Bug 21162 - Corrupted output from SSI files if multiviews used
Summary: Corrupted output from SSI files if multiviews used
Status: RESOLVED DUPLICATE of bug 17629
Alias: None
Product: Apache httpd-2
Classification: Unclassified
Component: mod_deflate (show other bugs)
Version: 2.0.46
Hardware: Other Linux
: P3 normal (vote)
Target Milestone: ---
Assignee: Apache HTTPD Bugs Mailing List
Depends on:
Reported: 2003-06-28 12:10 UTC by Sönke Tesch
Modified: 2004-11-16 19:05 UTC (History)
0 users


Note You need to log in before you can comment on or make changes to this bug.
Description Sönke Tesch 2003-06-28 12:10:04 UTC
Directory with option multiviews enabled, "AddOutputFilter INCLUDES .shtml" and
"AddOutputFilterByType DEFLATE text/html".
Apache 2.0.46 with the mod_deflate patch from
http://nagoya.apache.org/bugzilla/show_bug.cgi?id=19913 applied.

  test.shtml, line 1
  <!--#include virtual="include" -->
  test.shtml, line 3

  include.html, line 1

If test.shtml is requested with "Accept-Encoding: gzip" Apache returns
"Content-Encoding: gzip" and the uncompressed contents of include.html, followed
by the compressed content of test.shtml. Notice that the order is "include.html,
line 1", "test.shtml, line 1", "test.shtml, line 3", so it doesn't even get the
SSI thing correct on test.shtml.

Works fine if virtual="include" is replaced with virtual="include.html", i.e.
without content negotiation in the subrequest for include.html.
Comment 1 André Malo 2003-06-28 14:26:28 UTC

*** This bug has been marked as a duplicate of 17629 ***
Comment 2 Sönke Tesch 2003-06-28 14:48:03 UTC
Sorry, mixed up bug numbers. Patch from
http://nagoya.apache.org/bugzilla/show_bug.cgi?id=17629 applied and does not work.

Please also note that there is no redirect involved as in 17629.
Comment 3 André Malo 2003-06-28 16:16:44 UTC
That is, because the patch is not ready for commit ;-) The Problem is deeper.
The patch provided there solves one part, but breaks other things.

Trust me [ ;-) ], it's the same issue; Multiviews causes a (fast) internal
redirect. That's not a normal redirect, rather than a merge of a subrequest with
the actual request, but suffers in this case from the same conceptual problem.
(it bypasses the resource filter chain and mixes that way the two contents

*** This bug has been marked as a duplicate of 17629 ***