[mod_python] Text appended to the page when refreshing

R. Pelizzi wrote:

>Wojtek Dabrowski wrote:
>
>  
>
>>>I'm using the publisher handler and when i refresh a page that returns
>>>some html code (a full page, enclosed in html tags) the output is
>>>appended to the page (duplicated, if the python script isn't changed).
>>>   
>>>
>>>      
>>>
>>I have had a similiar problem. In my case, the output was generated as a
>>list of lines, kind of like:
>>
>>def handler(req):
>> page=bla()
>> for line in page.x:
>>   req.write(line)
>>
>>class bla:
>> x = list()
>> def __init__(self):
>>   self.x.append("HTML lines")
>>
>>Now x=list didn't generate an empty list, but x sometimes still held the
>>output generated last time the page was called. I corrected this by
>>adding an x=[] in the __init__ to get:
>>
>>class bla:
>> x = list()
>> def __init__(self):
>>   x=[]
>>   self.x.append("HTML lines")
>>
>>This cleared x of any garbage that might have been there, and it works
>>fine now.
>>
>>I don't know why x=list() didn't generate an empty list, I just know it
>>didn't ;) If there's anybody who can actually explain why this happened,
>>I'd be grateful.
>> 
>>
>>    
>>
>Strange: you're using a class variable but every instance deletes it.
>What's the purpose of such a list anyway? Couldn't you just set self.x
>in __init__? Also, since i think mod_python reuses the same interpreter
>(python session) for multiple requests (opposed to cgi), it is
>predictable that class attributes are shared among requests. My class,
>though, has only instance variables. These shouldn't be (and arent')
>shared in normal-python scripting. Now, i wonder how using mod_python
>messes everything up.
>  
>
Ok, forget about the whole thing: the behaviour has been observed in the
standard interpreter too, and i found a way to fix it. Looks like i
still need to wet my feet with python...