On Fri, 28 May 2004, Ted Cabeen wrote:
No, it is not possible without changing the code. In many cases, its better to give too many search results than too few.
How substantial of a code change would it be? I might be up for writing a patch if it didn't require widespread changes to the caching and database systems, which is probably over my head.
It requires widespread changes to how string canonicalization is done. It also requires the more substantial work of back-converting all of the on-disk data that is served by cyrus (i.e. regenerating all the cache files during an upgrade, which is more than a little obnoxious, especially for large sites) or dealing with this problem in another way.
Of course, were are fast approaching the day where the search and sort comparitors will require that we do stringprep on all the strings, and that day will be very sad.
When that day comes, will this no longer be a problem? How long is "fast approaching"?
It really depends on when we get development cycles for it really. However, the drafts to which I refer are being worked on in the IETF IMAPEXT working group at the moment.
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Rob Siemborski * Andrew Systems Group * Cyert Hall 207 * 412-268-7456 Research Systems Programmer * /usr/contributed Gatekeeper
--- Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
