Subject: [Tutor] Re: Tutor Testing for response from a Web site

[Jeff Shannon]

Anna Ravenscroft wrote:

>On Thursday 10 April 2003 11:48, Jeff Shannon wrote:
>
>>This actually doesn't *quite* work.  You can't have an 'except:' clause
>>and a 'finally:' clause hanging off the same 'try:'.  You need to nest
>>them:
>>
>>try:
>>    try:
>>        print "Begin Downloading..."
>>        metall = urllib.urlopen(metURL).read()
>>        f_object.write(metall)
>>    except IOError:
>>        print "Server connection could not be made."
>>    except:
>>        print "Something has gone wrong."
>>        print sys.exc_info()
>>finally:
>>    f_object.close()
>>    
>>
>
>Just to make sure *I'm* understanding things here... this 'finally' will 
>*always* execute regardless of whether there were any exceptions. Right?
>  
>
>[...]
>So, f_object gets closed whether download was successful or not, as a 
>"clean-up" action after the download has been attempted. Yes?
>
>Just trying to make sure I understand.
>

Yes, exactly.  The 'finally' clause will always execute, no matter how 
the 'try' block is exited (whether by exception, by return, by 
sys.exit(), or by simply reaching the end and continuing).  The only 
thing that will stop a 'finally' clause from executing is a complete 
crash of the Python interpreter, which almost never happens unless 
there's a problem in an extension module.  (Or if the interpreter is 
killed from outside, as might be necessary if it's caught in an infinite 
loop or a deadlock while inside the try block.)

This makes a 'finally' clause an ideal way to make sure that cleanup 
always happens.

Jeff Shannon
Technician/Programmer
Credit International