From issues-reply at bitbucket.org  Mon Jul  1 11:35:49 2019
From: issues-reply at bitbucket.org (Anthony Sottile)
Date: Mon, 01 Jul 2019 15:35:49 +0000 (UTC)
Subject: [pypy-issue] Issue #3034: egg __spec__ is `None` if queried while
 importing using `importlib.util.find_spec` (pypy/pypy)
Message-ID: <20190701153549.26958.72462@app-137.ash1.bb-inf.net>

New issue 3034: egg __spec__ is `None` if queried while importing using `importlib.util.find_spec`
https://bitbucket.org/pypy/pypy/issues/3034/egg-__spec__-is-none-if-queried-while

Anthony Sottile:

this is a bit of a rabbit hole, apologies ? originally from [https://github.com/pallets/flask/issues/3278](https://github.com/pallets/flask/issues/3278)

?

An egg which queries itself using `importlib.util.find_spec` is unable to retrieve its own information. \(This works fine for normal modules it appears\)

?

Here?s my minimal setup, shoved into one script:

?

```python
import os.path
import tempfile
import subprocess
import sys

with tempfile.TemporaryDirectory() as t:
    with open(os.path.join(t, 'setup.py'), 'w') as setup:
        setup.write(
            'from setuptools import setup\n'
            'setup(packages=["site_egg"], zip_safe=True)'
        )
    os.mkdir(os.path.join(t, 'site_egg'))
    with open(os.path.join(t, 'site_egg/__init__.py'), 'w') as f:
        f.write(
            'import importlib.util\n'
            'assert importlib.util.find_spec(__name__) is not None\n'
        )
    open(os.path.join(t, 'site_egg/__init__.py'), 'a').close()

    subprocess.check_call(
        (sys.executable, 'setup.py', 'bdist_egg'),
        stdout=subprocess.DEVNULL,
        cwd=t,
    )

    egg, = os.listdir(os.path.join(t, 'dist'))
    sys.path.insert(0, os.path.join(t, 'dist', egg))

    import site_egg
```

?

in cpython I get the following output:

?

```shell
$ python3.6 t.py
$
```

?

However with pypy3 I get:

?

```
$ pypy3 t.py
Traceback (most recent call last):
  File "t.py", line 29, in <module>
    import site_egg
  File "/tmp/tmp1c9axaoe/dist/UNKNOWN-0.0.0-py3.6.egg/site_egg/__init__.py", line 2, in <module>
  File "/tmp/y/venvpp/lib-python/3/importlib/util.py", line 102, in find_spec
    raise ValueError('{}.__spec__ is None'.format(name))
ValueError: site_egg.__spec__ is None
```

?

```
$ pypy3 --version --version
Python 3.6.1 (784b254d6699, Apr 14 2019, 10:22:42)
[PyPy 7.1.1-beta0 with GCC 6.2.0 20160901]
```



From issues-reply at bitbucket.org  Mon Jul  1 16:09:04 2019
From: issues-reply at bitbucket.org (hugovk)
Date: Mon, 01 Jul 2019 20:09:04 +0000 (UTC)
Subject: [pypy-issue] Issue #3035: CPython old-style Unicode API
 deprecations (pypy/pypy)
Message-ID: <20190701200904.1072.8649@celery-worker-111.ash1.bb-inf.net>

New issue 3035: CPython old-style Unicode API deprecations
https://bitbucket.org/pypy/pypy/issues/3035/cpython-old-style-unicode-api-deprecations

hugovk:

CPython deprecated `PyUnicode_AS_UNICODE` \(amongst others\) in CPython 3.3, and will remove it in CPython 4.0.

[https://docs.python.org/3/c-api/unicode.html#c.PyUnicode\_AS\_UNICODE](https://docs.python.org/3/c-api/unicode.html#c.PyUnicode_AS_UNICODE)

Are there any plans to deprecate/remove/replace it in PyPy?

More generally, how does PyPy deal with deprecations made in CPython?

---

We had a [PR](https://github.com/python-pillow/Pillow/pull/3780) in our project to replace `PyUnicode_AS_UNICODE` with `PyUnicode_AsUCS4Copy` or `PyUnicode_READ_CHAR`. Whilst this was fine for CPython, we got an [error](https://github.com/python-pillow/Pillow/issues/3835#issuecomment-507351734) on PyPy3:

`_imagingft.obj : error LNK2019: unresolved external symbol _PyUnicode_AsUCS4Copy referenced in function _text_layout_raqm`

With:

Python 3.6.1 \(784b254d6699, Apr 16 2019, 12:10:48\)  
\[PyPy 7.1.1-beta0 with MSC v.1910 32 bit\] on win32

and:

Python 3.5.3 \(928a4f70d3de, Feb 08 2019, 12:56:35\)  
\[PyPy 7.0.0 with MSC v.1910 32 bit\] on win32

Thank you!



From issues-reply at bitbucket.org  Wed Jul  3 08:54:19 2019
From: issues-reply at bitbucket.org (Sasquatch)
Date: Wed, 03 Jul 2019 12:54:19 +0000 (UTC)
Subject: [pypy-issue] Issue #3036: Tkinter UnicodeEncodeError: 'mbcs' codec
 can't encode characters in position 0--1 (pypy/pypy)
Message-ID: <20190703125419.6070.45843@celery-worker-110.ash1.bb-inf.net>

New issue 3036: Tkinter UnicodeEncodeError: 'mbcs' codec can't encode characters in position 0--1
https://bitbucket.org/pypy/pypy/issues/3036/tkinter-unicodeencodeerror-mbcs-codec-cant

Sasquatch:

I am trying to get matplotlib working on my machine with pypy, but it seems tkinter is causing problems. I installed pypy from the nightly builds as recommend from here [https://stackoverflow.com/questions/56826170/value-error-invalid-argument-with-pypy-when-importing-pandas-and-plotly](https://stackoverflow.com/questions/56826170/value-error-invalid-argument-with-pypy-when-importing-pandas-and-plotly) When I import tkinter however, I get: 

?

```python
>>>> import tkinter
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "C:\pypy-c-jit-96765-24757bd09ed9-win32\lib-python\3\tkinter\__init__.py"
, line 36, in <module>
    import _tkinter # If this fails your Python may not be configured for Tk
  File "C:\pypy-c-jit-96765-24757bd09ed9-win32\lib_pypy\_tkinter\__init__.py", l
ine 55, in <module>
    tklib.Tcl_FindExecutable(os.fsencode(sys.executable))
  File "C:\pypy-c-jit-96765-24757bd09ed9-win32\lib-python\3\os.py", line 800, in
 fsencode
    return filename.encode(encoding, errors)
UnicodeEncodeError: 'mbcs' codec can't encode characters in position 0--1: mbcs
encoding does not support errors='surrogateescape'
```

And thus:

```
>>>> import matplotlib.pyplot
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "C:\pypy-c-jit-96765-24757bd09ed9-win32\site-packages\matplotlib\pyplot.p
y", line 2355, in <module>
    switch_backend(rcParams["backend"])
  File "C:\pypy-c-jit-96765-24757bd09ed9-win32\site-packages\matplotlib\pyplot.p
y", line 221, in switch_backend
    backend_mod = importlib.import_module(backend_name)
  File "C:\pypy-c-jit-96765-24757bd09ed9-win32\lib-python\3\importlib\__init__.p
y", line 126, in import_module
    return _bootstrap._gcd_import(name[level:], package, level)
  File "<frozen importlib._bootstrap>", line 979, in _gcd_import
  File "<frozen importlib._bootstrap>", line 962, in _find_and_load
  File "<frozen importlib._bootstrap>", line 951, in _find_and_load_unlocked
  File "<frozen importlib._bootstrap>", line 656, in _load_unlocked
  File "<builtin>/frozen importlib._bootstrap_external", line 691, in exec_modul
e
  File "<frozen importlib._bootstrap>", line 206, in _call_with_frames_removed
  File "C:\pypy-c-jit-96765-24757bd09ed9-win32\site-packages\matplotlib\backends
\backend_tkagg.py", line 1, in <module>
    from . import _backend_tk
  File "C:\pypy-c-jit-96765-24757bd09ed9-win32\site-packages\matplotlib\backends
\_backend_tk.py", line 6, in <module>
    import tkinter as tk
  File "C:\pypy-c-jit-96765-24757bd09ed9-win32\lib-python\3\tkinter\__init__.py"
, line 36, in <module>
    import _tkinter # If this fails your Python may not be configured for Tk
  File "C:\pypy-c-jit-96765-24757bd09ed9-win32\lib_pypy\_tkinter\__init__.py", l
ine 55, in <module>
    tklib.Tcl_FindExecutable(os.fsencode(sys.executable))
  File "C:\pypy-c-jit-96765-24757bd09ed9-win32\lib-python\3\os.py", line 800, in
 fsencode
    return filename.encode(encoding, errors)
UnicodeEncodeError: 'mbcs' codec can't encode characters in position 0--1: mbcs
encoding does not support errors='surrogateescape'
```



From issues-reply at bitbucket.org  Thu Jul  4 09:30:31 2019
From: issues-reply at bitbucket.org (Pierre Augier)
Date: Thu, 04 Jul 2019 13:30:31 +0000 (UTC)
Subject: [pypy-issue] Issue #3037: PyPy3.6 Release Schedule? New beta
 quality PyPy3.6 soon? (pypy/pypy)
Message-ID: <20190704133031.6519.33871@app-137.ash1.bb-inf.net>

New issue 3037: PyPy3.6 Release Schedule? New beta quality PyPy3.6 soon?
https://bitbucket.org/pypy/pypy/issues/3037/pypy36-release-schedule-new-beta-quality

Pierre Augier:

I'd like to use PyPy with packages needing Python >= 3.6. In science, it is now quite standard. For example, conda, which only uses 2 Python versions, is now limited to Python 3.6 and 3.7 \(and quite soon 3.7 and 3.8 :slight_smile: \). So people tend to use f-strings and `pathlib.Path` and to assume that the code will be executed with Python >= 3.6.

So having a better PyPy3.6 soon \(for example before EuroScipy in September\) would be desirable. PyPy3.6 7.1.1 has bugs that are now solved in the PyPy36 branch. For example, \[Transonic\]\([https://transonic.readthedocs.io](https://transonic.readthedocs.io)\) tests pass with the nightly builds but fail with PyPy3.6 7.1.1.

Do you plan to release a new beta quality PyPy3.6 soon? I didn?t find anything like the release schedule of CPython for PyPy [https://www.python.org/dev/peps/pep-0596/](https://www.python.org/dev/peps/pep-0596/) ? Is there anything similar?



From issues-reply at bitbucket.org  Fri Jul  5 06:17:25 2019
From: issues-reply at bitbucket.org (=?utf-8?q?Karsten_Fl=C3=BCgge?=)
Date: Fri, 05 Jul 2019 10:17:25 +0000 (UTC)
Subject: [pypy-issue] Issue #3038: WebAssembly as target (pypy/pypy)
Message-ID: <20190705101725.26274.35852@app-147.ash1.bb-inf.net>

New issue 3038: WebAssembly as target
https://bitbucket.org/pypy/pypy/issues/3038/webassembly-as-target

Karsten Fl?gge:

Do you have any plans targeting WASM?



From issues-reply at bitbucket.org  Sat Jul  6 08:46:53 2019
From: issues-reply at bitbucket.org (Shlomi Fish)
Date: Sat, 06 Jul 2019 12:46:53 +0000 (UTC)
Subject: [pypy-issue] Issue #3039: cython / array.array bindings generate
 errors (pypy/pypy)
Message-ID: <20190706124652.24502.90513@celery-worker-110.ash1.bb-inf.net>

New issue 3039: cython / array.array bindings generate errors
https://bitbucket.org/pypy/pypy/issues/3039/cython-arrayarray-bindings-generate-errors

Shlomi Fish:

Why [https://github.com/shlomif/primesieve-python/tree/array-pypy3](https://github.com/shlomif/primesieve-python/tree/array-pypy3) \(note the branch\) gives me <<ValueError: array.array size changed, may indicate binary incompatibility. Expected 72 from C header, got 24 from PyObject>> in pypy3 and works fine in cpython3? After I ?#if 0? the check, I get a segfault:

?

```plaintext
shlomif[ProjEuler]:~/progs/riddles/project-euler/git/project-euler/641$ gdb -args pypy3 641-v1.py
GNU gdb (GDB) 8.2-8.mga7 (Mageia release 7)
Copyright (C) 2018 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.
Type "show copying" and "show warranty" for details.
This GDB was configured as "x86_64-mageia-linux-gnu".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>.
Find the GDB manual and other documentation resources online at:
    <http://www.gnu.org/software/gdb/documentation/>.

For help, type "help".
Type "apropos word" to search for commands related to "word"...
Reading symbols from pypy3...Reading symbols from /home/shlomif/Download/unpack/prog/python/pypy3.5-6.0.0-linux_x86_64-portable/bin/pypy3.5.debug...(no debugging symbols found)...done.                                                                                       
(no debugging symbols found)...done.
(gdb) r
Starting program: /home/shlomif/Download/unpack/prog/python/pypy3.5-6.0.0-linux_x86_64-portable/bin/pypy3 641-v1.py
Missing separate debuginfos, use: debuginfo-install glibc-2.29-14.mga8.x86_64
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib64/libthread_db.so.1".

Program received signal SIGSEGV, Segmentation fault.
resize (n=26355869, self=0x8e3350) at primesieve/array/_array.cpp:1190
1190        PyMem_Resize(items, char, (size_t)(n * self->ob_descr->itemsize));
Missing separate debuginfos, use: debuginfo-install lib64bz2_1-1.0.6-13.mga7.x86_64 lib64xcrypt1-4.4.6-1.mga7.x86_64 lib64zlib1-1.2.11-7.mga7.x86_64 libgomp1-8.3.1-0.20190628.2.mga8.x86_64 libstdc++6-8.3.1-0.20190628.2.mga8.x86_64                                         
(gdb) p self
$1 = (arrayobject *) 0x8e3350
(gdb) p self[0]
$2 = {ob_refcnt = 2305843009213693953, ob_pypy_link = 140737275159976,
  ob_type = 0x8bb840, ob_size = 262193, data = {ob_item = 0x0,
    as_floats = 0x0, as_doubles = 0x0, as_ints = 0x0, as_uints = 0x0,
    as_uchars = 0x0, as_schars = 0x0, as_chars = 0x0, as_ulongs = 0x0,
    as_longs = 0x0, as_ulonglongs = 0x0, as_longlongs = 0x0, as_shorts = 0x0,
    as_ushorts = 0x0, as_pyunicodes = 0x0, as_voidptr = 0x0}, allocated = 0,
  ob_descr = 0x4100015a90, weakreflist = 0x4000, ob_exports = -309255392}
(gdb) p self[0].ob
ob_descr      ob_pypy_link  ob_size
ob_exports    ob_refcnt     ob_type
(gdb) p self[0].ob_descr
$3 = (arraydescr *) 0x4100015a90
(gdb) p self[0].ob_descr [0]
Cannot access memory at address 0x4100015a90
(gdb) p items
$4 = (void *) 0x0
(gdb) p n
$5 = 26355869
(gdb) p self[0].ob_descr
$6 = (arraydescr *) 0x4100015a90
(gdb) q
A debugging session is active.

```

Thanks!

?

Here is the offending code:

?

```python
# import array
import heapq
import sys
import time
from bisect import bisect_right
# from subprocess import PIPE, Popen

# import primesieve
import primesieve.array

from six import print_
from six.moves import range

# Tuple is (n,k) where n is the sum and k is the count.
# pipe = Popen(["primesieve", "-p1", "2", "500000029"], stdout=PIPE).stdout
# pipe = Popen("primesieve -p1 2 1000", shell=True, stdout=PIPE).stdout

t1 = time.time()
primes = primesieve.array.primes(500000029)
print_(primes[0])
t2 = time.time()
```



From issues-reply at bitbucket.org  Mon Jul  8 08:56:14 2019
From: issues-reply at bitbucket.org (Amit Ripshtos)
Date: Mon, 08 Jul 2019 12:56:14 +0000 (UTC)
Subject: [pypy-issue] Issue #3040: ssl module in pypy3.6 does not have
 ssl.OP_NO_SSLv2 (pypy/pypy)
Message-ID: <20190708125613.34835.30877@celery-worker-112.ash1.bb-inf.net>

New issue 3040: ssl module in pypy3.6 does not have ssl.OP_NO_SSLv2
https://bitbucket.org/pypy/pypy/issues/3040/ssl-module-in-pypy36-does-not-have

Amit Ripshtos:

Hey guys, thanks for working on pypy!

I found an issue when using aiohttp with pypy3.6 and doing verify=False , the aiohttp code creates SSLContext by doing:

```
sslcontext = ssl.SSLContext(ssl.PROTOCOL_SSLv23)
sslcontext.options |= ssl.OP_NO_SSLv2
sslcontext.options |= ssl.OP_NO_SSLv3
sslcontext.options |= ssl.OP_NO_COMPRESSION
sslcontext.set_default_verify_paths()
return sslcontext
```

?

However, this code fails with the following exception:

```
AttributeError: module 'ssl' has no attribute 'OP_NO_SSLv2'
```

?

I tried to take a look and I saw that in pypy3.6, the attribute/option ?OP\_NO\_SSLv2?, therefore the aiohttp package fails when using the verify\_ssl=False setting.

?

**How to mimic the issue:**

Using pypy3.6, try to run this code:

```
import aiohttp

async with aiohttp.ClientSession(connector=aiohttp.TCPConnector(verify_ssl=False)) as session:
    async with session.request(method='get', url='https://google.com') as response:
        response.raise_for_status()
        await response.read()
```

?

This code will fail with the exception:

`AttributeError: module 'ssl' has no attribute 'OP_NO_SSLv2'`

?

**How it should be:**

There should be be OP\_NO\_SSLv2 attribute under ssl module.

?

I will try to fix the issue on my own, however I?m a newbie to pypy code , therefore I might need help.

?

Thanks and have a great week,

Amit.



From issues-reply at bitbucket.org  Mon Jul  8 09:43:48 2019
From: issues-reply at bitbucket.org (=?utf-8?q?Johan_J=C3=B6nsson?=)
Date: Mon, 08 Jul 2019 13:43:48 +0000 (UTC)
Subject: [pypy-issue] Issue #3041: pypy3 import numpy fails unknown symbol
 PyGILState_Check (pypy/pypy)
Message-ID: <20190708134347.30209.12837@celery-worker-110.ash1.bb-inf.net>

New issue 3041: pypy3 import numpy fails unknown symbol PyGILState_Check
https://bitbucket.org/pypy/pypy/issues/3041/pypy3-import-numpy-fails-unknown-symbol

Johan J?nsson:

I am trying to install numpy \(both from source and via pip\). With pypy-7.1.1-beta0 \(from the gentoo repositories\) I can build/install numpy, however when I try to import numpy i get the following output

```
Python 3.6.1 (784b254d669919c872a505b807db8462b6140973, Jul 07 2019, 15:05:37)
[PyPy 7.1.1-beta0 with GCC 8.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
And now for something completely different: ``it's nonsense all the way down,
and also all the way up''
>>>> import numpy
Traceback (most recent call last):
  File ".local/lib/pypy3.6/site-packages/numpy/core/__init__.py", line 17, in <module>
    from . import multiarray
  File ".local/lib/pypy3.6/site-packages/numpy/core/multiarray.py", line 14, in <module>
    from . import overrides
  File ".local/lib/pypy3.6/site-packages/numpy/core/overrides.py", line 7, in <module>
    from numpy.core._multiarray_umath import (
ImportError: .local/lib/pypy3.6/site-packages/numpy/core/_multiarray_umath.pypy3-71-x86_64-linux-gnu.so: undefined symbol: PyGILState_Check

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "~/.local/lib/pypy3.6/site-packages/numpy/__init__.py", line 142, in <module>
    from . import core
  File ".local/lib/pypy3.6/site-packages/numpy/core/__init__.py", line 47, in <module>
    raise ImportError(msg)
ImportError: 

IMPORTANT: PLEASE READ THIS FOR ADVICE ON HOW TO SOLVE THIS ISSUE!

Importing the numpy c-extensions failed.
- Try uninstalling and reinstalling numpy.
- If you have already done that, then:
  1. Check that you expected to use Python3.6 from "/usr/bin/pypy3",
     and that you have no directories in your PATH or PYTHONPATH that can
     interfere with the Python and numpy version "1.18.0.dev0+48efcc7" you're trying to use.
  2. If (1) looks fine, you can open a new issue at
     https://github.com/numpy/numpy/issues.  Please include details on:
     - how you installed Python
     - how you installed numpy
     - your operating system
     - whether or not you have multiple versions of Python installed
     - if you built from source, your compiler versions and ideally a build log

- If you're working with a numpy git repository, try `git clean -xdf`
  (removes all files not under version control) and rebuild numpy.

Note: this error has many possible causes, so please don't comment on
an existing issue about this - open a new one instead.

Original error was: ~/.local/lib/pypy3.6/site-packages/numpy/core/_multiarray_umath.pypy3-71-x86_64-linux-gnu.so: undefined symbol: PyGILState_Check
```



From issues-reply at bitbucket.org  Tue Jul  9 04:47:22 2019
From: issues-reply at bitbucket.org (Jacob Lifshay)
Date: Tue, 09 Jul 2019 08:47:22 +0000 (UTC)
Subject: [pypy-issue] Issue #3042: crash when reading non-blocking
 subprocess output (pypy/pypy)
Message-ID: <20190709084722.30273.4186@app-137.ash1.bb-inf.net>

New issue 3042: crash when reading non-blocking subprocess output
https://bitbucket.org/pypy/pypy/issues/3042/crash-when-reading-non-blocking-subprocess

Jacob Lifshay:

Output:

```
RPython traceback:
  File "pypy_interpreter.c", line 37926, in BuiltinCode_funcrun_obj
  File "pypy_module__io_1.c", line 11207, in W_BufferedReader_readline_w
  File "implement_2.c", line 32105, in dispatcher_70
  File "pypy_module__io.c", line 15998, in W_BufferedReader__raw_read
Traceback (most recent call last):
  File "pypy_bug.py", line 10, in <module>
    s = p.stdout.readline()
SystemError: unexpected internal exception (please report a bug): <BlockingIOError object at 0x7f9481569f00>; internal traceback was dumped to stderr
```

Versions:  
Python 3.6.1 \(7a2e437acfce, Mar 21 2019, 13:39:41\)  
\[PyPy 7.2.0-beta0 with GCC 6.2.0 20160901\]  
Ubuntu 18.04 x86\_64

reproducable using:

```python
import subprocess
import fcntl
import os
from select import select

p = subprocess.Popen(["/bin/bash", "-c", "printf ''"], stdin=subprocess.DEVNULL, stdout=subprocess.PIPE, stderr=None)
fl = fcntl.fcntl(p.stdout, fcntl.F_GETFL)
fcntl.fcntl(p.stdout, fcntl.F_SETFL, fl | os.O_NONBLOCK)
while True:
    s = p.stdout.readline()
    print(s)
    if len(s) == 0:
        if p.poll() is not None:
            break
        try:
            select([p.stdout], [], [], 1.0)
        except InterruptedError:
            pass
```



From issues-reply at bitbucket.org  Fri Jul 12 04:50:09 2019
From: issues-reply at bitbucket.org (Christoph Deil)
Date: Fri, 12 Jul 2019 08:50:09 +0000 (UTC)
Subject: [pypy-issue] Issue #3043: Support Py_RETURN_NOTIMPLEMENTED
 (pypy/pypy)
Message-ID: <20190712085008.3648.42109@celery-worker-108.ash1.bb-inf.net>

New issue 3043: Support Py_RETURN_NOTIMPLEMENTED
https://bitbucket.org/pypy/pypy/issues/3043/support-py_return_notimplemented

Christoph Deil:

Hi, I?d like to try Astropy on PyPy.

\(related ticket from last year in the Astropy tracker: [https://github.com/astropy/astropy/issues/7768](https://github.com/astropy/astropy/issues/7768)\)

At  the moment, `pip_pypy3 install astropy` fails like this:

```
astropy/wcs/src/unit_list_proxy.c:199:5: error: use of undeclared identifier 'Py_RETURN_NOTIMPLEMENTED'
```

Full log: [https://gist.github.com/cdeil/74d46beebdabf5b40cccedde759bdec1#file-gistfile1-txt-L1503](https://gist.github.com/cdeil/74d46beebdabf5b40cccedde759bdec1#file-gistfile1-txt-L1503)

>From [https://docs.python.org/3/c-api/object.html](https://docs.python.org/3/c-api/object.html) :

> `Py_RETURN_NOTIMPLEMENTED` - Properly handle returning `Py_NotImplemented` from within a C function \(that is, increment the reference count of NotImplemented and return it\).

Is `Py_RETURN_NOTIMPLEMENTED` something you could support in PyPy?



From issues-reply at bitbucket.org  Sat Jul 13 11:04:58 2019
From: issues-reply at bitbucket.org (Zsolt Cserna)
Date: Sat, 13 Jul 2019 15:04:58 +0000 (UTC)
Subject: [pypy-issue] Issue #3044: py3.6: str.encode should not work with
 not-text encoders (pypy/pypy)
Message-ID: <20190713150457.33657.87563@celery-worker-109.ash1.bb-inf.net>

New issue 3044: py3.6: str.encode should not work with not-text encoders
https://bitbucket.org/pypy/pypy/issues/3044/py36-strencode-should-not-work-with-not

Zsolt Cserna:

With Cpython, using a non-text encoder such as ?hex?, the following happens:

```
>>> "foo".encode("hex")
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
LookupError: 'hex' is not a text encoding; use codecs.encode() to handle arbitrary codecs
```

In pypy this call gets to the codec itself, where `TypeError` is raised \(which is correct as this encoding works on bytes, not on unicode\):

```
>>>> "foo".encode("hex")
Traceback (most recent call last):
  File "/home/zsolt/src/pypy/lib-python/3/encodings/hex_codec.py", line 15, in hex_encode
    return (binascii.b2a_hex(input), len(input))
TypeError: a bytes-like object is required, not str
```

The root cause of the problem is that when looking up a codec in `str.encode`, `lookup_text_codec()` function is not called. Thereby the CodecInfo's `_is_text_encoding` is not checked at all as `str.encode` uses `codes.encode`under the hood.

The solution would be adding a check for this method somewhere between `str.encode` and using the encoder.



From issues-reply at bitbucket.org  Mon Jul 22 09:23:20 2019
From: issues-reply at bitbucket.org (H. Turgut Uyar)
Date: Mon, 22 Jul 2019 13:23:20 +0000 (UTC)
Subject: [pypy-issue] Issue #3045: pathlib path resolving incorrectly
 (pypy/pypy)
Message-ID: <20190722132319.12948.95063@app-147.ash1.bb-inf.net>

New issue 3045: pathlib path resolving incorrectly
https://bitbucket.org/pypy/pypy/issues/3045/pathlib-path-resolving-incorrectly

H. Turgut Uyar:

On PyPy 3.6.1 \(7.1.1-beta0\), resolving a pathlib path that contains an intermediate parent step produces an incorrect result \(at least, different from CPython\). For example, if the current directory is `/home/uyar/Projects/myproject` which contains a `docs` directory and a `setup.py` file, the following code produces `PosixPath('/home/uyar/Projects/myproject/docs')`

```python
from pathlib import Path
Path("docs/../setup.py").resolve()
```

Whereas in CPython the same code produces `PosixPath('/home/uyar/Projects/myproject/setup.py')`



From issues-reply at bitbucket.org  Wed Jul 24 07:37:41 2019
From: issues-reply at bitbucket.org (Alex Edwards)
Date: Wed, 24 Jul 2019 11:37:41 +0000 (UTC)
Subject: [pypy-issue] Issue #3046: Raspberry Pi Observations - RAM and
 Library Checks? (pypy/pypy)
Message-ID: <20190724113741.29832.64857@celery-worker-110.ash1.bb-inf.net>

New issue 3046: Raspberry Pi Observations - RAM and Library Checks?
https://bitbucket.org/pypy/pypy/issues/3046/raspberry-pi-observations-ram-and-library

Alex Edwards:

Hi,

I?ve just been compiling PyPy pypy2.7-v7.1.1-src.zip on a Raspberry Pi ?stretch? - quite a slow process! :slight_smile: Thought I?d share my installation experience and niggles and hope it helps.  
  
After quite a long time, 30-60mins perhaps, it bombed out with insufficient memory.  I added some more swap and its working.  
  
It could be handy to simply warn if memory less than 500MB or 1GB spare perhaps?  Adding 1GB swap worked, though I didn?t try smaller amounts.  
  
Then installing again, after perhaps an hour, the build failed again because of a missing -dev library / package  libbz2-dev.  Then twice more on libncurses5-dev.  
  
I then found your helpful page - [http://doc.pypy.org/en/latest/build.html](http://doc.pypy.org/en/latest/build.html) - and installed all the packages I could in advance.  
  
sudo apt-get install libexpat1-dev  
sudo apt-get install libncurses-dev  
sudo apt-get install libncursesw-dev  
sudo apt-get install libgdbm-dev  
sudo apt-get install tk-dev   -- installs X, so avoid if possible  
sudo apt-get install liblzma-dev

It would have helped me a lot if these dependencies had been spotted and warned/failed sooner.  Also a link to your installation page would help too, rather than failing on one package per long compile.  
  
PS - I only more recently noticed that a newer version of pypy is in Raspbian buster - I should probably have just installed that! :smiley:   
  
Cheers

Alex



From issues-reply at bitbucket.org  Thu Jul 25 11:30:46 2019
From: issues-reply at bitbucket.org (Zihan Zheng)
Date: Thu, 25 Jul 2019 15:30:46 +0000 (UTC)
Subject: [pypy-issue] Issue #3047: Internal exception when loading json with
 unicode escapes (pypy/pypy)
Message-ID: <20190725153046.7061.3313@app-137.ash1.bb-inf.net>

New issue 3047: Internal exception when loading json with unicode escapes
https://bitbucket.org/pypy/pypy/issues/3047/internal-exception-when-loading-json-with

Zihan Zheng:

```
>>>> import json
>>>> json.loads('"\\udc00\\udc00"')
RPython traceback:
  File "pypy_interpreter.c", line 42074, in BuiltinCode2_fastcall_2
  File "pypy_module__pypyjson.c", line 221, in loads_1
  File "pypy_module__pypyjson.c", line 647, in JSONDecoder_decode_any
  File "pypy_module__pypyjson.c", line 5884, in JSONDecoder_decode_string_escaped
  File "pypy_module__pypyjson.c", line 8763, in JSONDecoder_decode_escape_sequence_unicode
  File "rpython_rlib.c", line 23815, in unichr_as_utf8
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/local/lib-python/3/json/__init__.py", line 361, in loads
    if _pypyjson else _default_decoder.decode(s))
SystemError: unexpected internal exception (please report a bug): <OutOfRange object at 0x7f787bc22db0>; internal traceback was dumped to stderr
```

Expected result: `'\udc00\udc00'`

PyPy version: PyPy 7.1.1-beta0



From issues-reply at bitbucket.org  Mon Jul 29 14:24:52 2019
From: issues-reply at bitbucket.org (Ronan Lamy)
Date: Mon, 29 Jul 2019 18:24:52 +0000 (UTC)
Subject: [pypy-issue] Issue #3048: tracemalloc (pypy/pypy)
Message-ID: <20190729182451.17315.41479@app-147.ash1.bb-inf.net>

New issue 3048: tracemalloc
https://bitbucket.org/pypy/pypy/issues/3048/tracemalloc

Ronan Lamy:

pypy doesn't implement the tracemalloc module because it?s closely tied to CPython internals. However, quoting from [What?s New In Python 3.6](https://docs.python.org/3/whatsnew/3.6.html#warnings): 

> When a `ResourceWarning` warning is logged, the `tracemalloc` module is now used to try to retrieve the traceback where the destroyed object was allocated.

which is a useful thing to do on any implementation and rather similar to what `pypy2 -X track-resources` does. So:

* is there a sensible way to create a `_tracemalloc` that would make this work like on CPython?
* or should we just enable `-X track-resources` on pypy3?