On Fri, 28 Jun 2002, Ian MacLean <[EMAIL PROTECTED]> wrote: > Stefan, do you recall the reasons for enforcing read-only on > properties in Ant ?
Do you want the full story? 8-) The short version: it happend to be that way in Ant 1.1 by accident and we sticked to it because the majority of people found good reasons why it is better that way. We added features to Ant that made most use cases for mutable properties go away. I'm not really convinced that property immutability (as we call it in Ant land) really is the best way to do it myself - just see what kind of hack I had to introduce in 1.5 to make it work correctly with <ant>, inheritall and nested properties. One of the most cited reasons is that you don't have to hunt down all build files involved just to see what value a property might get at one point - but then again, this means I don't have to look at my down-stream build files but my get hit by build-files calling mine. The reason I can follow best is that you need a policy that is simple to understand. "Properties are always mutable" and "properties are never mutable" are the two policies easiest to explain - and the "always mutable" policy immediately needs exception rules for properties defined on the command line. So we've picked the easiest to explain, comprehend and memorize rule. Stefan ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek No, I will not fix your computer. http://thinkgeek.com/sf _______________________________________________ Nant-developers mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/nant-developers
