Back Home
  Oracle Notes

Let me just start by saying that I pity you. While I realize that Oracle is a huge commercial success, and that it is considered to be a very stable product, I personally could not live with some of its short-comings and idiosyncrasies (bugs).

Having gotten that out of the way, it turns out that the modifications necessary for BugRat to support Oracle appear manageable.

The first problem - empty string being considered NULL values - was eliminated by removing the one dependency that BugRat had for empty strings. Namely, the category definitions. The empty string was used to signify the level of the category. Now, the "dot string" ( "." ) is used in place of empty strings. This makes Oracle happy.

The second, and more difficult, issue was the problem of large strings. Oracle seems to have real problems with string fields over 4000 characters. They will huff and puff all day long about performance, but this is just stupid baggage that they are having trouble eliminating from their now ancient design and prehistoric code base. To solve this problem, we had to funnel all of the JDBC string code through common methods in BugRatJDBC. This is not a bad thing. What this means, however, is that Orable users must define a special JDBC property in the web.xml config file in order for this code to be executed as opposed to the default code. There are more comments regarding this in web.xml and in BugRatJDBC.java.

Otherwise, please let me know if you run into any more Oracle issues.


Back Home