Possibly moving to InnoDB

[Myk Melez]

Jeffrey Hulten wrote:
> The other possibility is to make the indexing of the full text be 
> programmatical instead of a function of the database operation.  In 
> other words, index it ourselves.
Right.  One way to do this would be to duplicate both summary and 
comment data in a separate table specifically designed for fulltext 
indexing, so that both bugs and longdescs can get transactions and 
referential integrity, i.e.:

CREATE TABLE bug_descriptions (
  bug_id INT NOT NULL,
  summary VARCHAR(255) NOT NULL,
  comment TEXT
);

Besides freeing those tables to participate in InnoDB goodness (or 
SolidDB goodness, for that matter), the other advantage to this approach 
is that it resolves a longstanding problem with our fulltext indexes 
(because of MySQL's inability to include columns from different tables 
in a fulltext index) that reduces their utility (even with our hacks to 
partially work around the problem).  The cost, however, is disk space.

Otherwise, we could roll our own fulltext engine or use something like 
Senna <http://qwik.jp/senna/>.

-myk

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.bugzilla.org/pipermail/developers/attachments/20060807/503c8138/attachment.html>