Mailman 3 July 2010 - Python-Dev

More C API abstraction for user defined types
by Petre Galan July 17, 2010

July 17, 2010

Hello, Defining a user type which implements number protocol should be treated as a number (long) in PyArg_ParseTuple using "l" format. In getargs.c: case 'l': {/* long int */ long *p = va_arg(*p_va, long *); long ival; if (float_argument_error(arg)) return converterr("integer<l>", arg, msgbuf, bufsize); ival = PyLong_AsLong(arg); if (ival == -1 && PyErr_Occurred()) return converterr("integer<l>", arg, msgbuf, bufsize); else *p = ival; break; } ival should not be resolved through PyLong_AsLong, but through functionality/interface like PyNumber_Long, thus allowing more diversity in accessing the PyArg_Parse interfaces. Petre

3 7

profiler decorator - is it worth for inclusion?
by Giampaolo Rodolà July 17, 2010

July 17, 2010

Today I was looking for a quick and dirty way to profile a method of a class. I was thinking that cProfile module had a decorator for this but I was wrong so I decided to write one based on hotshot. Would it be worth for inclusion? #!/usr/bin/env python import hotshot import hotshot.stats import tempfile import pstats def profile(sort='cumulative', lines=30, strip_dirs=False): """A decorator which profiles a callable. Example usage: >>> @profile() ... def factorial(n): ... n = abs(int(n)) ... if n < 1: ... n = 1 ... x = 1 ... for i in range(1, n+1): ... x = i * x ... return x ... >>> factorial(5) 1 function calls in 0.000 CPU seconds Ordered by: internal time, call count ncalls tottime percall cumtime percall filename:lineno(function) 1 0.000 0.000 0.000 0.000 profiler.py:60(factorial) 0 0.000 0.000 profile:0(profiler) 120 >>> """ def outer(fun): def inner(*args, **kwargs): file = tempfile.NamedTemporaryFile() prof = hotshot.Profile(file.name) try: ret = prof.runcall(fun, *args, **kwargs) finally: prof.close() stats = hotshot.stats.load(file.name) if strip_dirs: stats.strip_dirs() if isinstance(sort, tuple): stats.sort_stats(*sort) else: stats.sort_stats(sort) stats.print_stats(lines) return ret return inner return outer

2 3

A suggestion
by Brandon Hayden July 17, 2010

July 17, 2010

In the Python language, or any other language for that matter, I have never understood why they don't have a loop function. Here's what I mean, every time someone wants something to repeat itself, they have to write a while loop like this: x = 0 while x < 10: function() x = x + 1 When, instead of writing all of that, you could do something similar to this: function().10 Where you can just type in the number 10 right next to it. It makes no since.

4 3

query: docstring formatting in python distutils code
by C. Titus Brown July 16, 2010

July 16, 2010

Hi all, over on the fellowship o' the packaging mailing list, one of our GSoC students (merwok) asked about how much formatting info should go into Python stdlib docstrings. Right now the stdlib docstrings are primarily text, AFAIK; but with the switch to Sphinx for the official Python docs, should we permit ReST-general and/or Sphinx-specific markup in docstrings? Hmm, I don't actually see that the stdlib docstrings are imported into the Python documentation anywhere, so maybe the use of Sphinx isn't that relevant. But how about ReST in general? See http://sphinx.pocoo.org/markup/index.html for sphinx-specific markup constructs. thanks, --titus -- C. Titus Brown, ctb(a)msu.edu

19 44

Peculiar import code in pickle.py
by Alexander Belopolsky July 16, 2010

July 16, 2010

When pickle.py needs to import a module by name, it goes through a peculiar dance of __import__(module, level=0) mod = sys.modules[module] As far as I can tell, unless builtins.__import__ is overridden or sys.modules clobbered by user code, the above should be equivalent to mod = __import__(module, level=0) Note that the optimized _pickle implementation does not do the sys.modules lookup and simply uses the module returned by __import__(..). This code goes back to 1999, so there was probably a good reason back then to write it this way. Presently, however, it seems to be just another obscure difference between C and Python implementation of pickle.

9 17

more details of list comprehension in tutorial than in language reference
by Peng Yu July 16, 2010

July 16, 2010

Hi, Since this is related to the document, I think that I should send this to the dev mailing list. Please let me know if this is not appropriate. If I don't miss anything, I feel that there are much less descriptions of list comprehensions in the language reference (python_2.6.5_reference.pdf, glossary) than in the tutorial (python_2.6.5_tutorial.pdf). I think that it deserves at least a subsection in the language reference to make the reference complete. May I suggest the maintainer of the document add such a subsection? -- Regards, Peng

3 2

Summary of Python tracker Issues
by Python tracker July 16, 2010

July 16, 2010

ACTIVITY SUMMARY (2010-07-09 - 2010-07-16) Python tracker at http://bugs.python.org/ To view or respond to any of the issues listed below, click on the issue number. Do NOT respond to this message. 2807 open (+44) / 18285 closed (+18) / 21092 total (+62) Open issues with patches: 1144 Average duration of open issues: 703 days. Median duration of open issues: 497 days. Open Issues Breakdown open 2765 (+42) languishing 14 ( +0) pending 27 ( +2) Issues Created Or Reopened (64) _______________________________ regrtest: add a minimal "progress bar" 2010-07-10 http://bugs.python.org/issue8560 reopened georg.brandl patch contextlib.ContextDecorator 2010-07-15 http://bugs.python.org/issue9110 reopened ncoghlan patch dict_keys purports to implement the Set ABC, but is missing th 2010-07-09 http://bugs.python.org/issue9212 created stutzbach range purports to implement the Sequence ABC, but is missing i 2010-07-09 http://bugs.python.org/issue9213 created stutzbach Most Set methods of KeysView and ItemsView do not work right 2010-07-09 http://bugs.python.org/issue9214 created stutzbach easy test_float crashes with assertion failure on Ubuntu buildbot. 2010-07-09 CLOSED http://bugs.python.org/issue9215 created mark.dickinson FIPS support for hashlib 2010-07-10 http://bugs.python.org/issue9216 created dmalcolm patch 2to3 crashes with some doctests 2010-07-14 http://bugs.python.org/issue9217 reopened tiagoantao pop multiple elements of a list at once 2010-07-10 http://bugs.python.org/issue9218 created jacobidiego Make reindent-rst.py preserve ^L 2010-07-10 CLOSED http://bugs.python.org/issue9219 created merwok easy Modules no longer usable as context managers 2010-07-10 CLOSED http://bugs.python.org/issue9220 created j1m 2to3 fails to parse a for construct 2010-07-10 CLOSED http://bugs.python.org/issue9221 created tiagoantao IDLE: Fix open/saveas 'Files of type' choices 2010-07-11 http://bugs.python.org/issue9222 created tjreedy patch, easy distutils Command docs linking 2010-07-11 CLOSED http://bugs.python.org/issue9223 created techtonik Distutls fails with MSVC++ 2008on Windows Vista 64bit 2010-07-11 CLOSED http://bugs.python.org/issue9224 created ipatrol Replace DUP_TOPX with DUP_TOP_TWO 2010-07-11 http://bugs.python.org/issue9225 created serprex patch erroneous behavior when creating classes inside a closure 2010-07-11 http://bugs.python.org/issue9226 created monsanto patch can't import Tkinter / use IDLE after installing Python 2.7 on 2010-07-11 http://bugs.python.org/issue9227 created michael.foord Make changes in the path and pathext on installation 2010-07-11 http://bugs.python.org/issue9228 created ipatrol memoryview of mmap object not working in 2.7 2010-07-12 CLOSED http://bugs.python.org/issue9229 created Matt.Gattis pdb.Pdb.checkline() throws AttributeError if called before sta 2010-07-12 http://bugs.python.org/issue9230 created __Vano patch IDLE interactive interpreter crash 2010-07-12 CLOSED http://bugs.python.org/issue9231 created pitrou Allow trailing comma in any function argument list. 2010-07-12 http://bugs.python.org/issue9232 created mark.dickinson patch json.load failure when C optimizations aren't built 2010-07-12 http://bugs.python.org/issue9233 created fdrake patch argparse: aliases for positional arguments (subparsers) 2010-07-12 http://bugs.python.org/issue9234 created asampson missing "import sys" in Tools/gdb/libpython.py 2010-07-12 CLOSED http://bugs.python.org/issue9235 created Yaroslav.Halchenko Invalid reads in fastsearch.h 2010-07-12 CLOSED http://bugs.python.org/issue9236 created skrah Add sys.call_tracing to on-line sys module documentation 2010-07-12 http://bugs.python.org/issue9237 created rocky zipfile incorrectly documented as not supporting archive comme 2010-07-12 http://bugs.python.org/issue9238 created rfk patch zipfile: truncating comment can corrupt the zipfile 2010-07-12 http://bugs.python.org/issue9239 created rfk patch different behaviour with for loop... interpreted vs scripted 2010-07-13 CLOSED http://bugs.python.org/issue9240 created jacobidiego SAXParseError on unicode (Japanese) file 2010-07-13 CLOSED http://bugs.python.org/issue9241 created gianzula unicodeobject.c: use of uninitialized values 2010-07-13 http://bugs.python.org/issue9242 created skrah sndhdr erroneously detects files as vox 2010-07-13 CLOSED http://bugs.python.org/issue9243 created jbit patch multiprocessing.pool: Worker crashes if result can't be encode 2010-07-13 http://bugs.python.org/issue9244 created asksol patch "from itertools imoprt *" crashes 2to3 2010-07-13 CLOSED http://bugs.python.org/issue9245 created ronaldoussoren os.getcwd() hardcodes max path len 2010-07-13 http://bugs.python.org/issue9246 created pitrou imp.reload() doesn't take import lock 2010-07-13 http://bugs.python.org/issue9247 created pitrou multiprocessing.pool: Proposal: "waitforslot" 2010-07-13 http://bugs.python.org/issue9248 created asksol patch struct.pack and Long Integer datatype should be 4, but is 8 by 2010-07-13 CLOSED http://bugs.python.org/issue9249 created hannes.reuter sys.modules changes size during iteration in regrtest module 2010-07-13 http://bugs.python.org/issue9250 created xdegaye patch Test for the import lock 2010-07-13 http://bugs.python.org/issue9251 created pitrou patch PyImport_Import calls __import__ with dummy fromlist 2010-07-13 http://bugs.python.org/issue9252 created belopolsky argparse: optional subparsers 2010-07-13 http://bugs.python.org/issue9253 created nvie __import__ docstring should recommend importlib.import_module( 2010-07-13 http://bugs.python.org/issue9254 created belopolsky document that the 'test' package is for core developers only 2010-07-13 http://bugs.python.org/issue9255 created brett.cannon easy plistlib should create non-naïve datetime objects 2010-07-14 http://bugs.python.org/issue9256 created wiml cElementTree iterparse requires events as bytes; ElementTree u 2010-07-14 http://bugs.python.org/issue9257 created eric-talevich Typos in docs for methods kqueue and kevent of module 'select' 2010-07-14 CLOSED http://bugs.python.org/issue9258 created Retro Python 2.7 breaks assigned __exit__s 2010-07-14 CLOSED http://bugs.python.org/issue9259 reopened arielbyd A finer grained import lock 2010-07-14 http://bugs.python.org/issue9260 created pitrou patch include higher (../../) dirs fails 2010-07-14 CLOSED http://bugs.python.org/issue9261 created indiedan IDLE: Use tabbed shell and edit windows 2010-07-14 http://bugs.python.org/issue9262 created tjreedy Try to print repr() when an C-level assert fails (in the garba 2010-07-14 http://bugs.python.org/issue9263 created dmalcolm patch trace.py documentation is incomplete 2010-07-15 http://bugs.python.org/issue9264 created eli.bendersky patch Incorrect name passed as arg[0] when shell=True and executable 2010-07-15 http://bugs.python.org/issue9265 created narnie patch, needs review ctypes "ValueError: NULL pointer access" on Win7 x64 2010-07-15 http://bugs.python.org/issue9266 created ghazel Update pickle opcode documentation in pickletools for 3.x 2010-07-15 http://bugs.python.org/issue9267 created belopolsky easy Add annotation option to pickletools.dis 2010-07-15 http://bugs.python.org/issue9268 created belopolsky patch, easy Cannot pickle self-referencing sets 2010-07-15 http://bugs.python.org/issue9269 created belopolsky distutils.util.getplatform and sysconfig.getplatform differ 2010-07-16 http://bugs.python.org/issue9270 created bugs-python(a)vendor.thewrittenword.com Python throws `IOError: [Errno 27] File too large' on long fil 2010-07-16 http://bugs.python.org/issue9271 created Jason.Gross CGIHTTPServer poisons os.environ 2010-07-16 http://bugs.python.org/issue9272 created techtonik patch 2to to simultaneously do code AND doctests 2010-07-16 http://bugs.python.org/issue9273 created tiagoantao Issues Now Closed (83) ______________________ epoll and kqueue wrappers for the select module 935 days http://bugs.python.org/issue1657 tjreedy ConfigParser.py do not allow leading (and trailing) space in v 922 days http://bugs.python.org/issue1714 tjreedy patch, easy Grammar change to prevent shift/reduce problem with varargslis 889 days http://bugs.python.org/issue2009 mark.dickinson patch MozillaCookieJar ignore HttpOnly cookies 868 days http://bugs.python.org/issue2190 jjlee patch MozillaCookieJar does not support Firefox3 cookie files 852 days http://bugs.python.org/issue2277 jjlee patch Augmenting the Windows build to support code signing. 787 days http://bugs.python.org/issue2850 tjreedy Pickle stream for unicode object may contain non-ASCII charact 774 days http://bugs.python.org/issue2980 tjreedy Let SimpleXMLRPCServer pass client_address to called functions 765 days http://bugs.python.org/issue3058 tjreedy Wrong size calculation in posix_execve 761 days http://bugs.python.org/issue3070 haypo The ValueError raised by failing to unpack sequence should hav 761 days http://bugs.python.org/issue3071 merwok patch Suggest change to glossary explanation: "Duck Typing" 744 days http://bugs.python.org/issue3214 merwok patch "make check" suggest a testing target under GNU coding standar 646 days http://bugs.python.org/issue3758 merwok patch test_output_textcalendar fails on non-englisch locale 644 days http://bugs.python.org/issue4044 r.david.murray test_mboxmmdf_to_maildir fails on non-englisch locale 644 days http://bugs.python.org/issue4045 r.david.murray Clarification needed for subprocess convenience functions in P 574 days http://bugs.python.org/issue4697 georg.brandl Little improvement on urlparse module, urlparse function. 547 days http://bugs.python.org/issue4932 orsenthil patch datetime lacks concrete tzinfo implementation for UTC 25 days http://bugs.python.org/issue5094 merwok patch expose _CrtSetReportMode via the msvcrt module 526 days http://bugs.python.org/issue5116 tjreedy IDLE to support reindent.py 521 days http://bugs.python.org/issue5150 merwok SocketIO redefines RawIOBase.__del__ 502 days http://bugs.python.org/issue5347 benjamin.peterson Subprocess.getstatusoutput Fails Executing 'dir' Command on Wi 451 days http://bugs.python.org/issue5808 matthieu.labbe Move test outside of urlparse module 445 days http://bugs.python.org/issue5842 merwok test_(zipfile|zipimport|gzip|distutils|sqlite) fail if zlib is 424 days http://bugs.python.org/issue6026 afarrell socket.shudown documentation: on some platforms, closing one h 321 days http://bugs.python.org/issue6774 georg.brandl patch system proxy not used for https (on windows) 311 days http://bugs.python.org/issue6853 orsenthil patch test_telnetlib gives spurious output 296 days http://bugs.python.org/issue6960 jackdied 2to3 execfile conversion changes behavior 235 days http://bugs.python.org/issue7268 benjamin.peterson patch test_memoryview test_setitem_writable failures with Intel ICC 191 days http://bugs.python.org/issue7616 pitrou patch Lib/ctypes/test/test_pep3118.py should not shadow the memoryvi 159 days http://bugs.python.org/issue7840 BreamoreBoy Cross-reference ast.literal_eval() from eval() docs 146 days http://bugs.python.org/issue7935 georg.brandl Problem with urlparse in Windows XP after a drag and drop 145 days http://bugs.python.org/issue7965 orsenthil test_ntpath fails on Windows XP 143 days http://bugs.python.org/issue7991 austin987 python 2.6.4 installation is not working for the single user m 124 days http://bugs.python.org/issue8114 BreamoreBoy Outdated information 94 days http://bugs.python.org/issue8338 georg.brandl patch sqlite3.connect documentation is incorrect 82 days http://bugs.python.org/issue8456 georg.brandl Update documentation on doctest/unittest2 integration 73 days http://bugs.python.org/issue8564 merwok patch cPickle dumps(tuple) != dumps(loads(dumps(tuple))) 59 days http://bugs.python.org/issue8738 belopolsky Inconsistency in behaviour of urllib and urllib2 with file:// 48 days http://bugs.python.org/issue8801 orsenthil svnmerge errors in msgfmt.py 30 days http://bugs.python.org/issue8974 loewis patch Bug in cookiejar 31 days http://bugs.python.org/issue8975 merwok Provide parseable repr to datetime.timezone 25 days http://bugs.python.org/issue9000 merwok patch, easy Documentation for comparing dictionaries is out of date 14 days http://bugs.python.org/issue9132 orsenthil patch, easy x.update(red=5, blue=6, other=7) doesn't work, where x is a Mu 10 days http://bugs.python.org/issue9137 mark.dickinson patch 2.7 sysconfig should handle arch duplicates while building uni 10 days http://bugs.python.org/issue9164 ronaldoussoren Add math.isfinite() 6 days http://bugs.python.org/issue9165 mark.dickinson patch Deprecate use of more than 3 positional arguments in timedelta 8 days http://bugs.python.org/issue9169 belopolsky unittest assertEqual error output sub-optimal for single strin 5 days http://bugs.python.org/issue9174 michael.foord patch open() doc: default value for buffering is -1, not None 5 days http://bugs.python.org/issue9184 georg.brandl patch Improve CFLAGS handling 2 days http://bugs.python.org/issue9189 jyasskin test_float crashes with assertion failure on Ubuntu buildbot. 3 days http://bugs.python.org/issue9215 mark.dickinson Make reindent-rst.py preserve ^L 0 days http://bugs.python.org/issue9219 merwok easy Modules no longer usable as context managers 0 days http://bugs.python.org/issue9220 j1m 2to3 fails to parse a for construct 0 days http://bugs.python.org/issue9221 merwok distutils Command docs linking 0 days http://bugs.python.org/issue9223 techtonik Distutls fails with MSVC++ 2008on Windows Vista 64bit 0 days http://bugs.python.org/issue9224 srid memoryview of mmap object not working in 2.7 0 days http://bugs.python.org/issue9229 pitrou IDLE interactive interpreter crash 0 days http://bugs.python.org/issue9231 ezio.melotti missing "import sys" in Tools/gdb/libpython.py 2 days http://bugs.python.org/issue9235 georg.brandl Invalid reads in fastsearch.h 3 days http://bugs.python.org/issue9236 skrah different behaviour with for loop... interpreted vs scripted 0 days http://bugs.python.org/issue9240 eric.smith SAXParseError on unicode (Japanese) file 0 days http://bugs.python.org/issue9241 amaury.forgeotdarc sndhdr erroneously detects files as vox 0 days http://bugs.python.org/issue9243 haypo patch "from itertools imoprt *" crashes 2to3 0 days http://bugs.python.org/issue9245 benjamin.peterson struct.pack and Long Integer datatype should be 4, but is 8 by 2 days http://bugs.python.org/issue9249 hannes.reuter Typos in docs for methods kqueue and kevent of module 'select' 0 days http://bugs.python.org/issue9258 georg.brandl Python 2.7 breaks assigned __exit__s 0 days http://bugs.python.org/issue9259 merwok include higher (../../) dirs fails 0 days http://bugs.python.org/issue9261 merwok commands.getstatusoutput(): cmd.exe support 3099 days http://bugs.python.org/issue506100 matthieu.labbe reading shelves is really slow 2420 days http://bugs.python.org/issue849662 rhettinger logging handlers raise exception on level 2373 days http://bugs.python.org/issue876421 vinay.sajip fails building C++ extensions when configured --without-cxx 2330 days http://bugs.python.org/issue902065 georg.brandl patch test test_codecs failed 2050 days http://bugs.python.org/issue1076790 lemburg patch Windows deadlock with PyEval_ReleaseLock 1964 days http://bugs.python.org/issue1147646 amaury.forgeotdarc urllib.urlencode provides two features in one param 1711 days http://bugs.python.org/issue1349732 orsenthil easy urllib/urllib2: HTTPS over (Squid) Proxy fails 1616 days http://bugs.python.org/issue1424152 tjreedy patch DOM tree inconsistency in expat XML parser 1603 days http://bugs.python.org/issue1434090 georg.brandl patch, easy "reindent.py" exposes bug in tokenize 1583 days http://bugs.python.org/issue1447633 merwok gettext.py breaks on plural-forms header (PATCH) 1587 days http://bugs.python.org/issue1448060 merwok patch mimetools module getencoding error 1545 days http://bugs.python.org/issue1471985 georg.brandl support all of strftime(3) 1465 days http://bugs.python.org/issue1520662 belopolsky socket.getaddrinfo() should take an address tuple. 1206 days http://bugs.python.org/issue1685962 giampaolo.rodola Expose callback API in readline module 1161 days http://bugs.python.org/issue1713877 georg.brandl move intern to sys, make intern() optionally warn 1136 days http://bugs.python.org/issue1728741 benjamin.peterson patch Top Issues Most Discussed (10) ______________________________ 26 Parent process hanging in multiprocessing if children terminate 8 days open http://bugs.python.org/issue9205 12 Incorrect name passed as arg[0] when shell=True and executable 1 days open http://bugs.python.org/issue9265 11 A finer grained import lock 2 days open http://bugs.python.org/issue9260 11 contextlib.ContextDecorator 1 days open http://bugs.python.org/issue9110 10 Condition.wait() doesn't raise KeyboardInterrupt 49 days open http://bugs.python.org/issue8844 10 msvc9compiler.py: ValueError: [u'path'] 213 days open http://bugs.python.org/issue7511 9 os.getcwd causes infinite loop on solaris 9 days open http://bugs.python.org/issue9185 9 PyImport_ImportModuleNoBlock() may solve problems but causes ot 129 days open http://bugs.python.org/issue8098 9 Finding programs in PATH, adding shutil.which 3278 days open http://bugs.python.org/issue444582 8 Python 2.7 breaks assigned __exit__s 0 days closed http://bugs.python.org/issue9259

1 0

Re: [Python-Dev] Removing IDLE from the standard library
by John Nagle July 15, 2010

July 15, 2010

On Sun, Jul 11, 2010 at 18:20, Tal Einat <taleinat(a)gmail.com> wrote: > > The (hopefully) compelling arguments were others, such as the sentence > > following the one you quoted: > > > > "I think that in its current state, IDLE may still be helpful for > > learning Python, but it is more likely to drive away users who run > > into its various quirks and problems." The underlying problem is that, given the way Python is currently maintained, there are only two options: 1) part of the main Python distribution and synchronized with it, and 2) supported (or not) by some third party who may or may not produce a working version in sync with what the Python distribution is doing. There's no intermediate tier, like "Python extras", with a code base under common source control and some quality control. (PyPi is mostly a link farm to projects elsewhere, not a source repository.) IDLE belongs in such an intermediate tier. There are no dependencies on IDLE. Nothing will stop working because a version of IDLE isn't available. There are a reasonable number of modules not in the Python distribution that are more mission-critical to large numbers of users than IDLE. IDLE is useful, but not critical. Given limited resources, it's necessary to cut back in some areas. I'd vote IDLE off the island. John Nagle

12 17

Whither 'trunk'
by Terry Reedy July 15, 2010

July 15, 2010

The 'trunk' branch appears to have been frozen 12 days ago when 2.7 was released. I presume py3k is now the main development branch. Correct? There are doc(s) on the site the directed people to the 'trunk' branch. If not updated (as seems from a python-list post today, but I asked the OP), it/they should be. -- Terry Jan Reedy

2 1

Re: [Python-Dev] importlib
by John Nagle July 15, 2010

July 15, 2010

anatoly techtonik<techtonik(a)gmail.com> wrote: > What about web-applications? Is that true that for FastCgi or mod_wsgi > deamon mode interpreter and application is started only once per say > 100 requests? Yes. Only CGI programs reload on every use. FCGI/WSGI programs run more or less forever, and are only loaded once, reading requests and handling them. Some FCGI systems reload their applications once in a while, and ask the idle ones to exit after periods of no traffic, but that's a form of automatic performance tuning. And, of course, there are the permanently resident systems like Twisted. (Incidentally, if you run FCGI/WSGI programs under Apache, make sure that you don't put .fcgi scripts in a directory from which Apache will execute .cgi scripts. Apache will run the .fcgi script with CGI, and the WSGI Python module will detect this and run the script one time and exit for each request. Everything appears to work and performance is terrible.) If import performance is important in a production server environment, it's more useful to look at minimizing the number of loads rather than trying to make them faster. If and when Unladen Swallow finally gets going, startup times will get worse as the JIT compiler cranks, but run time will (hopefully) decrease. John Nagle

1 0