Honestly, I'm a little torn about the interaction between restricting the range of properties and an explicit "set-property" call. It feels like the task of trying to restrict the range of valid properties shouldn't clobber a previous pinpointing of exactly what property you want, because they seem like different operations. I think I'm slightly in favor of Bob's original proposal, with one modification:
- A redefinition of a deferred binding property does not invalidate a previous "set-property" declaration. If the previous "set-property" declaration is no longer a valid value in the redefined property, it is a static error. On Wed, Sep 10, 2008 at 10:25 AM, Bruce Johnson <[EMAIL PROTECTED]> wrote: > Sorry, typo. See correction below. > > On Wed, Sep 10, 2008 at 10:24 AM, Bruce Johnson <[EMAIL PROTECTED]> wrote: > >> I propose instead of allow <set-property> to have a "values" attribute as >> a mutually-exclusive alternative to the "value" attribute: >> > > I propose instead to allow <set-property> to have a "values" attribute... > (ditto above) > > --~--~---------~--~----~------------~-------~--~----~ http://groups.google.com/group/Google-Web-Toolkit-Contributors -~----------~----~----~----~------~----~------~--~---
