Re: [Zope-dev] PDF-specific Bug in the ZServer implementation??? Orjust strange behavoiur of IE?

2002-01-07 Thread Derek Simkowiak 

-> > Any hints on that?

The failure to open a PDF can be caused by the browser never 
knowing that file is all done downloading.

If Zope is running on WindowsNT and you're using Solaris (either
as the workstation or else as an HTTP proxy) then you may be getting a
rare bug between the WindowsNT and Solaris TCP/IP stacks.  The Solaris
packet sends an out-of-order FIN (or was it just a FIN with no data?  
Can't remember) and the WindowsNT stack never closes the connection, NOR
does it acknowledge the last packet before the FIN, so you wind up just
stuck.  I.e., you'll see a failed PDF download.  Sun and MS both know
about this, but neither will claim responsibility.

I don't remember the exact details, but we saw this problem a few
months ago.  It was more prevalent with PDFs than anything else because of
large filesize and the fact that I.E. must spawn the PDF "control" (aka
'widget') with the browser.

Just a thought.

--Derek



___
Zope-Dev maillist  -  [EMAIL PROTECTED]
http://lists.zope.org/mailman/listinfo/zope-dev
**  No cross posts or HTML encoding!  **
(Related lists - 
 http://lists.zope.org/mailman/listinfo/zope-announce
 http://lists.zope.org/mailman/listinfo/zope )

RE: [Zope-dev] PDF-specific Bug in the ZServer implementation??? Orjust strange behavoiur of IE?

2002-01-07 Thread sean . upton 

Acrobat uses something called byte-serving, which is a pain to support
sometimes on the server-side, because it essentially amounts to serving
parts of a file up to a given byte offset; when Acrobat acts as an HTTP
client, it requests the PDF file, and once it has enough of the top of the
doc to get header info from the file, it is able to omit downloading the
whole thing and request portions of the file between 2 byte offsets, which
represent the start and end of a page, as part of the http request.
Byte-serving, unfortunately is not implemented in all http servers (or is
broken in some, i.e. Apache 1.3.14).  I'm not sure what the story for
ZServer is, but I am curious to find out if there have been any developments
in this regard...

More info in the IETF RFC for HTTP 1.1:
http://www.w3.org/Protocols/rfc2616/rfc2616-sec3.html#sec3.12

Sean

-Original Message-
From: Joachim Werner [mailto:[EMAIL PROTECTED]]
Sent: Monday, January 07, 2002 12:28 PM
To: [EMAIL PROTECTED]
Subject: [Zope-dev] PDF-specific Bug in the ZServer implementation??? Or
just strange behavoiur of IE?


Hi!

This is my first post in the new year, so best wishes for 2002 from Germany!

And on we go with my problem:

I have a ZClass-based Class derived from File. It holds PDF documents. SOME
(not all) of them do not auto-open in Acrobat Reader as they are supposed to
if MS IE 5.x is used. Opera etc. seem to work. The files themselves are
o.k., and opening in Acrobat Reader AFTER downloading (via right
mouseclick/save as) works fine, too.

If the same document that doesn't auto-open when downloaded fom the Zope
server is downloaded from plain Apache, it works ...

The configuration is Zope 2.4.1 (binary release, python 2.1, linux2-x86),
python 2.1.0, linux2 on a SuSE Linux with Apache as Proxy server (via
proxy_pass). I checked the HTTP header with WGET, and they are identical for
working and non-working PDF docs. The mime type is set to application/pdf
...

Any hints on that?

Joachim


___
Zope-Dev maillist  -  [EMAIL PROTECTED]
http://lists.zope.org/mailman/listinfo/zope-dev
**  No cross posts or HTML encoding!  **
(Related lists - 
 http://lists.zope.org/mailman/listinfo/zope-announce
 http://lists.zope.org/mailman/listinfo/zope )

___
Zope-Dev maillist  -  [EMAIL PROTECTED]
http://lists.zope.org/mailman/listinfo/zope-dev
**  No cross posts or HTML encoding!  **
(Related lists - 
 http://lists.zope.org/mailman/listinfo/zope-announce
 http://lists.zope.org/mailman/listinfo/zope )