On 09.07.2008 10:45:35 Max Berger wrote:
> Jeremias,
> Jeremias Maerki schrieb:
> > Am I the only one concerned about backwards-compatibility here?
> No. I am also concerned about backwards-compatibility, but in a
> different way:
> This change changed the semantics without changing the API, therefore
> code still compiled, but crashed (such as the bug I encountered). This
> is a type of api change I am not happy with.
> What would be ok with me is if the interface had changed (in this case,
> the signature of the functions). My code would no longer compile, and
> I'd have to prepare a new plugin for the new version (which i currently
> have to do anyways).
> My favorite solution would be: Provide the new semantics with a new
> signature (or method name), and keep the old one as "deprecated" for at
> least 1 release (Then all plugin developers have enough time to adjust),
> then remove it.

+1 to that approach. I volunteer to do the necessary changes if we reach
a consensus.

> http://java.sun.com/j2se/1.5.0/docs/guide/javadoc/deprecation/deprecation.html
> Max

Jeremias Maerki

Reply via email to