------- Additional Comments From [hidden email] 2005-05-24 22:47 -------
Daniel, you are correct that the cost of the exception catching is only incurred
if the exception is actually thrown. I was a bit stunned to see this crop up in
performance analysis. See Bug 34930 where some performance profiling highlights
this area as one of the higher CPU consumers, so it _must_ be getting called.
now over a large Document set being indexed, if it is invoked quite a bit, it
ends up causing a reasonable impact on the overall timeline for indexing.
Isn't the proposed patch a better way to do the same thing even if it's NOT
called that often?