` it writeable again.

I have already started work on this, but don't expect it to be ready quickly -- I think it's going to be pretty hard.

Thanks

Responsible: vext01

From issues-reply at bitbucket.org  Sat Aug 13 09:18:20 2016
From: issues-reply at bitbucket.org (Naoki INADA)
Date: Sat, 13 Aug 2016 13:18:20 -0000
Subject: [pypy-issue] Issue #2367: del bytearray[:n] should be optimized
 (pypy/pypy)
Message-ID: <20160813131820.6604.87841@celery-worker-102.ash1.bb-inf.net>

New issue 2367: del bytearray[:n] should be optimized
https://bitbucket.org/pypy/pypy/issues/2367/del-bytearray-n-should-be-optimized

Naoki INADA:

del bytearray[:n] is optimized from Python 3.4
ref: https://bugs.python.org/issue19087

AsyncIO uses bytearray for I/O buffers.
PyPy3 compatible to CPython 3.5 should have same optimization.



From issues-reply at bitbucket.org  Mon Aug 15 09:42:59 2016
From: issues-reply at bitbucket.org (Edd Barrett)
Date: Mon, 15 Aug 2016 13:42:59 -0000
Subject: [pypy-issue] Issue #2368: cpu_info() should be at runtime,
 not translation time. (pypy/pypy)
Message-ID: <20160815134259.26977.45930@celery-worker-101.ash1.bb-inf.net>

New issue 2368: cpu_info() should be at runtime, not translation time.
https://bitbucket.org/pypy/pypy/issues/2368/cpu_info-should-be-at-runtime-not

Edd Barrett:

As discussed with @arigo In `rpython/jit/backend/x86/detect_feature.py` there is a function `cpu_info()` which collects information from the host cpu by executing binary code. Currently this is run at translation time, but actually it should be run at runtime.

Found last week, when looking into W^X.



From issues-reply at bitbucket.org  Tue Aug 16 01:02:38 2016
From: issues-reply at bitbucket.org (David Naylor)
Date: Tue, 16 Aug 2016 05:02:38 -0000
Subject: [pypy-issue] Issue #2369: PyPy does not compile FreeBSD/PowerPC
 (pypy/pypy)
Message-ID: <20160816050238.11054.43370@celery-worker-101.ash1.bb-inf.net>

New issue 2369: PyPy does not compile FreeBSD/PowerPC
https://bitbucket.org/pypy/pypy/issues/2369/pypy-does-not-compile-freebsd-powerpc

David Naylor:

After applying the attached patches PyPy successfully translates, but fails in compiling one of the generated C files:
```
data_rpython_jit_metainterp.c:54770:2: error: 'pypy__is_nonnull_longdouble' undeclared here (not in a function)
  pypy__is_nonnull_longdouble, /* 1466 */
  ^
gmake[1]: *** [Makefile:890: data_rpython_jit_metainterp.o] Error 1
```
pypy__is_nonnull_longdouble appears to come from pypy/module/_cffi_backend/misc.py, unless pypy__is_nonnull_longdouble means something else in a different context.  



From issues-reply at bitbucket.org  Wed Aug 17 10:27:57 2016
From: issues-reply at bitbucket.org (Brian Reihnard)
Date: Wed, 17 Aug 2016 14:27:57 -0000
Subject: [pypy-issue] Issue #2370: pypy doesn't start if the primary group
 gid > max(int) 2147483647 (pypy/pypy)
Message-ID: <20160817142757.23562.46432@celery-worker-102.ash1.bb-inf.net>

New issue 2370: pypy doesn't start if the primary group gid > max(int) 2147483647
https://bitbucket.org/pypy/pypy/issues/2370/pypy-doesnt-start-if-the-primary-group-gid

Brian Reihnard:

pypy doesn't start if the primary group gid >  max(int) 

For instance if gid=4294967294, we got a crash:

RPython traceback:
  File "pypy_goal.c", line 2093, in entry_point
  File "rpython_rlib_rvmprof.c", line 255, in execute_frame_rvmprof__star_3
  File "pypy_interpreter_2.c", line 4932, in execute_frame
  File "rpython_jit_metainterp.c", line 46977, in ll_portal_runner__Unsigned_Bool_pypy_interpreter
  File "pypy_module_pypyjit.c", line 340, in portal_40
  File "pypy_interpreter_3.c", line 35262, in handle_bytecode__AccessDirect_None
  File "pypy_interpreter_3.c", line 41587, in dispatch_bytecode__AccessDirect_None
  File "pypy_interpreter_4.c", line 2081, in call_function__AccessDirect_None
  File "rpython_rlib_rvmprof.c", line 255, in execute_frame_rvmprof__star_3
  File "pypy_interpreter_2.c", line 4932, in execute_frame
  File "rpython_jit_metainterp.c", line 46977, in ll_portal_runner__Unsigned_Bool_pypy_interpreter
  File "pypy_module_pypyjit.c", line 340, in portal_40
  File "pypy_interpreter_3.c", line 35262, in handle_bytecode__AccessDirect_None
  File "pypy_interpreter_3.c", line 42670, in dispatch_bytecode__AccessDirect_None
  File "pypy_interpreter_3.c", line 59793, in IMPORT_NAME__AccessDirect_None
  File "pypy_interpreter.c", line 26116, in BuiltinCode_funcrun_obj
  File "pypy_interpreter.c", line 51709, in BuiltinCode_handle_exception
Fatal RPython error: OSError
Aborted (core dumped)




From issues-reply at bitbucket.org  Sun Aug 21 04:38:24 2016
From: issues-reply at bitbucket.org (Maciej Fijalkowski)
Date: Sun, 21 Aug 2016 08:38:24 -0000
Subject: [pypy-issue] Issue #2371: Unicode history does not work very well
 in pyrepl (at least on OS X) (pypy/pypy)
Message-ID: <20160821083824.59727.11311@celery-worker-102.ash1.bb-inf.net>

New issue 2371: Unicode history does not work very well in pyrepl (at least on OS X)
https://bitbucket.org/pypy/pypy/issues/2371/unicode-history-does-not-work-very-well-in

Maciej Fijalkowski:

At least for characters that are wider than one byte, if I do the following:

>>>> x=u'??? ?'

then when I press "enter" and "up arrow" and then "Ctrl-a" it goes somewhere in the middle of the line. I think editing works just fine, but the visuals are screwed. Works fine on cpython



From issues-reply at bitbucket.org  Sun Aug 21 22:48:58 2016
From: issues-reply at bitbucket.org (Daniel Holth)
Date: Mon, 22 Aug 2016 02:48:58 -0000
Subject: [pypy-issue] Issue #2372: bug in
 wheel.tests.test_wheelfile:test_verifying_zipfile() (pypy/pypy)
Message-ID: <20160822024858.7762.78723@celery-worker-104.ash1.bb-inf.net>

New issue 2372: bug in wheel.tests.test_wheelfile:test_verifying_zipfile()
https://bitbucket.org/pypy/pypy/issues/2372/bug-in-wheelteststest_wheelfile

Daniel Holth:

I'm having a strange issue with the wheel tests, failing in pypy2-v5.3.1-linux64 but not pypy 4.0.0 or CPython.

Wheel replaces _update_crc on individual instances of ZipExtFile to do sha256 hashing instead of crc32.

The third time it tries to do this at vzf.open("three") in the test, the ZipExtFile instance calls _update_crc from the class instead of the one from the instance. You can notice this from the failing test or by playing with the debugger around line 672 of the standard Python zipfile.py. Look at self._update_crc, step into it, wrong function.

```
#!python

            if self._compress_type == ZIP_STORED:
                self._update_crc(data, eof=(self._compress_left==0))
```

```
#!shell

python -c "import wheel.test.test_wheelfile ; wheel.test.test_wheelfile.test_verifying_zipfile()"
```

The test passes if I comment out the first two vzf.open("one") and vzf.open("two") .



From issues-reply at bitbucket.org  Mon Aug 22 06:19:01 2016
From: issues-reply at bitbucket.org (Joost Shao)
Date: Mon, 22 Aug 2016 10:19:01 -0000
Subject: [pypy-issue] Issue #2373: what's the diff between microbenchmarks
 and minibenchmarks, why merge them to one ? (pypy/pypy)
Message-ID: <20160822101901.10350.30766@celery-worker-101.ash1.bb-inf.net>

New issue 2373: what's the diff between microbenchmarks and minibenchmarks, why merge them to one ?
https://bitbucket.org/pypy/pypy/issues/2373/whats-the-diff-between-microbenchmarks-and

Joost Shao:

i review the code in these two directory, and run them with pytho2.7 and pypy .

but i don't understand the reason two diff as mini and micro, can you help me ?



From issues-reply at bitbucket.org  Mon Aug 22 10:50:33 2016
From: issues-reply at bitbucket.org (avi)
Date: Mon, 22 Aug 2016 14:50:33 -0000
Subject: [pypy-issue] Issue #2374: How to solve open shared object error
 while trying to install pypy on linux? (pypy/pypy)
Message-ID: <20160822145033.11363.85764@celery-worker-103.ash1.bb-inf.net>

New issue 2374: How to solve open shared object error while trying to install pypy on linux?
https://bitbucket.org/pypy/pypy/issues/2374/how-to-solve-open-shared-object-error

avi:

I am following this link: http://pypy.readthedocs.io/en/latest/install.html

And I am trying to install pypy. I have downloaded 'pypy-c-jit-latest-linux64.tar.bz2' extracted the files and run

 ./pypy-2.1/bin/pypy
However I get the following error:

./pypy: error while loading shared libraries: libffi.so.6: cannot open shared object file: No such file or directory
How can I solve this?

Thank



From issues-reply at bitbucket.org  Tue Aug 23 12:38:15 2016
From: issues-reply at bitbucket.org (Ronan Lamy)
Date: Tue, 23 Aug 2016 16:38:15 -0000
Subject: [pypy-issue] Issue #2375: Missing posix attributes in py3k
 (pypy/pypy)
Message-ID: <20160823163815.5624.71114@celery-worker-101.ash1.bb-inf.net>

New issue 2375: Missing posix attributes in py3k
https://bitbucket.org/pypy/pypy/issues/2375/missing-posix-attributes-in-py3k

Ronan Lamy:

Comparing CPython3.3 with current pypy3 (py3k branch), the following attributes are missing from `dir(posix)`:
```
['CLD_CONTINUED', 'CLD_DUMPED', 'CLD_EXITED', 'CLD_TRAPPED', 
'F_LOCK', 'F_TEST', 'F_TLOCK', 'F_ULOCK', 'O_ACCMODE', 'O_CLOEXEC',
'POSIX_FADV_DONTNEED', 'POSIX_FADV_NOREUSE', 'POSIX_FADV_NORMAL',
'POSIX_FADV_RANDOM', 'POSIX_FADV_SEQUENTIAL', 'POSIX_FADV_WILLNEED',
'PRIO_PGRP', 'PRIO_PROCESS', 'PRIO_USER', 'P_ALL', 'P_PGID', 'P_PID',
'SCHED_BATCH', 'SCHED_FIFO', 'SCHED_IDLE', 'SCHED_OTHER', 'SCHED_RESET_ON_FORK', 'SCHED_RR',
'SEEK_DATA', 'SEEK_HOLE', 'WEXITED', 'WNOWAIT', 'WSTOPPED',
'XATTR_CREATE', 'XATTR_REPLACE', 'XATTR_SIZE_MAX',
'get_terminal_size', 'getgrouplist', 'getpriority', 'getxattr', 'listxattr', 'lockf', 'pipe2',
'posix_fadvise', 'posix_fallocate', 'pread', 'pwrite', 'readv', 'removexattr',
'sched_get_priority_max', 'sched_get_priority_min', 'sched_getaffinity', 'sched_getparam',
'sched_getscheduler', 'sched_param', 'sched_rr_get_interval', 'sched_setaffinity',
'sched_setparam', 'sched_setscheduler', 'sched_yield', 'sendfile', 'setpriority', 'setxattr',
'sync', 'terminal_size', 'times_result', 'waitid', 'waitid_result', 'writev']
```



From issues-reply at bitbucket.org  Tue Aug 23 15:00:32 2016
From: issues-reply at bitbucket.org (Amber Brown)
Date: Tue, 23 Aug 2016 19:00:32 -0000
Subject: [pypy-issue] Issue #2376: buildbot.pypy.org is not available over
 HTTPS (pypy/pypy)
Message-ID: <20160823190032.25668.98057@celery-worker-101.ash1.bb-inf.net>

New issue 2376: buildbot.pypy.org is not available over HTTPS
https://bitbucket.org/pypy/pypy/issues/2376/buildbotpypyorg-is-not-available-over

Amber Brown:

This is an issue if you are downloading, say, nightly builds.

For Twisted, we install txacme from PyPI and pass the following to the WebStatus:

```
#!python

http_port='le:certs:tcp:port=443:interface=0.0.0.0',
```

We then have a service that redirects all HTTP traffic to HTTPS with 302s.

I am guessing this would go here: https://bitbucket.org/pypy/buildbot/src/879f3ec8e19a3d5b21c138248bd6fec93005e3ae/bot2/pypybuildbot/master.py?at=default&fileviewer=file-view-default#master.py-31




From issues-reply at bitbucket.org  Tue Aug 23 16:07:28 2016
From: issues-reply at bitbucket.org (Daniil Yarancev)
Date: Tue, 23 Aug 2016 20:07:28 -0000
Subject: [pypy-issue] Issue #2377: #2377 (pypy/pypy)
Message-ID: <20160823200728.7199.21837@celery-worker-103.ash1.bb-inf.net>

New issue 2377: #2377
https://bitbucket.org/pypy/pypy/issues/2377/2377

Daniil Yarancev:

Using last version of PyPy on Windows (5.3.1), issuing a "pypy -m ensurepip" command gives strange error:

```
#!python


File "C:\pypy2-v5.3.1-win32\lib-python\2.7\mimetypes.py", line 267, in read_windows_registry
      mimetype = mimetype.encode(default_encoding)
  UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 0: ordinal not in range(128)
```


I use Russian version of Windows 10 (10.0.10586 build) 

But I found a hack for resolving this error: changed 
```
#!python

mimetype = mimetype.encode(default_encoding)

to

mimetype = mimetype
```




From issues-reply at bitbucket.org  Wed Aug 24 20:13:07 2016
From: issues-reply at bitbucket.org (llvtt)
Date: Thu, 25 Aug 2016 00:13:07 -0000
Subject: [pypy-issue] Issue #2378: NaN with payload in struct.(un)pack with
 endian-ness given (pypy/pypy)
Message-ID: <20160825001307.4745.48108@celery-worker-101.ash1.bb-inf.net>

New issue 2378: NaN with payload in struct.(un)pack with endian-ness given
https://bitbucket.org/pypy/pypy/issues/2378/nan-with-payload-in-struct-un-pack-with

llvtt:

PyPy3 does not appear to handle NaN with a payload when calling struct.pack or struct.unpack if endianness is given as part of the format string:


```
#!python

import struct

# result: b'\x12\x00\x00\x00\x00\x00\xf8\x7f'
struct.pack('d', struct.unpack('d', b'\x12\x00\x00\x00\x00\x00\xf8\x7f')[0])
# result: b'\x00\x00\x00\x00\x00\x00\xf8\x7f' (starts with b'\x00' instead of b'x\12')
struct.pack('d', struct.unpack('<d', b'\x12\x00\x00\x00\x00\x00\xf8\x7f')[0])

```

```
$ pypy3 --version
Python 3.2.5 (b2091e973da69152b3f928bfaabd5d2347e6df46, Sep 23 2015, 09:21:00)
[PyPy 2.4.0 with GCC 4.2.1 Compatible Apple LLVM 7.0.0 (clang-700.0.72)]
```

PyPy2 does not appear to have this behavior (NaN payload is part of the output whether endian-ness is specified or not).



From issues-reply at bitbucket.org  Sun Aug 28 02:37:32 2016
From: issues-reply at bitbucket.org (Armin Rigo)
Date: Sun, 28 Aug 2016 06:37:32 -0000
Subject: [pypy-issue] Issue #2382: sys.meta_path not working like CPython's
 (pypy/pypy)
Message-ID: <20160828063732.27725.77283@celery-worker-101.ash1.bb-inf.net>

New issue 2382: sys.meta_path not working like CPython's
https://bitbucket.org/pypy/pypy/issues/2382/sysmeta_path-not-working-like-cpythons

Armin Rigo:

``sys.meta_path`` is not working like CPython's:

http://stackoverflow.com/questions/39186850/import-hooks-custom-module-loaders-for-pypy-do-not-work



From issues-reply at bitbucket.org  Sun Aug 28 17:49:52 2016
From: issues-reply at bitbucket.org (Hexcode Technologies)
Date: Sun, 28 Aug 2016 21:49:52 -0000
Subject: [pypy-issue] Issue #2383: Testing PokemonGo-Map on pypy and have an
 unexpected error. (pypy/pypy)
Message-ID: <20160828214952.14936.85081@celery-worker-101.ash1.bb-inf.net>

New issue 2383: Testing PokemonGo-Map on pypy and have an unexpected error.
https://bitbucket.org/pypy/pypy/issues/2383/testing-pokemongo-map-on-pypy-and-have-an

Hexcode Technologies:

https://github.com/PokemonGoMap/PokemonGo-Map
it happens on : `install_aliases()`
```
Traceback (most recent call last):
  File "runserver.py", line 25, in <module>
    from pogom.search import search_overseer_thread
  File "/home/v3ss/github/PokemonGo-Map/pogom/search.py", line 33, in <module>
    from pgoapi import PGoApi
  File "/home/v3ss/pypy/site-packages/pgoapi/__init__.py", line 50, in <module>
    from pgoapi.pgoapi import PGoApi
  File "/home/v3ss/pypy/site-packages/pgoapi/pgoapi.py", line 35, in <module>
    from pgoapi.auth_ptc import AuthPtc
  File "/home/v3ss/pypy/site-packages/pgoapi/auth_ptc.py", line 28, in <module>
    install_aliases()
  File "/home/v3ss/pypy/site-packages/future/standard_library/__init__.py", line 465, in install_aliases
    from future.backports.urllib import request
  File "/home/v3ss/pypy/site-packages/future/backports/urllib/request.py", line 97, in <module>
    from future.backports.http import client as http_client
  File "/home/v3ss/pypy/site-packages/future/backports/http/client.py", line 76, in <module>
    from future.backports.email import parser as email_parser
  File "/home/v3ss/pypy/site-packages/future/backports/email/parser.py", line 15, in <module>
    from future.backports.email.feedparser import FeedParser, BytesFeedParser
  File "/home/v3ss/pypy/site-packages/future/backports/email/feedparser.py", line 32, in <module>
    from future.backports.email import message
  File "/home/v3ss/pypy/site-packages/future/backports/email/message.py", line 20, in <module>
    from future.backports.email import utils
  File "/home/v3ss/pypy/site-packages/future/backports/email/utils.py", line 40, in <module>
    from future.backports.urllib.parse import quote as url_quote, unquote as url_unquote
  File "/home/v3ss/pypy/site-packages/future/backports/urllib/parse.py", line 633, in <module>
    _ALWAYS_SAFE_BYTES = bytes(_ALWAYS_SAFE)
  File "/home/v3ss/pypy/site-packages/future/types/newbytes.py", line 108, in __new__
    raise ValueError('bytes must be in range(0, 256)')
ValueError: bytes must be in range(0, 256)
```

It works fine in python 2.7.12



From issues-reply at bitbucket.org  Sat Aug 27 13:11:57 2016
From: issues-reply at bitbucket.org (Daniil Yarancev)
Date: Sat, 27 Aug 2016 17:11:57 -0000
Subject: [pypy-issue] Issue #2380: PyPy 3.5: Implement PEP 485 - A function
 for testing approximate equality (pypy/pypy)
Message-ID: <20160827171157.17212.19492@celery-worker-103.ash1.bb-inf.net>

New issue 2380: PyPy 3.5: Implement PEP 485 - A function for testing approximate equality
https://bitbucket.org/pypy/pypy/issues/2380/pypy-35-implement-pep-485-a-function-for

Daniil Yarancev:

Example code:


```
#!python

>>> import math
>>> a = 5.0
>>> b = 4.99998
>>> math.isclose(a, b, rel_tol=1e-5)
True
>>> math.isclose(a, b, rel_tol=1e-6)
False
```




From issues-reply at bitbucket.org  Sat Aug 27 13:09:02 2016
From: issues-reply at bitbucket.org (Daniil Yarancev)
Date: Sat, 27 Aug 2016 17:09:02 -0000
Subject: [pypy-issue] Issue #2379: PyPy 3.5: Implement PEP 461 (pypy/pypy)
Message-ID: <20160827170902.32203.70059@celery-worker-104.ash1.bb-inf.net>

New issue 2379: PyPy 3.5: Implement PEP 461
https://bitbucket.org/pypy/pypy/issues/2379/pypy-35-implement-pep-461

Daniil Yarancev:



PEP 461 adds support for the % interpolation operator to bytes and bytearray.


```
#!python

b'Hello %b!' % b'World'
```




From issues-reply at bitbucket.org  Sat Aug 27 13:25:27 2016
From: issues-reply at bitbucket.org (Daniil Yarancev)
Date: Sat, 27 Aug 2016 17:25:27 -0000
Subject: [pypy-issue] Issue #2381: PyPy 3.5 ensurepip module failure
 (pypy/pypy)
Message-ID: <20160827172527.25507.21358@celery-worker-103.ash1.bb-inf.net>

New issue 2381: PyPy 3.5 ensurepip module failure
https://bitbucket.org/pypy/pypy/issues/2381/pypy-35-ensurepip-module-failure

Daniil Yarancev:


```
#!python
ubuntu at ip-172-31-26-119:~/pypy35/bin$ ./pypy3 -m ensurepip
Ignoring ensurepip failure: pip 8.1.2 requires SSL/TLS

```
Is it a bug, or just work-in-progress?



From issues-reply at bitbucket.org  Mon Aug 29 02:58:14 2016
From: issues-reply at bitbucket.org (Daniil Yarancev)
Date: Mon, 29 Aug 2016 06:58:14 -0000
Subject: [pypy-issue] Issue #2384: pyglet gdiplus.py fatal error under Win32
 (pypy/pypy)
Message-ID: <20160829065814.40701.33840@celery-worker-102.ash1.bb-inf.net>

New issue 2384: pyglet gdiplus.py fatal error under Win32
https://bitbucket.org/pypy/pypy/issues/2384/pyglet-gdipluspy-fatal-error-under-win32

Daniil Yarancev:

When trying to execute simple example from pyglet:

```
#!python

import pyglet

window = pyglet.window.Window()
image = pyglet.image.load('texture.png')
label = pyglet.text.Label('Hello, World!!',
                      font_name='Times New Roman',
                      font_size=36,
                      x=window.width//2, y=window.height//2,
                      anchor_x='center', anchor_y='center')

@window.event
def on_draw():
    window.clear()
    label.draw()
    #    image.blit(0,0)

pyglet.app.run()
```
App starts, white screen, and then it's crash with this error:


```
#!python

Traceback (most recent call last):
  File "main.py", line 5, in <module>
    image = pyglet.image.load('texture.png')
  File "C:\PyPy\site-packages\pyglet\image\__init__.py", line 193, in load
    image = decoder.decode(file, filename)
  File "C:\PyPy\site-packages\pyglet\image\codecs\gdiplus.py", line 277, in decode
    self._delete_bitmap(bitmap)
  File "C:\PyPy\site-packages\pyglet\image\codecs\gdiplus.py", line 272, in _delete_bitmap
    self.stream.Release()
  File "C:\PyPy\site-packages\pyglet\com.py", line 125, in <lambda>
    self.method.get_field()(self.i, self.name)(obj, *args)
  File "C:\PyPy\lib_pypy\_ctypes\function.py", line 345, in __call__
    newargs.insert(0, thisvalue.value)
AttributeError: 'IUnknown' object has no attribute 'value'
```
PyPy version:

```
#!python

Python 2.7.10 (77392ad26350, Aug 28 2016, 10:44:29)
[PyPy 5.4.0 with MSC v.1500 32 bit] on win32
```
I think there's some issues with com.py in pyglet





From issues-reply at bitbucket.org  Tue Aug 30 06:22:05 2016
From: issues-reply at bitbucket.org (David MacIver)
Date: Tue, 30 Aug 2016 10:22:05 -0000
Subject: [pypy-issue] Issue #2385: Missing __self__ from super objects
 (pypy/pypy)
Message-ID: <20160830102205.23499.23458@celery-worker-103.ash1.bb-inf.net>

New issue 2385: Missing __self__ from super objects
https://bitbucket.org/pypy/pypy/issues/2385/missing-__self__-from-super-objects

David MacIver:

```python
class Foo(object):
    pass

x = Foo()

assert super(Foo, x).__self__ is x
```

This works on CPython 2.7 but on pypy 3.5.1 it errors with 

```
Traceback (most recent call last):
  File "superior.py", line 6, in <module>
    assert super(Foo, x).__self__ is x
AttributeError: 'super' object has no attribute '__self__'
```

What's weird is this seems to be a regression. I don't currently have an old pypy around to test on but I found this because code that was working on pypy 2.6.1 started breaking when I upgraded my build to 3.5.1



From issues-reply at bitbucket.org  Tue Aug 30 19:48:08 2016
From: issues-reply at bitbucket.org (toby dacre)
Date: Tue, 30 Aug 2016 23:48:08 -0000
Subject: [pypy-issue] Issue #2386: unicode.format() has issue with unicode
 keys (pypy/pypy)
Message-ID: <20160830234808.29107.96592@celery-worker-105.ash1.bb-inf.net>

New issue 2386: unicode.format() has issue with unicode keys
https://bitbucket.org/pypy/pypy/issues/2386/unicodeformat-has-issue-with-unicode-keys

toby dacre:

The following code works with python 2.7.11 and  3.4.3 outputting `? extremely evil`

pypy 2.7.10 throws a KeyError this appears to be a bug

```
# -*- coding: utf-8 -*-
x = u'{? Very bad name}'
d = {u'? Very bad name': u'? extremely evil'}

print(x.format(**d))
```




From issues-reply at bitbucket.org  Wed Aug 31 19:06:35 2016
From: issues-reply at bitbucket.org (Armin Rigo)
Date: Wed, 31 Aug 2016 23:06:35 -0000
Subject: [pypy-issue] Issue #2387: ctypes.c_longdouble is not usable
 (pypy/pypy)
Message-ID: <20160831230635.25160.38907@celery-worker-107.ash1.bb-inf.net>

New issue 2387: ctypes.c_longdouble is not usable
https://bitbucket.org/pypy/pypy/issues/2387/ctypesc_longdouble-is-not-usable

Armin Rigo:

The type ctypes.c_longdouble is not usable at all.  For example, ``ctypes.POINTER(ctypes.c_longdouble)`` crashes.  See comments in 191a4cce5363.



From issues-reply at bitbucket.org  Wed Aug 31 21:03:44 2016
From: issues-reply at bitbucket.org (Stefano Rivera)
Date: Thu, 01 Sep 2016 01:03:44 -0000
Subject: [pypy-issue] Issue #2388: Sandbox broken in 5.4.0 os.access()
 called with float mode (pypy/pypy)
Message-ID: <20160901010344.3077.60134@celery-worker-107.ash1.bb-inf.net>

New issue 2388: Sandbox broken in 5.4.0 os.access() called with float mode
https://bitbucket.org/pypy/pypy/issues/2388/sandbox-broken-in-540-osaccess-called-with

Stefano Rivera:

The sandbox fails to get to a REPL in 5.4.0, it seems something is making a broken call to `os.access()` with a floating point `mode` argument:

```
[sandlib:call] ll_os.ll_os_getenv('PYPY_GENERATIONGC_NURSERY')
[sandlib:result] None
[sandlib:call] ll_os.ll_os_getenv('PYPY_GC_DEBUG')
[sandlib:result] None
[sandlib:call] ll_os.ll_os_envitems()
[sandlib:result] []
[sandlib:call] ll_os.ll_os_stat('/bin/pypy-c')
[sandlib:vpath] '/bin/pypy-c' => <RealFile /usr/lib/pypy-sandbox/x86_64-linux-gnu/pypy-c-sandbox>
[sandlib:result] posix.stat_result(st...ctime=0)
[sandlib:call] ll_os.ll_os_access('/bin/pypy-c', 1.0)
[sandlib:vpath] '/bin/pypy-c' => <RealFile /usr/lib/pypy-sandbox/x86_64-linux-gnu/pypy-c-sandbox>
Traceback (most recent call last):
  File "/usr/bin/pypy-sandbox", line 12, in <module>
    execfile(fn)
  File "/usr/lib/python2.7/dist-packages/pypy/sandbox/pypy_interact.py", line 133, in <module>
    main()
  File "/usr/lib/python2.7/dist-packages/pypy/sandbox/pypy_interact.py", line 128, in main
    sandproc.interact()
  File "/usr/lib/python2.7/dist-packages/pypy/sandbox/rpython/translator/sandbox/sandlib.py", line 325, in interact
    returncode = self.handle_until_return()
  File "/usr/lib/python2.7/dist-packages/pypy/sandbox/rpython/translator/sandbox/sandlib.py", line 252, in handle_until_return
    write_exception(child_stdin, e, tb)
  File "/usr/lib/python2.7/dist-packages/pypy/sandbox/rpython/translator/sandbox/sandlib.py", line 249, in handle_until_return
    answer, resulttype = self.handle_message(fnname, *args)
  File "/usr/lib/python2.7/dist-packages/pypy/sandbox/rpython/translator/sandbox/sandlib.py", line 288, in handle_message
    return handler(*args), resulttype
  File "/usr/lib/python2.7/dist-packages/pypy/sandbox/rpython/translator/sandbox/sandlib.py", line 457, in do_ll_os__ll_os_access
    return node.access(mode)
  File "/usr/lib/python2.7/dist-packages/pypy/sandbox/rpython/translator/sandbox/vfs.py", line 47, in access
    return (e_mode & mode) == mode
TypeError: unsupported operand type(s) for &: 'int' and 'float'
```