Re: AW: [Httpunit-develop] Problem with TC4.0 and redirection

[Glenn A. M. <gl...@so...>]

On Mon, 25 Mar 2002, Joi Ellis wrote:

> On Mon, 25 Mar 2002, Glenn A. McAllister wrote:
> 
> [snip]
> > A point in favour of this solution is that a runtime like HotSpot should
> > detect the "dead" areas after the first pass of the code and simply not
> > bother with them again.
> 
> Except that, HotSpot's JIT can interfere with running the code under
> a debugger, so one has to turn it off to make the debugger function.
> So the JIT's ability to skip over the dead areas is of no help.  Of course,
> if you're single-stepping through a section, you won't notice any
> degradation because you're waiting for the debugger gui.  You would 
> notice when running to breakpoints, method returns, etc.

Its not during interactive debugging that this would be an issue, its
during normal operation.  If I've got a debugger stepping through the
code, I've already exhausted the options logging will have provided me.

Admittedly, part of this discussion is moot.  Performance is not that big 
a deal in a tool like HttpUnit.  None the less, there is no reason to 
sacrifice performance if you don't have to.

Glenn McAllister
SOMA Networks, Inc.