> From: Stephen Ng [mailto:[EMAIL PROTECTED]] > > So CacheValidity objects should always take an extra Object parameter, > which contains the string or whatever key is being hashed. The isValid > function would then not just do its usual validity checking, it should > also check the equality of those objects. > > To be explicit, we could rewrite TimeStampCacheValidity to be hash clash > safe as follows: > > private Object realKey; // the other key is "just" a hash which > fits in a long. > > public TimeStampCacheValidity(long timeStamp, Object realKey) { > this.timeStamp = timeStamp; > this.realKey = realKey; > } > > public boolean isValid(CacheValidity validity) { > if (validity instanceof TimeStampCacheValidity) { > return this.timeStamp == > ((TimeStampCacheValidity)validity).getTimeStamp() && > this.realKey.equals(validity.realKey); > } > return false; > } > > Would this (or something like it since I haven't even compiled this) > work?
This has good chances to work; also I'm not (yet?) sure I like this solution. > Or is something better planned for a future (imminent) release of > Cocoon? (I didn't have any luck searching for the bug or TODO that you > mentioned) It is planned to be fixed, also it is not discussed how it will be fixed. Bug related to the caching is: http://nagoya.apache.org/bugzilla/show_bug.cgi?id=4641 And TODO item (related to caching) is: http://cvs.apache.org/viewcvs.cgi/xml-cocoon2/todo.xml?rev=1.28&content- type=text/vnd.viewcvs-markup: <action context="code" assigned-to="open"> Fix lastModificationDate() method misuse. Instead of modificationDate, sometimes generated hash value is being returned, and instead of comaring dates on before/after, they are compared on equality. See http://marc.theaimsgroup.com/?l=xml-cocoon-dev&m=99780708222079& w=2 and http://marc.theaimsgroup.com/?l=xml-cocoon-dev&m=99814216923705& w=2 for more information on this issue. </action> If you have time, you can try to analyze the whole caching picture, including aggregation, sources, and C/X includes. Vadim > Many thanks, > > Steve > > > -----Original Message----- > > From: Vadim Gritsenko [mailto:[EMAIL PROTECTED]] > > Sent: Friday, April 19, 2002 2:05 PM > > To: [EMAIL PROTECTED]; [EMAIL PROTECTED] > > Subject: RE: Cocoon cache hash clash > > > > > > > From: Stephen Ng [mailto:[EMAIL PROTECTED]] > > > > > > I'm confused by coocoon caching--what if I have use HashUtil in my > > > generateKey, and two different pages/request happen to generate the > > same > > > hash key? Won't cocoon erroneously return the same page from the > > cache? > > > > Only if validity matches. But it usually does not - result is > > that these two pages won't be cached at all. Luckily, this > > (almost) does not happen. > > > > There is a bug and a TODO item on this theme. > > > > Vadim > > > > > > > Steve Ng > > > Lumigent --------------------------------------------------------------------- Please check that your question has not already been answered in the FAQ before posting. <http://xml.apache.org/cocoon/faqs.html> To unsubscribe, e-mail: <[EMAIL PROTECTED]> For additional commands, e-mail: <[EMAIL PROTECTED]>
