🌐 Mailman 3 April 2000 - Python-list - python.org

Good style for multi-valued returns
by claird＠starbase.neosoft.com May 1, 2000

May 1, 2000

Examples of multi-valued returns: (head, tail); (quotient, remainder); (left, right); (prologue, body, epilogue); (data, code); (CAR, CDR); (count, size); (handle, status); (decomposition; remainder); ... Let me guess: folks generally pass tuples for small and determinate-in-number multi-values, and dictionaries other- wise, 'cept some do it all with tuples. Accurate? -- Cameron Laird <claird(a)NeoSoft.com> Business: http://www.Phaseit.net Personal: http://starbase.neosoft.com/~claird/home.html

8 9

win32com: _NewEnum being cached
by Tom Hines May 1, 2000

May 1, 2000

Hello. As you know, to implement "for x in my_list" functionality in COM, you must code a get__NewEnum() function, which I have done and it works fine under Visual Basic. But when called from Python, the call seems to get cached, as subsequent Python "for" blocks don't result in calls to get__NewEnum(). I have a COM class that implements a dynamic collection, so I need "for" blocks to result in calls to get__NewEnum() every time. Does anybody know how to force this? Thanks. -- Tom Hines Rockville, MD Sent via Deja.com http://www.deja.com/ Before you buy.

2 4

Using BaseHTTPServer
by Lars von Wedel May 1, 2000

May 1, 2000

Hello, my CGI project suffers from a lengthy delay for starting up python and loading all required modules. I'm thinking about setting up a small HTTP server in python itself so that the interpreter and all the required modules have to be loaded only once. >From the module documentation it is not clear to me, what I actually have to do to catch the GET calls arriving at the server. I assume that I can use BaseHTTPServer as it is. I just need to implement a working request handler, right? Now comes the more fuzzy part: Since I do not want to execute external files but internal methods, I think I could modify CGIHTTPServer by subclassing it and overriding its do_GET() method to decode the request directly? Or would I rather use SimpleHTTPServer, here? Within any of the do_GET methods, how do I access data that has been sent by the client (...&data1=val1?data2=val2?...)? Thanks, Lars

4 4

ZODB troubles
by Nolan Darilek May 1, 2000

May 1, 2000

For hack value, I've semi-revived some old code which I wrote and have been attempting to rewrite it to work with ZODB. I'm trying to create an environment in which python code is stored alongside objects. So, you'd have a database of objects which could be used, and you'd be able to 'dump' their code to text. But, I've been having some ZODB issues and was hoping that someone could help. I'm writing small test scripts to ensure that my code is behaving the way it should. So far I have three tests: 1. Open the ZODB and create a test object, closing it again. 2. Re-open the ZODB, create a persistent function' and run it to prove that it works. 3. Open the ZODB, access the object, run the persistent function and print the code. Test 1 works flawlessly. Test 2 runs, but in order to save the newly-added function, I have to delete the object from the database and re-insert it; it can't be modified live it seems. Test 3 works depending on which route is taken for 2; the long "delete and re-add the object" way works, but the shorter "add an instance attribute to the object and commit" doesn't. I tried duplicating this problem on ints and other simple types, but I couldn't. It only seems to appear under my circumstances. So, in order to help you help me, I've flattened my hierarchy of scripts into a single Python script which requires the ZODB modules. I have three functions entitled one()/three(). Since two() determines whether this script fails or succeeds, I've added a few instructions on how to switch between the success and failure cases. To run this, 'python test.py 1', or 2-3 to run the other tests. They should be run in order, since they lay groundwork for the next test. Another question; is it all right to embed get_transaction().commit()'s (or commit(1)'s) deep within my code, in places where fundamental changes are made? (See Function.compile() for an example of what I mean. I'm inserting the script below. If anyone could help with this, I'd greatly appreciate it. Thanks. -------------------------------------------------------------------------------- #!/usr/bin/env python import Persistence import re import rexec import sys from ZODB import * from ZODB.FileStorage import * class Database: """Simple wrapper around ZODB. This class provides a simple wrapper around the lower-level ZODB functions. Some functionality is not yet implemented. """ def __init__(self, file='Data.fs'): """ Create a database, connect to it and get the root object out of it. """ self.db = DB(FileStorage(file)) self.connection = self.db.open() self.root = self.connection.root() def addObject(self, object): """Add an object to the database.""" dbref = `len(self.root.keys())` while self.root.has_key(dbref): dbref = `int(dbref)+1` self.root[dbref] = object get_transaction().commit() return dbref def dump(self): """Commit all transactions.""" get_transaction().commit() def __getitem__(self, item): return self.root[item] class Function(Persistence.Persistent, rexec.RExec): """This is a 'persistent function.' A persistent function acts like a regular function, but code exists alongside compiled code. This also neatly handles recompiling functions, since the function objects don't play nice with ZODB/Pickle. """ def __init__(self, code = ""): """Initialize a new instance. If `code' is specified, compile it. """ self.code = code if code != "": self.compile(code) def __call__(self, *args): """Execute the function code. If specified, `args' are passed directly to the function code. """ # setstate() can't seem to find self.code. # Since the actual function code objects don't persist, the uncompiled # code is kept until the function is initially called, then we compile. try: self._v_func except AttributeError: self.compile(self.code) apply(self._v_func, args) def compile(self, code): """Compiles the function. `code' is a list of strings consisting of this function's Python code. """ self.code = code exec(code) counter = 0 match = re.match("def\s+(.*?)\s*$.*$:", self.code) if not match: return fname = match.string[match.regs[1][0]:match.regs[1][1]] exec("self._v_func = %s" % fname) # Will this commit the modified function? # Is it a good idea to nestle commits within places like this where # changes are made? get_transaction().commit() def __repr__(self): return self.code class Object(Persistence.Persistent): """A persistent object with some changes. This acts like a normal object with a notable exception: all attributes are stored in 'Attribute' objects, with setattr/getattr handling access. This will allow certain attributes to be public, private, etc. """ def __init__(self): self.__dict__["__attributes__"] = {} def __setattr__(self, name, value): self.__attributes__[name] = value get_transaction().commit() def __getattr__(self, name): return self.__attributes__[name] def one(): print "Test 1: Initializing a new database and adding test object 0." db = Database() t = Object() db.addObject(t) print "Done!" def two(): print "Test 2: Retrieving test object 0 and adding a function." db = Database() # t = db.root['0'] # t.hw = Function("def hw():\n print 'Hello, world.'\n") # del db.root['0'] # db.addObject(t) # Uncomment the previous 4 lines and comment out the next to fix this. db.root['0'].hw = Function("def hw():\n print 'Hello, world.'\n") # As you can see, the previous line is much more aesthetically pleasing. # How can I get it to do what the above 4 lines do? db.root['0'].hw() db.dump() print "Done!" def three(): print "Test 3: Retrieving test object and function." db = Database() db.root['0'].hw() print db.root['0'].hw print "Done!" if sys.argv[1] == '1': one() if sys.argv[1] == '2': two() if sys.argv[1] == '3': three() --------------------------------------------------------------------------------

2 1

Linux Processes from Python HOW-TO?
by chris_barker＠my-deja.com May 1, 2000

May 1, 2000

Hi all, Has anyone written a HOW-TO or tutorial that explains how to start and manipulate other processes with Python. I imagine it's trivial if you are familiar with the Linux/Unix process model, but I get very confused about when to use: system popen and it's variants exec and it's variants fork etc. I'm really hoping someopne has written this up! If you just have a subset of the info, send it to me, and maybe I'll write it up! -Chris Barker cbarker(a)jps.net Sent via Deja.com http://www.deja.com/ Before you buy.

6 6

Python & clipboard
by Hans Nowak May 1, 2000

May 1, 2000

Howdy y'all, I am trying to find a function which gets clipboard data (on Win98). I have been looking at the win32clipboard module, but it seems that the best I can get is a handle to an object containing clipboard data. I also looked into Tkinter, and I found clipboard_append and clipboard_clear methods, but no such thing as clipboard_get. Does anyone know how to do this? Blessed be, --Hans Nowak (zephyrfalcon(a)hvision.nl) Homepage: http://www.hvision.nl/~ivnowa/newsite/ Python questions? See http://tor.dhs.org/~zephyrfalcon/snippets You call me a masterless man. You are wrong. I am my own master.

2 1

What does "first class" mean?
by Tom Loredo May 1, 2000

May 1, 2000

Hi folks- I've often seen reference here (and on the Ruby page) to Python's treatment of functions and methods as "first class objects." What does this mean? I thought I had most of the OO lingo in hand after a year of EE/CS courses at MIT in the early 80s (back then most it fell under the rubric of "data abstraction"). But "first class" in this context must have been in one of the lectures I missed. (There were quite a few of those, actually!) Thanks, Tom Loredo

5 5

How to make compile work for strings that contain ^M's?
by Alex April 30, 2000

April 30, 2000

Hi. I just noticed that the command compile(file, 'test', 'exec') can fail if 'file' contains '^M^J' at the end of each line. Specifically, if file is WDhandler.py, in the webdebug package (http://www.cyberclip.com/webdebug/webdebug-1.15.tar.gz), then I get the error compile(file, 'test', 'exec') Traceback (innermost last): File "<stdin>", line 1, in ? File "<string>", line 29 """ ^ SyntaxError: invalid syntax ... and sys.exc_info()[1].offset == 1043. Is there any way that I can get compile to work properly in this case? Neither python nor emacs seem to have a problem with '^M^J' line-endings otherwise, but I use the compile(file, 'test', 'exec') command in a script called by emacs to systematically test my scripts for syntax errors before I save them. It just occurred to me to strip out the ^M's in the script that does the syntax checking. I guess that'll do, but I'd still be interested in whether there's a nicer way. Alex.

1 0

FSBO Affiliate Program
by mtanzer＠bigfoot.com April 30, 2000

April 30, 2000

Would you like to increase the income and hits on your website...at no cost to you? You are invited to join our FSBO affiliate program. You were personally selected because, looking at your website, we see that you are active in working with FSBOs. Please take a minute and visit http://www.home-buyer-seller-tips.com/affiliates.htm and discover how this program works and how it can bring you a sizable profit each month...and it's FREE! ******************************************************* If you are not involved in FSBOs and wish to be removed from our mailing list, please type "remove" in the subject box of this message and return it to us. ******************************************************* Milt Tanzer, Email: mtanzer(a)bigfoot.com Real estate broker and counselor since 1973

1 0

Environmental variable $PYTHONPATH
by cestluk April 30, 2000

April 30, 2000

Please excuse me if this rather dumb question has been answered before. It does not seem to be in the FAQ, though. It seems that each time you add a new library/module to your current python installation, you have to make it known to python manually, for example, by editing $PYTHONPATH or including it in the search path by invoking sys.path(). I am using the W32 version of python 1.6 and the sys.path output looks something like '', 'C\\Python16\\win32', 'C\\Python16\\win32\\lib', 'C\\Python16', 'C\\Python16\\Pythonwin', 'C\\PYTHON16\\DLLs', 'C\\PYTHON16\\lib', 'C\\PYTHON16\\lib\\plat-win', 'C\\PYTHON16\\lib\\lib-tk', 'C\\PYTHON16\\PYTHONWIN', 'C\\PYTHON16'] It seems you have to specify the name of _every_ directory and subdirectory. Doing this manually would be quite tedious if you have a lot of modules to install. Is there a simpler/less manual way to do it? I am thinking of the case of TeX (in tetex) where the system installation is under a certain texmf tree and the user's own packages are under another and running kpathsea will set up the search path for the packages. Regards, ST --

2 1