httpd-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Rob Szarka <>
Subject [users@httpd] PDF plug-in issues under 2.0.48
Date Mon, 27 Sep 2004 13:49:44 GMT

I am running the stock SUSE build of the 2.0.48 version of httpd2-prefork 
with all the latest updates applied. On migrating sites from an old 1.3.22 
install on another server, I saw problems with PDF files on clients running 
MSIE. I have a workaround in place, but not really a solution.


The MSIE client tries to load the PDF (presumably using the Adobe plug-in), 
but fails with the message "File does not begin with '%PDF'". The log file 
shows two consecutive hits, the first with a 200 error code and the second 
with a 206. There are no problems loading the file in Mozilla.

Things That Don't Seem To Work:

I found several similar problem reports on this list, in bugzilla, and 
elsewhere, some with proposed solutions.

Some were due to using mod_deflate on the content, but I am not. I tested 
with the module not loaded at all to be sure.

Others suggested keeping IE from using chunked encoding through a directive 
sucj as "SetEnvIf Request_URI "\.pdf$" nokeepalive" or "<FilesMatch "\.pdf$" >
    BrowserMatch MSIE nokeepalive downgrade-1.0 force-response-1.0
</FilesMatch>". This seemed to be on the right track (see below), but, 
assuming I applied those directives correctly, it didn't work.

Rob Baily reported a problem with large PDF files in 2.0.49, but he wrote 
that the problem doesn't occur in 2.0.48. The files I'm testing are <1MB; 
not sure what "large" means in this context. Bug #25960 suggests that 
"large" is >90kb, but I haven't tested with <90kb files yet.


On both the old and new servers my mime type for .pdf files was 
"application/pdf". Changing this to "application/x-pdf" (also tried 
"application/pdf pdf", which makes no difference in either MSIE or Mozilla) 
results in MSIE downloading the entire file and loading it in Acrobat 
(apparently as a helper rather than a plug-in) instead of trying to request 
and display it a piece at a time.

Unfortunately, this also results in Mozilla (understandably) reporting an 
unknown MIME type and prompting the user for action (the default action 
being to load into Acrobat).


Has anyone experienced and solved this problem? Should I submit a bug report?

Until this problem is solved, is there a way to get Apache to send 
different MIME types to different browsers as a better workaround?
Surf the Sea of Noise!

The official User-To-User support forum of the Apache HTTP Server Project.
See <URL:> for more info.
To unsubscribe, e-mail:
   "   from the digest:
For additional commands, e-mail:

View raw message