Re: Improved Javadoc Task

[James Duncan Davidson <ja...@eng.sun.com>]

Stefano Mazzocchi wrote:
> (question: should we have a standard run-time way to indicate task
> deprecation so that one can go ahead and remove the warnings?)

I don't think that we want tasks to get that tightly bound. Since we
still (we do, don't we?) have the <taskdef..> mechanism, we don't want
to necessarily prejudice the namespace.

> 6) made javadoc extend exec, so that it spanws another JVM. Note: this
> is required since javadoc calls system.exit() and breaks ant execution
> if executed in the same JVM. performance penalties turn out to be very
> small.

Cool enough. Javadoc takes enough time anyway that the vm startup isn't
that big a deal.


-- 
James Davidson                                     duncan@eng.sun.com 
Java + XML / Portable Code + Portable Data                 !try; do()