Mailman 3 November 2016 - pypy-issue

Issue #2431: Is there a pypy3.3-5.5.0 release? (pypy/pypy)
by Josh Friend 10 Nov '16

10 Nov '16

New issue 2431: Is there a pypy3.3-5.5.0 release? https://bitbucket.org/pypy/pypy/issues/2431/is-there-a-pypy33-550-release Josh Friend: The [downloads page](http://pypy.org/download.html) only has 5.5.0-alpha. However, there is a blog post from october [announcing the release of 5.5.0](https://morepypy.blogspot.com/2016/10/pypy3-550-released.html) There is also a [`release-pypy3.3-5.5.0` tag](https://bitbucket.org/pypy/pypy/src/ea9979b550eeae87924dc4bef06070e8f8…. Did someone forget to create and upload zips of the release?

1 0

Issue #2430: Use of print via embedded code crashes python (pypy/pypy)
by Whitney Young 07 Nov '16

07 Nov '16

New issue 2430: Use of print via embedded code crashes python https://bitbucket.org/pypy/pypy/issues/2430/use-of-print-via-embedded-code-… Whitney Young: I have issues trying to use `print` both when [embedding PyPy via CFFI](http://cffi.readthedocs.io/en/latest/embedding.html) and when using the [legacy embedding approach](http://doc.pypy.org/en/latest/embedding.html). **CFFI** I've taken the vanilla [embedding code](http://cffi.readthedocs.io/en/latest/embedding.html#usage) and modified it slightly to try to trace the error: ```c /* file plugin.h */ typedef struct { int x, y; } point_t; extern int do_stuff(point_t *); ``` ```python # file plugin_build.py import cffi ffibuilder = cffi.FFI() with open('plugin.h') as f: ffibuilder.embedding_api(f.read()) ffibuilder.set_source("my_plugin", ''' #include "plugin.h" ''') ffibuilder.embedding_init_code(r""" from my_plugin import ffi @ffi.def_extern() def do_stuff(p): import sys with open("/tmp/out", "a") as f: f.write("running embedded\n") sys.stderr.write(b"stderr works\n") sys.stdout.write(b"stdout works\n") try: print("print is broken\n") except Exception as e: f.write("exception\n") f.write(repr(e) + " " + str(e) + "\n") sys.stderr.write(bytes(repr(e) + "\n", encoding="utf8")) print("adding %d and %d" % (p.x, p.y)) return p.x + p.y """) ffibuilder.compile(target="plugin-1.5.*", verbose=True) ``` Compilation of this is like so for me: ``` $ python plugin.py python: /home/wyoung/.linuxbrew/lib/libssl.so.1.0.0: no version information available (required by /home/wyoung/dev/api/env/bin/libpypy-c.so) python: /home/wyoung/.linuxbrew/lib/libssl.so.1.0.0: no version information available (required by /home/wyoung/dev/api/env/bin/libpypy-c.so) python: /home/wyoung/.linuxbrew/lib/libcrypto.so.1.0.0: no version information available (required by /home/wyoung/dev/api/env/bin/libpypy-c.so) generating ./my_plugin.c running build_ext building 'my_plugin' extension cc -O2 -fPIC -Wimplicit -I/home/wyoung/dev/api/env/include -c my_plugin.c -o ./my_plugin.o my_plugin.c:825:5: warning: initialization from incompatible pointer type [enabled by default] _CFFI_PYTHON_STARTUP_FUNC, ^ my_plugin.c:825:5: warning: (near initialization for ‘_cffi_pypy_init.func’) [enabled by default] cc -shared ./my_plugin.o -L/home/wyoung/dev/api/env/bin -L/home/wyoung/dev/api/env/pypy/goal -lpypy-c -o ./plugin-1.5.so ``` I have a small driver program, `test.c` that I'm using to load this: ```c #define _GNU_SOURCE #include "plugin.h" #include <string.h> #include <stdio.h> #include <stdlib.h> #include <dlfcn.h> int main() { void *handle; handle = dlopen("/home/wyoung/.pythons/pypy3-v5.5.0-linux64/bin/libpypy-c.so", RTLD_NOW | RTLD_GLOBAL); if (!handle) { printf("cant find libpypy\n"); exit(1); } handle = dlopen("/home/wyoung/dev/api/plugin-1.5.so", RTLD_NOW | RTLD_GLOBAL); if (!handle) { printf("could not load plugin :(\n"); } point_t the_point; the_point.x = 12; the_point.y = 14; int (*stuff)(point_t *) = dlsym(handle, "do_stuff"); int res = stuff(&the_point); printf("res is %d through function at %p\n", res, stuff); return 0; } ``` When compiled & run: ``` $ gcc test.c -o test -ldl && LD_LIBRARY_PATH="/home/wyoung/.pythons/pypy3-v5.5.0-linux64/bin:$LD_LIBRARY_PATH" ./test stderr works stdout works TypeError('must be bytes or buffer, not str',) res is 0 through 0x7f4d9ef7dcc0 ``` **Legacy Approach** I get the same error as [described on this uWSGI issue](https://github.com/unbit/uwsgi/issues/869#issue-61051504): ``` debug: OperationError: debug: operror-type: AttributeError debug: operror-value: 'module' object has no attribute 'stdout' ``` I would imagine this is related to the above issue.

1 0

Issue #2429: cPickle issue with revdb (pypy/pypy)
by Tristan Seligmann 05 Nov '16

05 Nov '16

New issue 2429: cPickle issue with revdb https://bitbucket.org/pypy/pypy/issues/2429/cpickle-issue-with-revdb Tristan Seligmann: Minimal reproduction: ``` import cPickle import pickle class Cls(object): def __init__(self): self.plugins = [] obj = {'key': Cls()} pickle.dumps(obj) # This works cPickle.dumps(obj) # This does not ``` produces the following traceback: ``` Traceback (most recent call last): File "picklebug.py", line 10, in <module> cPickle.dumps(obj) # This does not File "/home/mithrandi/code/pypy/lib_pypy/cPickle.py", line 124, in dumps Pickler(file, protocol).dump(obj) File "/home/mithrandi/code/pypy/lib-python/2.7/pickle.py", line 224, in dump self.save(obj) File "/home/mithrandi/code/pypy/lib-python/2.7/pickle.py", line 286, in save f(self, obj) # Call unbound method with explicit self File "/home/mithrandi/code/pypy/lib-python/2.7/pickle.py", line 653, in save_dict self._batch_setitems(obj.iteritems()) File "/home/mithrandi/code/pypy/lib-python/2.7/pickle.py", line 667, in _batch_setitems save(v) File "/home/mithrandi/code/pypy/lib-python/2.7/pickle.py", line 331, in save self.save_reduce(obj=obj, *rv) File "/home/mithrandi/code/pypy/lib-python/2.7/pickle.py", line 401, in save_reduce save(args) File "/home/mithrandi/code/pypy/lib-python/2.7/pickle.py", line 286, in save f(self, obj) # Call unbound method with explicit self File "/home/mithrandi/code/pypy/lib-python/2.7/pickle.py", line 562, in save_tuple save(element) File "/home/mithrandi/code/pypy/lib-python/2.7/pickle.py", line 286, in save f(self, obj) # Call unbound method with explicit self File "/home/mithrandi/code/pypy/lib-python/2.7/pickle.py", line 812, in save_global self.memoize(obj) File "/home/mithrandi/code/pypy/lib_pypy/cPickle.py", line 112, in memoize return PythonPickler.memoize(self, obj) File "/home/mithrandi/code/pypy/lib-python/2.7/pickle.py", line 244, in memoize assert id(obj) not in self.memo AssertionError ``` My pypy-revdb was built from a740348ea339 on the reverse-debugger branch; regular pypy 5.4.1 doesn't seem to have the same bug.

1 0

Issue #2428: platform.uname() fails on Windows (pypy/pypy)
by Daniil Yarancev 05 Nov '16

05 Nov '16

New issue 2428: platform.uname() fails on Windows https://bitbucket.org/pypy/pypy/issues/2428/platformuname-fails-on-windows Daniil Yarancev: Using latest nigthly build of PyPy for Win32. This code fails: ``` #!python import platform print(platform.uname()) ``` With this error: ``` #!python Traceback (most recent call last): File "<stdin>", line 1, in <module> File "C:\PyPy\lib-python\2.7\platform.py", line 1159, in uname release,version,csd,ptype = win32_ver() File "C:\PyPy\lib-python\2.7\platform.py", line 635, in win32_ver maj, min, build = _get_real_winver(*winver[:3]) File "C:\PyPy\lib-python\2.7\platform.py", line 600, in _get_real_winver actual_len = kernel32.GetModuleFileNameW(HANDLE(kernel32._handle), File "C:\PyPy\lib_pypy\_ctypes\primitive.py", line 319, in __init__ self.value = value File "C:\PyPy\lib_pypy\_ctypes\primitive.py", line 188, in _setvalue self._buffer[0] = value TypeError: unsupported operand type for long(): '_rawffi.alt.WinDLL' ``` With PyPy v5.4.1 everything works fine

1 0

Issue #2427: pypy3.3: building extensions: environment variables CC etc - not honored (pypy/pypy)
by Danny Milosavljevic 04 Nov '16

04 Nov '16

New issue 2427: pypy3.3: building extensions: environment variables CC etc - not honored https://bitbucket.org/pypy/pypy/issues/2427/pypy33-building-extensions-envi… Danny Milosavljevic: I'm trying to package pypy3.3 for the Guix distribution. In the course of that I found that pypy3.3-v5.2.0-alpha1-src/lib-python/3/distutils/sysconfig_pypy.py doesn't honor the compiler environment variables ("CC" etc) like sysconfig_cpython.py does. Is that on purpose? If not, find attached a patch which makes it honor the variables (I have tested it and it works). Responsible: cfbolz

1 0