Mailman 3 December 2012 - Python-Dev

Re: [Python-Dev] [Python-checkins] cpython (3.3): Issue #16045: add more unit tests for built-in int()
by Terry Reedy Dec. 24, 2012

Dec. 24, 2012

> + # For example, PyPy 1.9.0 raised TypeError for these cases because it > + # expects x to be a string if base is given. > + @support.cpython_only > + def test_base_arg_with_no_x_arg(self): > + self.assertEquals(int(base=6), 0) > + # Even invalid bases don't raise an exception. > + self.assertEquals(int(base=1), 0) > + self.assertEquals(int(base=1000), 0) > + self.assertEquals(int(base='foo'), 0) I think the above behavior is buggy and should be changed rather than frozen into CPython with a test. According to the docs, PyPy does it right. The current online doc gives the signature as int(x=0) int(x, base=10) <<where x is s string>> The 3.3.0 docstring says "When converting a string, use the optional base. It is an error to supply a base when converting a non-string." Certainly, accepting any object as a base, violating "The allowed values are 0 and 2-36." just because giving a base is itself invalid is crazy. -- Terry Jan Reedy

3 4

Running GUI and "GObject.mainloop.run()" together?
by Ajay Garg Dec. 24, 2012

Dec. 24, 2012

Hi all. This is another question that arises as part of my efforts to run a GUI application, as well as a dbus-service within the same process. (The other question being at http://mail.python.org/pipermail/python-dev/2012-December/123287.html) For a recap of the brief history, I have a parent process, that is spawning a child process via "subprocess". Currently, the child-process is a GUI process; however, I intend to "behave" it as a dbus-service as well. Thus, a) I subclassed the child-process "main" class with "dbus.service.Object"; however, I then got the error "metaclass conflict: the metaclass of a derived class must be a (non-strict)subclass of the metaclasses of all its bases " b) I then used composition, wherein another class, "RemoteListener" deriving from "dbus.service.Object" was made an attribute of the "main" class. That worked. However, when I do dbus.mainloop.glib.DBusGMainLoop(set_as_default=True) GObject.mainloop.run() in the "RemoteListener"'s __init__ method, the GUI of the "main" class fails to load (apparently, because the "mainloop.run()" causes the singular, main-thread to go into busy-wait). c) I tried option b), but now instantiating "RemoteListener" in a separate thread; however, no improvement :( Is there a way to run GUI and a dbus-service together? Or is a dbus-service pure "backend" process? :P I will be grateful for any pointers; if you need me to test anything else, please let me know, I will be more than willing :) Regards, Ajay

1 0

[Distutils, Python3] Incorrect shared library extension on linux
by Sandro Mani Dec. 23, 2012

Dec. 23, 2012

Hello, First: I'm using Python3 as available in Fedora rawhide (python3-3.3.0-2.fc19.x86_64). Attempting to build a project using python3/distutils, I noticed that find_library_file would not find any library at all. Some investigation showed that this was due to the fact that libraries were searched with the ".cpython-33m.so" extension. Even more investigation showed that the library extension was read being overridden by the one defined in the /usr/lib64/python3.3/config-3.3m/Makefile shipped by python3-libs. See below for the detailed analysis. The python-versioned extension obviously makes no sense for regular shared objects which are not python binary modules, so this is clearly wrong. As a workaround I commented sysconfig.py@customize_compiler::235 (compiler.shared_lib_extension = so_ext, see below), and things seem to work. Is this a distribution bug or an upstream bug? Thanks, Sandro Detailed analysis: setup.py: def _find_library_file(self, library): return self.compiler.find_library_file(self.compiler.library_dirs, library) --- In function /usr/lib64/python3.3/distutils/unixcompiler.py@find_library_file::266: shared_f = self.library_filename(lib, lib_type='shared') In function /usr/lib64/python3.3/distutils/ccompiler.py@library_filename::882: ext = getattr(self, lib_type + "_lib_extension") -> Where does shared_lib_extension get defined? * At /usr/lib64/python3.3/distutils/ccompiler.py::66 shared_lib_extension = None -> default for abstract class * At /usr/lib64/python3.3/distutils/unixcompiler.py::77 shared_lib_extension = ".so" -> this is the correct value * In function /usr/lib64/python3.3/distutils/sysconfig.py@customize_compiler::235 by /usr/lib64/python3.3/distutils/sysconfig.py@customize_compiler::235 compiler.shared_lib_extension = so_ext by /usr/lib64/python3.3/distutils/sysconfig.py@customize_compiler::194 (cc, cxx, opt, cflags, ccshared, ldshared, so_ext, ar, ar_flags) = \ get_config_vars('CC', 'CXX', 'OPT', 'CFLAGS', 'CCSHARED', 'LDSHARED', 'SO', 'AR', 'ARFLAGS')) by /usr/lib64/python3.3/distutils/sysconfig.py@get_config_vars::530 526 global _config_vars 527 if _config_vars is None: 528 func = globals().get("_init_" + os.name) # -> os.name = posix 529 if func: 530 func() # -> _init_posix, populates _config_vars by /usr/lib64/python3.3/distutils/sysconfig.py@_init_posix::439 435 g = {} 436 # load the installed Makefile: 437 try: 438 filename = get_makefile_filename() # /usr/lib64/python3.3/config-3.3m/Makefile 439 parse_makefile(filename, g) ... 476 global _config_vars 477 _config_vars = g # -> _config_vars["SO"] = ".cpython-33m.so" by /usr/lib64/python3.3/config-3.3m/Makefile::122 SO= .cpython-33m.so

2 1

Re: [Python-Dev] [Python-checkins] Cron <docs@dinsdale> /home/docs/build-devguide
by Barry Warsaw Dec. 22, 2012

Dec. 22, 2012

On Dec 22, 2012, at 03:36 PM, Terry Reedy wrote: >I always reject the requests as I don't believe these messages belong here. I >even asked, some months ago, on pydev who was responsible for the robot that >sends these but got no answer. Today, apparently, another list admin decided >on the opposite response and gave root(a)python.org blanket permission to flood >this list with irrelavancy. It it not my responsibility and I have no idea >how to fix it. > >While people with push priviliges are supposed to subscribe to this list, I >know there is at least one who unsubscribed because of the volume. This will >only encourage more to leave, so I hope someone can stop it. Actually, I made docs(a)dinsdale.python.org an acceptable alias so these messages won't just fill up the hold queue of the list. It did get the problem fixed, didn't it? ;) I don't remember the previous conversation. If folks really don't want those messages hitting the checkins list, then errors should probably be sent to some address that can do something about the problem when they occur. Maybe that's not docs(a)dinsdale.python.org, or maybe that alias should point somewhere else (is that address ever used in the good path?). I have no idea where that address is used. As for the noise issue, well, I hope such failures shouldn't happen very often. We can set up an auto-discard, but then I worry that problems will just go unnoticed for days. -Barry

2 1

Re: [Python-Dev] [Python-checkins] Cron <docs@dinsdale> /home/docs/build-devguide
by M.-A. Lemburg Dec. 22, 2012

Dec. 22, 2012

On 22.12.2012 21:36, Terry Reedy wrote: > > On 12/22/2012 1:30 PM, Cron Daemon wrote: >> abort: error: Connection timed out >> _______________________________________________ >> Python-checkins mailing list >> Python-checkins(a)python.org >> http://mail.python.org/mailman/listinfo/python-checkins > > As a volunteer checkin-list admin, I occasionally get messages like this: > ''' > As list administrator, your authorization is requested for the > following mailing list posting: > > List: Python-checkins(a)python.org > From: root(a)python.org > Subject: Cron <docs@dinsdale> /home/docs/build-devguide > Reason: Message has implicit destination > > At your convenience, visit: > > http://mail.python.org/mailman/admindb/python-checkins > > to approve or deny the request. > ''' > > I always reject the requests as I don't believe these messages belong here. I even asked, some > months ago, on pydev who was responsible for the robot that sends these but got no answer. Today, > apparently, another list admin decided on the opposite response and gave root(a)python.org blanket > permission to flood this list with irrelavancy. It it not my responsibility and I have no idea how > to fix it. You can add a sender filter to have the messages automatically discarded. > While people with push priviliges are supposed to subscribe to this list, I know there is at least > one who unsubscribed because of the volume. This will only encourage more to leave, so I hope > someone can stop it. I think such messages should go to a sys admin list. -- Marc-Andre Lemburg eGenix.com Professional Python Services directly from the Source (#1, Dec 22 2012) >>> Python Projects, Consulting and Support ... http://www.egenix.com/ >>> mxODBC.Zope/Plone.Database.Adapter ... http://zope.egenix.com/ >>> mxODBC, mxDateTime, mxTextTools ... http://python.egenix.com/ ________________________________________________________________________ 2012-12-14: Released mxODBC.Connect 2.0.2 ... http://egenix.com/go38 2013-01-22: Python Meeting Duesseldorf ... 31 days to go ::::: Try our mxODBC.Connect Python Database Interface for free ! :::::: eGenix.com Software, Skills and Services GmbH Pastor-Loeh-Str.48 D-40764 Langenfeld, Germany. CEO Dipl.-Math. Marc-Andre Lemburg Registered at Amtsgericht Duesseldorf: HRB 46611 http://www.egenix.com/company/contact/

1 0

hg.python.org down
by Stefan Krah Dec. 22, 2012

Dec. 22, 2012

Hi, hg.python.org seems to be unreachable (tested from various IP addresses). Stefan Krah

3 2

Testing the tests by modifying the ordering of dict items.
by Mark Shannon Dec. 22, 2012

Dec. 22, 2012

Hi, Python code should not depend upon the ordering of items in a dict. Unfortunately it seems that a number of tests in the standard library do just that. Changing PyDict_MINSIZE from 8 to either 4 or 16 causes the following tests to fail: test_dis test_email test_inspect test_nntplib test_packaging test_plistlib test_pprint test_symtable test_trace test_sys also fails, but this is a legitimate failure in sys.getsizeof() Changing the collision resolution function from f(n) = 5n + 1 to f(n) = n + 1 results in the same failures, except for test_packaging and test_symtable which pass. Finally, changing the seed in unicode_hash() from (implicit) 0 to an arbitrary value (12345678) causes the above tests to fail plus: test_json test_set test_ttk_textonly test_urllib test_urlparse I think this is a real issue as the unicode_hash() function is likely to change soon due to http://bugs.python.org/issue13703. Should I: 1. Submit one big bug report? 2. Submit a bug report for each "failing" test separately? 3. Ignore it, since the tests only fail when I start messing about? Cheers, Mark.

4 4

Re: [Python-Dev] peps: Specify start_serving(). Add Post-History.
by Antoine Pitrou Dec. 21, 2012

Dec. 21, 2012

On Fri, 21 Dec 2012 20:34:18 +0100 (CET) guido.van.rossum <python-checkins(a)python.org> wrote: > > - In either case, once it has a socket, it will wrap it in a > - transport, and then enter a loop accepting connections (the best way > - to implement such a loop depends on the platform). Each time a > - connection is accepted, a transport and protocol are created for it. > + TBD: Support SSL? I don't even know how to do that synchronously, > + and I suppose it needs a certificate. You need a SSLContext, and that SSLContext must have a cert / key pair defined using the `load_cert_chain()` method. I supposed you meant "asynchronously", not "synchronously". The listening socket doesn't have to be a SSL socket, only the connected sockets need to be wrapped. The non-blocking handshake shouldn't be different (AFAICT) for a server SSL socket than for a client SSL socket. Regards Antoine.

3 4

Summary of Python tracker Issues
by Python tracker Dec. 21, 2012

Dec. 21, 2012

ACTIVITY SUMMARY (2012-12-14 - 2012-12-21) Python tracker at http://bugs.python.org/ To view or respond to any of the issues listed below, click on the issue. Do NOT respond to this message. Issues counts and deltas: open 3844 (+18) closed 24677 (+46) total 28521 (+64) Open issues with patches: 1686 Issues opened (38) ================== #16682: Document that audioop works with bytes, not strings http://bugs.python.org/issue16682 opened by serhiy.storchaka #16684: Unicode property value abbreviated names and long names http://bugs.python.org/issue16684 opened by PanderMusubi #16685: Deprecate accepting strings as arguments in audioop functions http://bugs.python.org/issue16685 opened by serhiy.storchaka #16686: audioop overflow issues http://bugs.python.org/issue16686 opened by serhiy.storchaka #16688: Backreferences make case-insensitive regex fail on non-ASCII s http://bugs.python.org/issue16688 opened by pyos #16689: stdout stderr redirection mess http://bugs.python.org/issue16689 opened by techtonik #16690: Reference leak with custom tp_dealloc in PyType_FromSpec http://bugs.python.org/issue16690 opened by bfroehle #16692: Support TLS 1.1 and TLS 1.2 http://bugs.python.org/issue16692 opened by pitrou #16694: Add pure Python operator module http://bugs.python.org/issue16694 opened by zach.ware #16695: Clarify fnmatch & glob docs about the handling of leading "."s http://bugs.python.org/issue16695 opened by hynek #16698: test_posix.test_getgroups fails on some systems http://bugs.python.org/issue16698 opened by rosslagerwall #16699: Mountain Lion buildbot lacks disk space http://bugs.python.org/issue16699 opened by pitrou #16700: Document that bytes OS API can returns unusable results on Win http://bugs.python.org/issue16700 opened by serhiy.storchaka #16701: Docs missing the behavior of += (in-place add) for lists. http://bugs.python.org/issue16701 opened by montysinngh #16702: Force urllib2_localnet test not to use http proxies http://bugs.python.org/issue16702 opened by jeffknupp #16705: Use concrete classes inherited from OSError instead of errno c http://bugs.python.org/issue16705 opened by asvetlov #16709: unittest discover order is filesystem specific - hard to repro http://bugs.python.org/issue16709 opened by rbcollins #16712: collections.abc.Sequence should not provide __reversed__ http://bugs.python.org/issue16712 opened by naoki #16713: "tel" URIs should support params http://bugs.python.org/issue16713 opened by pitrou #16715: Get rid of IOError. Use OSError instead http://bugs.python.org/issue16715 opened by asvetlov #16716: Deprecate OSError aliases in the doc http://bugs.python.org/issue16716 opened by asvetlov #16718: Mysterious atexit fail http://bugs.python.org/issue16718 opened by techtonik #16720: Get rid of os.error. Use OSError instead http://bugs.python.org/issue16720 opened by serhiy.storchaka #16721: configure incorrectly adds -OPT:Olimit=0 for clang http://bugs.python.org/issue16721 opened by Vladimir.Timofeev #16723: io.TextIOWrapper on urllib.request.urlopen terminates prematur http://bugs.python.org/issue16723 opened by mdehoon #16726: expat ParseFile expects bytes, not string http://bugs.python.org/issue16726 opened by mdehoon #16728: Missing cross-reference in sequence glossary entry http://bugs.python.org/issue16728 opened by naoki #16729: Document how to provide defaults for setup.py commands options http://bugs.python.org/issue16729 opened by techtonik #16730: _fill_cache in _bootstrap.py crashes without directory execute http://bugs.python.org/issue16730 opened by David.Pritchard #16731: xxlimited/xxmodule docstrings ambiguous http://bugs.python.org/issue16731 opened by danielsh #16732: setup.py support for xxmodule without tkinker http://bugs.python.org/issue16732 opened by danielsh #16733: Solaris ctypes_test failures http://bugs.python.org/issue16733 opened by yippi #16737: Different behaviours in script run directly and via runpy.run_ http://bugs.python.org/issue16737 opened by vinay.sajip #16739: texttestresult should decorate the stream with _WritelnDecorat http://bugs.python.org/issue16739 opened by elopio #16741: `int()`, `float()`, etc think python strings are null-terminat http://bugs.python.org/issue16741 opened by gangesmaster #16742: PyOS_Readline drops GIL and calls PyOS_StdioReadline, which is http://bugs.python.org/issue16742 opened by trent #16743: mmap accepts files > 1 GB, but processes only 1 GB http://bugs.python.org/issue16743 opened by schlamar #16744: sys.path.append causes wrong behaviour http://bugs.python.org/issue16744 opened by rappy Most recent 15 issues with no replies (15) ========================================== #16744: sys.path.append causes wrong behaviour http://bugs.python.org/issue16744 #16743: mmap accepts files > 1 GB, but processes only 1 GB http://bugs.python.org/issue16743 #16742: PyOS_Readline drops GIL and calls PyOS_StdioReadline, which is http://bugs.python.org/issue16742 #16741: `int()`, `float()`, etc think python strings are null-terminat http://bugs.python.org/issue16741 #16733: Solaris ctypes_test failures http://bugs.python.org/issue16733 #16732: setup.py support for xxmodule without tkinker http://bugs.python.org/issue16732 #16729: Document how to provide defaults for setup.py commands options http://bugs.python.org/issue16729 #16726: expat ParseFile expects bytes, not string http://bugs.python.org/issue16726 #16720: Get rid of os.error. Use OSError instead http://bugs.python.org/issue16720 #16715: Get rid of IOError. Use OSError instead http://bugs.python.org/issue16715 #16712: collections.abc.Sequence should not provide __reversed__ http://bugs.python.org/issue16712 #16702: Force urllib2_localnet test not to use http proxies http://bugs.python.org/issue16702 #16699: Mountain Lion buildbot lacks disk space http://bugs.python.org/issue16699 #16695: Clarify fnmatch & glob docs about the handling of leading "."s http://bugs.python.org/issue16695 #16692: Support TLS 1.1 and TLS 1.2 http://bugs.python.org/issue16692 Most recent 15 issues waiting for review (15) ============================================= #16739: texttestresult should decorate the stream with _WritelnDecorat http://bugs.python.org/issue16739 #16732: setup.py support for xxmodule without tkinker http://bugs.python.org/issue16732 #16731: xxlimited/xxmodule docstrings ambiguous http://bugs.python.org/issue16731 #16730: _fill_cache in _bootstrap.py crashes without directory execute http://bugs.python.org/issue16730 #16720: Get rid of os.error. Use OSError instead http://bugs.python.org/issue16720 #16702: Force urllib2_localnet test not to use http proxies http://bugs.python.org/issue16702 #16694: Add pure Python operator module http://bugs.python.org/issue16694 #16688: Backreferences make case-insensitive regex fail on non-ASCII s http://bugs.python.org/issue16688 #16686: audioop overflow issues http://bugs.python.org/issue16686 #16682: Document that audioop works with bytes, not strings http://bugs.python.org/issue16682 #16679: Wrong URL path decoding http://bugs.python.org/issue16679 #16674: Faster getrandbits() for small integers http://bugs.python.org/issue16674 #16672: improve tracing performances when f_trace is NULL http://bugs.python.org/issue16672 #16669: Docstrings for namedtuple http://bugs.python.org/issue16669 #16667: timezone docs need "versionadded: 3.2" http://bugs.python.org/issue16667 Top 10 most discussed issues (10) ================================= #16688: Backreferences make case-insensitive regex fail on non-ASCII s http://bugs.python.org/issue16688 13 msgs #16718: Mysterious atexit fail http://bugs.python.org/issue16718 13 msgs #16728: Missing cross-reference in sequence glossary entry http://bugs.python.org/issue16728 9 msgs #16612: Integrate "Argument Clinic" specialized preprocessor into CPyt http://bugs.python.org/issue16612 7 msgs #16685: Deprecate accepting strings as arguments in audioop functions http://bugs.python.org/issue16685 7 msgs #16694: Add pure Python operator module http://bugs.python.org/issue16694 7 msgs #15533: subprocess.Popen(cwd) documentation http://bugs.python.org/issue15533 5 msgs #16618: Different glob() results for strings and bytes http://bugs.python.org/issue16618 5 msgs #16669: Docstrings for namedtuple http://bugs.python.org/issue16669 4 msgs #16716: Deprecate OSError aliases in the doc http://bugs.python.org/issue16716 4 msgs Issues closed (43) ================== #8853: getaddrinfo should accept port of type long http://bugs.python.org/issue8853 closed by petri.lehtinen #10155: Add fixups for encoding problems to wsgiref http://bugs.python.org/issue10155 closed by aclover #11175: allow argparse FileType to accept encoding and errors argument http://bugs.python.org/issue11175 closed by petri.lehtinen #14901: Python Windows FAQ is Very Outdated http://bugs.python.org/issue14901 closed by brian.curtin #15743: test_urllib2/test_urllib use deprecated urllib.Request methods http://bugs.python.org/issue15743 closed by jeffknupp #15783: decimal: Support None default values in the C accelerator modu http://bugs.python.org/issue15783 closed by skrah #16298: httplib.HTTPResponse.read could potentially leave the socket o http://bugs.python.org/issue16298 closed by pitrou #16480: pyvenv 3.3 fails to create symlinks for <virtualenv>/local/{bi http://bugs.python.org/issue16480 closed by doko #16488: Add context manager support to epoll object http://bugs.python.org/issue16488 closed by pitrou #16597: file descriptor not being closed with context manager on IOErr http://bugs.python.org/issue16597 closed by python-dev #16626: Infinite recursion in glob.glob('*:') on Windows http://bugs.python.org/issue16626 closed by pitrou #16646: FTP.makeport() loses socket error details http://bugs.python.org/issue16646 closed by giampaolo.rodola #16647: LMTP.connect() loses socket error details http://bugs.python.org/issue16647 closed by asvetlov #16661: test_posix.test_getgrouplist fails on some systems - incorrect http://bugs.python.org/issue16661 closed by rosslagerwall #16664: Test Glob: files starting with . http://bugs.python.org/issue16664 closed by hynek #16670: Point class may be not be a good example for namedtuple http://bugs.python.org/issue16670 closed by terry.reedy #16678: optparse: parse only known options http://bugs.python.org/issue16678 closed by r.david.murray #16681: Documentation 'bidirectional category' should be 'bidirectiona http://bugs.python.org/issue16681 closed by ezio.melotti #16683: Resort audioop documentation http://bugs.python.org/issue16683 closed by ezio.melotti #16687: Fix small gramatical error and add reference link in hashlib d http://bugs.python.org/issue16687 closed by python-dev #16691: How to use ctypes.windll.user32.MessageBoxW http://bugs.python.org/issue16691 closed by ned.deily #16693: Assertion error in ceval if Chainmap(object()) used as locals http://bugs.python.org/issue16693 closed by python-dev #16696: BytesWarning in glob.glob http://bugs.python.org/issue16696 closed by pitrou #16697: argparse kwarg 'choices' documentation http://bugs.python.org/issue16697 closed by r.david.murray #16703: except statement turns defined variable into undefined http://bugs.python.org/issue16703 closed by brett.cannon #16704: Get rid of select.error in stdlib. Use OSError instead http://bugs.python.org/issue16704 closed by asvetlov #16706: Get rid of os.error. Use OSError instead http://bugs.python.org/issue16706 closed by asvetlov #16707: --with-pydebug and --without-pymalloc are incompatible http://bugs.python.org/issue16707 closed by pitrou #16708: Module: shutil will not import when writen in the text editor http://bugs.python.org/issue16708 closed by r.david.murray #16710: json encode/decode error http://bugs.python.org/issue16710 closed by amaury.forgeotdarc #16711: s/next()/__next__/ in collections.abc.Iterator document. http://bugs.python.org/issue16711 closed by asvetlov #16714: Raise exceptions, don't throw http://bugs.python.org/issue16714 closed by asvetlov #16717: Get rid of socket.error. Use OSError instead http://bugs.python.org/issue16717 closed by asvetlov #16719: Get rid of WindowsError. Use OSError instead http://bugs.python.org/issue16719 closed by asvetlov #16722: __index__() overrides __bytes__() when bytes() is called http://bugs.python.org/issue16722 closed by python-dev #16724: Define `binary data` representation in Python http://bugs.python.org/issue16724 closed by r.david.murray #16725: Add 'ident' property to SysLogHandler like in Python 3.x http://bugs.python.org/issue16725 closed by r.david.murray #16727: Windows installers for 2.7.3 don't install python27.dll correc http://bugs.python.org/issue16727 closed by loewis #16734: Delay interpreter startup phase until script is read http://bugs.python.org/issue16734 closed by christian.heimes #16735: zipfile.is_zipfile wrongly recognizes non-zip as zip http://bugs.python.org/issue16735 closed by r.david.murray #16736: select.poll() converts long to int without checking for overfl http://bugs.python.org/issue16736 closed by sbt #16738: Comparisons difference: bytes with bytes, str with str http://bugs.python.org/issue16738 closed by christian.heimes #16740: Types created with PyType_FromSpec lack a __module__ attribute http://bugs.python.org/issue16740 closed by bfroehle

1 0

Possible GIL/threading issue involving subprocess and PyMem_MALLOC...
by Trent Nelson Dec. 21, 2012

Dec. 21, 2012

This seems odd to me so I wanted to see what others think. The unit test Lib/unittest/test/test_runner.py:Test_TextRunner.test_warnings will eventually hit subprocess.Popen._communicate. The `mswindows` implementation of this method relies on threads to buffer stdin/stdout. That'll eventually result in PyOs_StdioReadline being called without the GIL being held. PyOs_StdioReadline calls PyMem_MALLOC, PyMem_FREE and possibly PyMem_REALLOC. On a debug build, these macros are redirected to their _PyMem_Debug* counterparts. The call hierarchy for _PyMem_DebugMalloc looks like this: void * _PyMem_DebugMalloc(size_t nbytes) { return _PyObject_DebugMallocApi(_PYMALLOC_MEM_ID, nbytes); } /* generic debug memory api, with an "id" to identify the API in use */ void * _PyObject_DebugMallocApi(char id, size_t nbytes) { uchar *p; /* base address of malloc'ed block */ uchar *tail; /* p + 2*SST + nbytes == pointer to tail pad bytes */ size_t total; /* nbytes + 4*SST */ bumpserialno(); ------------^^^^^^^^^^^^^^^ total = nbytes + 4*SST; if (total < nbytes) /* overflow: can't represent total as a size_t */ return NULL; p = (uchar *)PyObject_Malloc(total); -------------------------^^^^^^^^^^^^^^^^^^^^^^^ if (p == NULL) return NULL; <snip> Both bumpserialno() and PyObject_Malloc affect global state. The latter also has a bunch of LOCK() and UNLOCK() statements, but these end up being no-ops: /* * Python's threads are serialized, * so object malloc locking is disabled. */ #define SIMPLELOCK_DECL(lock) /* simple lock declaration */ #define SIMPLELOCK_INIT(lock) /* allocate (if needed) and ... */ #define SIMPLELOCK_FINI(lock) /* free/destroy an existing */ #define SIMPLELOCK_LOCK(lock) /* acquire released lock */ #define SIMPLELOCK_UNLOCK(lock) /* release acquired lock */ ... /* * This malloc lock */ SIMPLELOCK_DECL(_malloc_lock) #define LOCK() SIMPLELOCK_LOCK(_malloc_lock) #define UNLOCK() SIMPLELOCK_UNLOCK(_malloc_lock) #define LOCK_INIT() SIMPLELOCK_INIT(_malloc_lock) #define LOCK_FINI() SIMPLELOCK_FINI(_malloc_lock) The PyObject_Malloc() one concerns me the most, as it affects huge amounts of global state. Also, I just noticed PyOs_StdioReadline() can call PyErr_SetString, which will result in a bunch of other calls that should only be made whilst the GIL is held. So, like I said, this seems like a bit of a head scratcher. Legit issue or am I missing something? Trent.

4 5