On Jun 7, 1:59 am, "Simon Pickles" <[EMAIL PROTECTED]> wrote: > Hi, > > Can someone help me leave the murky c++ world and enter shiny pythonland? > > I have a problem with importing and global variables, here's my code: [snip!] > When run, I come unstuck here: > > self.clientSocket, self.clientAddress = network.accept() > > I get a nameError on 'network', yet it is one in the global namespace, > defined in server.py before CServerThread.Listen() is called. > > In c++, I understood how to declare variables, here I have a problem. Its > like I need an extern statement from c++. > > Can anyone clarify this for me?
The others have covered your problem. I'd just like to add a little detail. As the others have mentioned, global variables reside in their module namespace. There is no truly "global" namespace, and that hasn't really shown up as a problem in Python. There are two tricks to this: 1) Modules are only executed once, at the first import. Every additional import merely gets the module object, but does not execute the code again. Try adding a print statement outside all class and function statements in a module, then import it multiple times. You'll only see the results of the print statement once. You can also start up python with the "-v" flag. Python will then tell you whenever it executes a module. This means that you can have a "global" module. Create a Python module with all the variables that you want, and just import that module in all places that need those variables. If you find yourself stuffing a huge number of variables into your global module, you probably need to rethink your program design. Otherwise, put the global data in the various modules that create and update it. Usually, your other code will only need to access that data under more rare conditions. If you absolutely need a module to be re-executed, the "reload" function will do what you want. 2) Modules do not have to be imported at the start of the file. A module import statement can occur any time code is executed. By convention, we place the import statement at the beginning of the block where we use it. Remember that Python isn't C++. If you place an import statement in a function, Python doesn't try to import that statement until the function is called. This is very different from C/C++'s #include preprocessor statements. If you're having problems with circular references, the import statement can usually be moved inside of functions or methods that need them, or the files can be refactored to get rid of the circular reference. Here's something that shows both points. I started python by typing in "python -v". There's a large number of imported modules that Python automatically loads. After that finishes: >>> def spam(x): ... x = 5 * x ... import math # Math won't be imported until this function is called ... return math.pow(x, 2) ... >>> spam(1) # This should import the math module, executing its contents import math # builtin 25.0 >>> spam(2) # spam() will now use the previously import math module. (This is >>> fast) 100.0 Hope this helps! --Jason -- http://mail.python.org/mailman/listinfo/python-list
