[Chicago] help from Django and Pylons developers

[Cosmin Stejerean]

> > Django - the 'recommended' way to deal with files is to not have django
> service
> > the request, but let apache or whatever web server handle those URLs.
> This is
> > good, because python seems to be a poor choice for moving 'large' amounts
> of
> > data 'like this.'
> >
>
>  The problem is that sometime you need to perform authentication of a large
> file (for example a video stream) and you cannot rely on apache. Some
> problem if the file you want to serve is actually generated. For example if
> you scan my machine for .php files you get an infinite file being streamed
> to you. I would not know how to do that with Django. Perhaps it is possible.
>
>

One option for streaming really large files that require
authentication is to create a symlink to the file with a random name
in a folder serviced by Apache (or nginex) and have it serve the file
(I can also think of creative ways to abuse basic auth).

-- 
Cosmin Stejerean
http://blog.offbytwo.com