cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ovidiu Predescu <>
Subject Re: [C1][Patch] Resolves 2 caching issues
Date Wed, 17 Jan 2001 19:43:53 GMT
These changes look good to me. Robin, would you please apply this patch before
the 1.8.1 goes out?


On Tue, 16 Jan 2001 23:19:34 -0600, "Wayne Bayever" <> wrote:

> I finally finished testing the patch on Tomcat, Resin and JServ!
> Attached are fixes to and which solve the =
> following issues:
> 1.  Caching doesn't work properly when XML file is included in JSP or =
> servlet.  Before the fix, Cocoon was caching based on the requested URI =
> and query string, not the included URI and query string.  So, if you =
> passed dynamically generated parameters into the XML file from =
> JSP/servlet, Cocoon didn't realize that the request had changed and the =
> cached version was sent.  Also, if 2 different XML files were included =
> in the same JSP/servlet, the first XML file was sent twice.
> With the fix, the included URI and all parameters are used in the cache =
> key.
> (Note, I moved the isValidName() method to Utils.isValidParamName())
> (See more detailed discussion with subject: Caching issues around =
> 1/10/2000)
> 2. While testing the fix for problem 1, I came across another problem.  =
> The files were being cached correctly, but the XSLT processor was =
> reporting that the XSL file had changed, thus invalidating the cache.  =
> This was because the monitor key that was being used was generated by =
> Utils.encode() which includes query string info in the key.  Then, when =
> the query string changed, the monitor couldn't find the key and reported =
> the file as changed.
> With the fix, the monitor key is generated with Utils.encode(req, false, =
> false, false) which only uses (protocol)://(server)/requestedXMLPage as =
> the key.  Thus, a change in query string doesn't matter.
> I generated these diffs with cvs diff -u.  Please let me know if there =
> is something wrong with them and I should regenerate them.
> Wayne.

View raw message