On Wed, 31 Jul 2002, Jose Alberto Fernandez <[EMAIL PROTECTED]>
wrote:

> I still do not understand why side issue of -projecthelp has to be
> the guiding principle for how the entire system will behave.

-projecthelp is the same as asking Ant to run a target that doesn't
exist here (at least to me).  In both cases I don't expect any
persistent side effects to happen, which could probably only be done
by not executing any tasks at all.

>> My list would be all property setting tasks as well as taskdef and
>> typedef.
> 
> Any task that can set a property should be in this list.

That's what I tried to say. 8-)

But nowadays, almost every task can set a property (<exec> can, for
example).  So maybe it shouldn't be as broad.

> I will try to present a different argument on how to view this
> problem: If you take a look at *MAKE* one option they have and we do
> not is (I think) -n which allows to run the makefile without
> executing the commands.

Yes, I know that - and users have asked for it of course.  But things
are a bit different in Ant (only talking about make's -n) as the
logging you get from tasks like <javac> if you don't perform a
directory scan is rather useless.

But I agree that the whole issue of which task should run under what
condition is closely related to -dryrun and may get resolved at the
same time.

Stefan

--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>

Reply via email to