ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Bruce Atherton <>
Subject Re: <available> / <condition> breaking immutability
Date Tue, 27 Nov 2001 18:55:47 GMT
At 08:05 PM 11/27/2001 +1100, Peter Donald wrote:

>If I was the user and I had fubared my build setup then I would want to know
>as early as possible.

Sure, but how is your setup fubared here? What is the downside for the user 
if they don't follow exactly the developers' vision of what a property 
should be?

I would agree with you if the property-changing behaviour of <available> 
ran counter to most peoples' expectations of how it would work, but I doubt 
anyone here would make that claim. Would they? It seems obvious to me that 
having <available> fail to set the property would create much more 
confusion among users than the current behaviour.

Again, this is just my preference. I can see the arguments in favour of 
making the change.

>Interesting idea ... needs exploration - I suspect it will be too complex.
>However the real problem is that ant1 is not fully capable of dealing
>with mutable properties and such a technique would need to wait till ant2.

I'm not suggesting any change to code here. You already have a system that 
matches up to what I described, although it doesn't enforce the policies 
through code. I'm just talking about changing the design perspective on 
what a property is by expanding to two types. I'm trying to make it clear 
why it is still good design for your current system to have some properties 
that can be changed by certain select tasks.

To unsubscribe, e-mail:   <>
For additional commands, e-mail: <>

View raw message