I agree that less permutations is better, but I would add a twist to the question: even if we could should we? if one could only predict all issues we will have with IE9 and then would we fix on safariimpl based on (is ua==msie)? fork then and implement implIE9 to fix something? I don't know of any test we can use to confidently assert that ie9 is and will be 100% safari friendly. I would vote for having it now and invest more time on trying to collapse if possible in increments.
On 2011/03/04 19:12:30, rjrjr wrote:
Does the new IE9 value for user.agent imply yet another permutation?
We
should really avoid that if we can, and so far it sounds like it might
not
be needed. Can we introduce IE9 without causing a new hard perm?
On Fri, Mar 4, 2011 at 11:07 AM,
<https://mail.google.com/mail/?view=cm&fs=1&tf=1&[email protected]> wrote:
> Mostly LGTM > > Needs a unit test for property fallback behavior. > > > http://gwt-code-reviews.appspot.com/1369807/ >
http://gwt-code-reviews.appspot.com/1369807/ -- http://groups.google.com/group/Google-Web-Toolkit-Contributors
