Zope large PDF question

Kyler Laird Kyler at news.Lairds.org
Mon Apr 7 18:23:12 CEST 2003


davec999 at yahoo.co.uk (David Currie) writes:

>I am having some problems serving large(ish) PDF files from Zope. 
>Basically, Acrobat Reader seems to hang trying to open a PDF directly
>from a Zope server when used as an IE plugin.  Downloading the file
>with right-click/save-as works fine, opening the file in the plugin
>when it is stored on an Apache server on the same machine and opening
>the file from the same place also works without any problems.  So it
>looks like the Reader plugin does not play well with Zope.  Does
>anyone have any ideas on why this could be happening?

Yes, it's almost certainly due to byterange reading.

>Does Zope have
>problems transferring large files under some circumstances?

Zope has problems dealing with byteranges the way that
Acrobat Reader specifies them for "large" files.  I 
reported this a long time ago, but haven't followed it
lately.  (I just use my own patch to disable it.)  

As I recall, Reader sends a request for pieces of a
large file
	1-500, 1001-1500, 501-1000
and Zope decides to be efficient and join all of the
contiguous regions.  Unfortunately, that defeats the
purpose of requesting the ranges *and* breaks Reader.

--kyler




More information about the Python-list mailing list