CGI script to count downloads

[Remco Gerlich]

Greg Ewing <greg at cosc.canterbury.ac.nz> wrote in comp.lang.python:
> I'm trying to write a script that will keep a count
> of the number of times a file is downloaded. So far,
> the script just takes the name of the required file,
> opens the file, and passes on its contents.
> 
> It works fine, except for one annoying thing:
> When Netscape pops up the box requesting a file
> name for saving, the default filename seems to be
> constructed in a screwy way from parts of the
> URL. For instance, if the URL used to retrieve the
> file is
> 
>   download.py?file=download/akaroa2.6.1doc.tar.gz
> 
> then the default filename that Netscape gives me
> is
> 
>   akaroa2.6.1doc.tar.py
> 
> So, my question is, can I put anything in the
> headers of the response to give the browser a
> hint as to what name the file should be saved
> under?

<clpm> This is a CGI question! </clpm> ;)

According to the CGI faq at http://www.htmlhelp.com/faq/cgifaq.html ,
chapter 4 question 22, you can use the URL

download.py/filename?options=anyextraoptions

And the filename should be in "PATH_INFO" (env var, I suppose).

Weeding through some other CGI scripts that Google dug up...

Content-type: compressed/gzip; name=%s

might work.

Another CGI script I found uses the header
Content-disposition: inline; filename=%s

And notes that it probably only works on Netscape.

Seen enough Perl for today now...

> Or is there a better way of going about this
> whole counting business?

Maybe analyze logs?

> This is the script I'm using:
> 
>   #!/usr/local/cosc/bin/python
>   print "Content-Type: compressed/gzip"
>   print
>   import cgi, sys
>   form = cgi.FieldStorage()
>   path = form["file"].value
>   f = open(path)
>   sys.stdout.write(f.read())
>   f.close() 

That needs a whole lot of extra checking on the filename of course...

-- 
Remco Gerlich