On 09 Mar 2011, at 17:44, Vincent Hennebert wrote: > I’d like to launch a vote for the integration of the patch from > Bugzilla #50763 [1] into the Trunk.
I saw the related message a few weeks ago, and am inclined to vote Yea, since I agree with the basics. FOP's current behavior is counter-intuitive, regardless of whether it is compliant. We have done it in the past for other situations, like indent-inheritance and relaxed validation, so there are enough precedents to justify this one. In light of Glenn's remarks, I am going to make it only a +0. (Not that the difference matters much, as the patch is ready anyway.) Especially relevant is the question what other implementations do. If they, too, exhibit the behavior that is now proposed as the default, then it could be a strong argument in favor. Often enough, a spec ultimately evolves to match the behavior of the majority of implementations. If they do what FOP currently does, then while awaiting a definitive answer from W3C, it could be kept non-standard, and enabled through a config-setting (as was done for indent-inheritance, and that turned out to be a correct choice; the intuitive behavior really was non-standard). Perhaps there is a compliant, pure XSL-FO solution to this problem, that we are overlooking (?) Regards, Andreas ---
