mod_mime/4097: mod_mime_magic unable to handle compressed content larger than 4k

[Marcin Cieslak <sa...@system.pl>]

>Number:         4097
>Category:       mod_mime
>Synopsis:       mod_mime_magic unable to handle compressed content larger than 4k
>Confidential:   no
>Severity:       non-critical
>Priority:       medium
>Responsible:    apache
>State:          open
>Class:          sw-bug
>Submitter-Id:   apache
>Arrival-Date:   Mon Mar 22 03:10:00 PST 1999
>Last-Modified:
>Originator:     saper@system.pl
>Organization:
apache
>Release:        1.3.4
>Environment:
FreeBSD 2.2.8-STABLE
>Description:
Due to the limit imposed by HOWMANY symbol, the zmagic() function in
mod_mime_magic.c attempts to uncompress first 4k of the file only,
and mod_mime_magic module is unable to detect the type of uncompressed
contents.

The following entry appears in the error_log:

httpd: [Sun Mar 21 18:17:04 1999] [error] [client 127.0.0.1] mod_mime_magic:
read failed /home/WWW/Home/cdrom/cc/td/doc/product/software/ssr91/rn_9_14/78116.htm

This error is generated in mod_mime_magic.c:2198, since ap_bread()
was unable to read any data from gzip process pipe.
>How-To-Repeat:
Discovered when browsing Cisco Documentation CD (it contains compressed
HTML files) -- some file (smaller) were properly displayed by
Netscape and lynx, while others were not. I guess that the module's
primary objective is to allow browsing of Cisco CD, since it was
submitted to you by Cisco:)

Sample files can be downloaded from http://www.system.pl/internal/mod_mime_magic/
(Note this server has mod_mime_magic disabled).

The file small_gzipped.htm should work ok with a standard module.

>Fix:
In http://www.system.pl/internal/mod_mime_magic/patch I enclose a patch
which allows to read large compressed files and fixes another bug
(files without ending ".Z" cannot be verified by module due to the way
"uncompress" utility works).

I have modified the options given to gzip utility - now they include
a file name. 
I have tried to fix it loading the whole compressed file into memory,
but this failed on large files (>100k) - like max_gzipped.htm on my site.
There may be another bug in the way apache handles pipes to- and from- 
process.
E-mail me for further details.
>Audit-Trail:
>Unformatted:
[In order for any reply to be added to the PR database, ]
[you need to include <ap...@Apache.Org> in the Cc line ]
[and leave the subject line UNCHANGED.  This is not done]
[automatically because of the potential for mail loops. ]
[If you do not include this Cc, your reply may be ig-   ]
[nored unless you are responding to an explicit request ]
[from a developer.                                      ]
[Reply only with text; DO NOT SEND ATTACHMENTS!         ]