🌐 Mailman 3 December 2006 - Python-list - python.org

Re: PEP 3107 Function Annotations for review and comment
by John Roth Dec. 31, 2006

Dec. 31, 2006

Tony Lownds wrote: > > First, it only handles functions/methods. Python FIT needs > > metadata on properties and assignable/readable attributes > > of all kinds. So in no sense is it a replacement. Parenthetically, > > neither is the decorator facility, and for exactly the same reason. > > > > I can't argue against docstrings and maybe annotations on attributes, > I'd like them myself. That should be a separate PEP because the scope > of this one is Function Annotations. > > The syntax for function annotations has been much more thoroughly > discussed than for annotations on attributes. See Guido's blog and > other references in the PEP. Syntax is always an issue. Looking at the history of annotations shows that people seem to prefer using a comment mechanism. This goes along with the notion of supplying minimal mechanism until you see what the actual usage is going to be. As far as I'm concerned, an annotation mechanism has to have several properties: 1. The annotation data needs to be available at run time without having the source available. 2. It needs to be syntax checkable by a mechanism supplied by the author of the annotation schema. I'd suggest a hook at the back end of import, since this doesn't get in the way of the compiler. The converse of this, of course, is that neither the language nor the compiler needs to have any idea of the actual syntax. This provides maximal freedom to experiment. 3. It needs to have a convention that will allow authors of different schemas to stay out of each other's way. Docstrings almost manage this. While they're certainly available at run time (at least if you don't compile in a way that strips them out) you can only have one in any module, class or method. This means you can't always put them where you want them, that is, close to the item that they're annotating. Parenthetically, I'd note that adding docstring capabilities to properties was a definite step forward. John Roth > > Thanks for the feedback from everyone so far, > -Tony

1 0

Progress Box or Bar in Windows
by tubby Dec. 31, 2006

Dec. 31, 2006

Hi guys, I have a Python script that I've prettied-up for Windows users by adding things like shell.SHBrowseForFolder and win32gui.MessageBox, etc. In short, it looks like this: 1. Pretty window where user can browse for folder. (instead of typing path into cmd prompt) 2. win32gui.MessageBox informing user the script will now begin. 3. 4. win32gui.MessageBox informing user the script has finished. I left 3 blank. This is where I'd like to use some sort of progress indicator (a box that says 'processing 3 of 30 files... 4 of 30, etc' or a progress bar that moves across the screen as the script does its job. Any ideas on how to do this quickly and easily (less than 20 lines of code)? I'd prefer to stick with the win32 extensions, but I'd be willing to use TK too if that would be simpler... and it looks like TK would be much simpler :) Thanks!

3 3

Re: PEP 3107 Function Annotations for review and comment
by Kay Schluehr Dec. 31, 2006

Dec. 31, 2006

I have two questions: 1) I don't understand the clause ('*' [tname] (',' tname ['=' test])* in the grammar rule of typedargslist. Does it stem from another PEP? 2) Is the func_annotation information for def foo(*c: list) stored as {"*c": list} preserving optional argument information or {"c":list} ? Regards, Kay

2 1

request for code : Py Tic Tac Toe in action
by Furman Smith Dec. 31, 2006

Dec. 31, 2006

Hi elainejackson7355_at_home.com, I read "I played against your script and got a win, which, as you know, is not possible with optimal play by the opponent. A while ago I wrote a script that plays optimal tic-tac-toe. Let me know if you want to look at it." I would certainly appreciate the code. --furman === === === extra of little or no interest: I'm learning Python in order to analyze a game that I invented (the last article listed at http://www.integers-ejcnt.org/vol2.html ) called naught or cross. Neither player "owns" a symbol -- "x" or "o" -- but wins when achieving three of the same symbol in a tic-tac-toe line. When you place your symbol on the board you tell me which symbol I am to place. I place the assigned symbol where-ever I choose and then tell you which symbol to place. I started a hand analysis yesterday and it took hours to determine part of the game tree. I estimated, assuming other parts of the tree would take as much work, that I had about 750 times as much work left to do. So it is time to program -- which I love and I've been intending to learn Python anyway. I've taught Fortran and Pascal (and BASIC, AWK, Logo, etc) years ago and need to learn the object oriented approach. Perhaps your work will make it make sense. So far I'm thinking of a game position as a tuple where the first component shows the locations occupied by the symbol which is to be placed and the second component shows the locations of the other symbol. I'm using the following scheme for cell locations in order to simplify translating a position to a standard position (rotating by adding mod 8 to the outside cells and flipping by subtracting the outside cells from 8): 0 1 2 7 8 3 6 5 4 This is already more than you probably want to know but let me give an example. If you have just placed the piece that I've named and the board looks like o . . x o . x . . and you tell me to place an "x" (which is position (67,08) with x's specified in the first component, which is equivalent to standard position (12,08)) then I might place the x in the middle right and tell you to place an x on your move. Therefore the board looks like o . . x o x x . . and you are to place an x (this is equivalent to (367,08)). Since you wrote a Python program to play tic-tac-toe optimally, I bet that you will place your x in the lower right and tell me to place an o. So the position is o . . x o x x . x and I am to place an o. (a position equivalent to (08,3467)). You have me. No matter where I place the o and no matter whether I tell you to place an x or an o after that, you will be able to make three in a line. I found a neat Web page which was cached by Google, http://209.85.165.104/search?q=cache:jyTWvS7w5asJ:en.literateprograms.org/Ti c_Tac_Toe_(Python)+Py+Tic+Tac+Toe&hl=en&gl=us&ct=clnk&cd=10 , and I'm enjoying reading it but I'd appreciate seeing your approach also.

1 0

Re: Question concerning this list [WebCrawler]
by Marc 'BlackJack' Rintsch Dec. 31, 2006

Dec. 31, 2006

In <mailman.2169.1167563637.32031.python-list(a)python.org>, Thomas Ploch wrote: > This is how my regexes look like: > > import re > > class Tags: > def __init__(self, sourceText): > self.source = sourceText > self.curPos = 0 > self.namePattern = "[A-Za-z_][A-Za-z0-9_.:-]*" > self.tagPattern = re.compile("<(?P<name>%s)(?P<attr>[^>]*)>" > % self.namePattern) > self.attrPattern = re.compile( > r"\s+(?P<attrName>%s)\s*=\s*(?P<value>\"[^\"]*\"|'[^']*')" > % self.namePattern) Have you tested this with tags inside comments? >>> You are probably right. For me it boils down to these problems: >>> - Implementing a stack for large queues of documents which is faster >>> than list.pop(index) (Is there a lib for this?) >> >> If you need a queue then use one: take a look at `collections.deque` or >> the `Queue` module in the standard library. > > Which of the two would you recommend for handling large queues with fast > response times? `Queue.Queue` builds on `collections.deque` and is thread safe. Speedwise I don't think this makes a difference as the most time is spend with IO and parsing. So if you make your spider multi-threaded to gain some speed go with `Queue.Queue`. Ciao, Marc 'BlackJack' Rintsch

2 1

Are all classes new-style classes in 2.4+?
by Isaac Rodriguez Dec. 31, 2006

Dec. 31, 2006

Hi, This is probably a very basic question, but I've been playing with new style classes, and I cannot see any difference in behavior when a declare a class as: class NewStyleClass(object): or class NewStyleClass: I declare property members in both and it seems to work the exact same way. I am using Python 2.4, and I was wondering if by default, all classes are assumed to be derived from "object". If not, can someone point me to some place where I can learn more about new-style classes and their advantages? All the documentation I've found is very vague. Thanks, - Isaac.

5 4

py2exe 0.6.6 released
by Jimmy Retzlaff Dec. 31, 2006

Dec. 31, 2006

py2exe 0.6.6 released ===================== py2exe is a Python distutils extension which converts Python scripts into executable Windows programs, able to run without requiring a Python installation. Console and Windows (GUI) applications, Windows NT services, exe and dll COM servers are supported. Changes in 0.6.6: * Better support for Python 2.5. * Experimental support for 64-bit builds of Python on win64. * Better ISAPI support. * New samples for ISAPI and COM servers. * Support for new "command-line styles" when building Windows services. Changes in 0.6.5: * Fixed modulefinder / mf related bugs introduced in 0.6.4. This will be most evident when working with things like win32com.shell and xml.xpath. * Files no longer keep read-only attributes when they are copied as this was causing problems with the copying of some MS DLLs. Changes in 0.6.4: * New skip-archive option which copies the Python bytecode files directly into the dist directory and subdirectories - no archive is used. * An experimental new custom-boot-script option which allows a boot script to be specified (e.g., --custom-boot-script=cbs.py) which can do things like installing a customized stdout blackhole. See py2exe's boot_common.py for examples of what can be done. The custom boot script is executed during startup of the executable immediately after boot_common.py is executed. * Thomas Heller's performance improvements for finding needed modules. * Mark Hammond's fix for thread-state errors when a py2exe created executable tries to use a py2exe created COM DLL. Changes in 0.6.3: * First release assembled by py2exe's new maintainer, Jimmy Retzlaff. Code changes in this release are from Thomas Heller and Gordon Scott. * The dll-excludes option is now available on the command line. It was only possible to specify that in the options argument to the setup function before. The dll-excludes option can now be used to filter out dlls like msvcr71.dll or even w9xpopen.exe. * Fix from Gordon Scott: py2exe crashed copying extension modules in packages. Changes in 0.6.2: * Several important bugfixes: - bundled extensions in packages did not work correctly, this made the wxPython single-file sample fail with newer wxPython versions. - occasionally dlls/pyds were loaded twice, with very strange effects. - the source distribution was not complete. - it is now possible to build a debug version of py2exe. Changes in 0.6.1: * py2exe can now bundle binary extensions and dlls into the library-archive or the executable itself. This allows to finally build real single-file executables. The bundled dlls and pyds are loaded at runtime by some special code that emulates the Windows LoadLibrary function - they are never unpacked to the file system. This part of the code is distributed under the MPL 1.1, so this license is now pulled in by py2exe. * By default py2exe now includes the codecs module and the encodings package. * Several other fixes. Homepage: <http://www.py2exe.org> Download from the usual location: <http://sourceforge.net/project/showfiles.php?group_id=15583> Enjoy, Jimmy

1 0

Re: python , Boost and straight (but complex) C code
by Osiris Dec. 31, 2006

Dec. 31, 2006

On Sat, 30 Dec 2006 23:35:22 +0200, "Roman Yakovenko" <roman.yakovenko(a)gmail.com> wrote: >On 12/30/06, Osiris <nono(a)hotmail.com> wrote: >> Visual C++ build log at: >> >> http://213.10.133.192/BuildLog.htm > >It is better to ask Boost.Python related questions on it mailing list: >http://mail.python.org/mailman/listinfo/c++-sig/ > >You should add to the link line boost_python.lib, thus you will eliminate >"unresolved reference symbol" errors. If you did not build Boost.Python >next page( http://boost.org/libs/python/doc/building.html ) contains pretty >good explanation how to do this. That explanation I find confusing: "To build boost_python, use Boost.Build...": Where is boost.build and what is it ? in my D:\boost\libs\python\build there is no boost.build... in D:\boost there is a build-boost.jam file.... Is that it ? I have a D:\boost\libs\python\build\bin-stage\boost_python.dll (204kB size) and a D:\boost\libs\python\build\bin-stage\boost_build.lib (163 KB) I made them with Visual Studio... I made a bjam, it is in C:\ In D:\boost\libs\python\build there are a bin_stage and a VisualStudio folder , the files jamfile jamfile.v2 python_v1.zip Is it so, that I can build the boost_python.dll and lib, either with bjam OR with VisualStudio C++ ? Why are all those other folders there in d:\boost ? people, regression, tools, status etc ? What do I do with d:\boost\boost ? a lot of headers there... what are all these folders in d:\boost\libs\ ? In short: it's all rather confusing.... I think it must be like this: To use my C/C++ code with Python, add some stuff in the C/C++ source and compile it into a DLL, that must be combined with some boost-DLL to make it accessible to Python 2.4. Therefore I need a boost DLL from boost.org, and some headerfiles that belong to the boost DLL. (where are those boost header files ) As you see, a lot of obfuscation.... MAybe I see too many bears on the road... > >http://boost.org/libs/python/doc/v2/scope.html - here you will find example how >to expose my_int variable to Python. my_int has type int, so C++ code will not >see changes to the variable that are done from Python. You will have to write >set_my_int( x ) function to achieve this. > >If you are new to Boost.Python try to use Py++ - the code generator for >the library. The Py++ GUI( http://tinyurl.com/ycwvwo ) will help you >to start\learn Boost.Python.

2 1

Re: Question concerning this list
by Marc 'BlackJack' Rintsch Dec. 31, 2006

Dec. 31, 2006

In <mailman.2166.1167535289.32031.python-list(a)python.org>, Thomas Ploch wrote: > Alright, my prof said '... to process documents written in structural > markup languages using regular expressions is a no-no.' (Because of > nested Elements? Can't remember) So I think he wants us to use regexes > to learn them. He is pointing to HTMLParser though. Problem is that much of the HTML in the wild is written in a structured markup language but it's in many cases broken. If you just search some words or patterns that appear somewhere in the documents then regular expressions are good enough. If you want to actually *parse* HTML "from the wild" better use the BeautifulSoup_ parser. .. _BeautifulSoup: http://www.crummy.com/software/BeautifulSoup/ > You are probably right. For me it boils down to these problems: > - Implementing a stack for large queues of documents which is faster > than list.pop(index) (Is there a lib for this?) If you need a queue then use one: take a look at `collections.deque` or the `Queue` module in the standard library. Ciao, Marc 'BlackJack' Rintsch

2 1

python , Boost and straight (but complex) C code
by Osiris Dec. 31, 2006

Dec. 31, 2006

I have these pieces of C-code (NOT C++ !!) I want to call from Python. I found Boost. I have MS Visual Studio 2005 with C++. is this the idea: I write the following C source file: ============================ #include <iostream> #include <stdafx.h> namespace { // Avoid cluttering the global namespace. int my_int; /* a global integer: or outside namespace ? */ double calc ( double f) { my_int = (int) (f/2); printf( "Half of %f is %d\n", f, my_int ); return f/2; } } #include <boost/python.hpp> using namespace boost::python; BOOST_PYTHON_MODULE( half ) { def("calc", calc ); } ================================ Which I put in a VC project and compile into a .DLL This DLL I put somewhere on my disk, where Python 2.4 can find it. then I write the following Python source: ===================== from half import * calc(34655.0) et voila ? Can I acces my_int too, this way ?

5 9