I think the Gump litmus test is a reasonable enough way to judge how much of an impact our changes might have on others when it comes to backwards compatibility. Its not a complete test, but a reasonably decent one to me.
I'm not so hardcore about breaking backwards compatibility that things like this bother me - since I'm always running the latest CVS version of Ant anyway :)
Erik
On Thursday, January 9, 2003, at 04:47 AM, Nicola Ken Barozzi wrote:
Conor MacNeill wrote:Nicola Ken Barozzi wrote:Nicola Ken,
Ok.
I don't see though what feature can be attained from "relying" on the XmlProperty failing for file not existing. Can you please post a use-case to compare with mine?
Backward compatability isn't about what is sensible or what should be. Someone, no matter how silly it might be, may be relying on the current behaviour. They may want their build to fail unless a user has provided a required set of properties - who knows?
Hmmm...
OTOH, we do sometimes break things in minor ways - I'm happy to have it as it is now, if the change is documented as potentially build breaking.
I'll be ok in both ways. I sometimes forget that Ant is a very particular project with regards to changes, because of it driving all other projects with builds.
Let's see what others think about this and do what most prefer.
Thanks :-)
-- Nicola Ken Barozzi [EMAIL PROTECTED] - verba volant, scripta manent - (discussions get forgotten, just code remains) ---------------------------------------------------------------------
--
To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
-- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
