I have been here for almost seven years and this has been said a few times. It is understandable though as the majority of people are not using PB for 3D games and have no interest in it, so the majority has to be satisfied first. It is kind of like how it keeps being said the Visual Designer will be fixed and improved, but it never actually happens. Not that many use it, so the priority isn't on it.When did they say this?
You then said:I know only from one case, and since that PB-OGRE is usable.
PB-OGRE is missing some really important features of OGRE.
It seems there is a difference between something being usable and actually being able to make use of it

You have reaffirmed what I said, everybody will use OGRE3D directly and skip PB's implementation. Like you suggest, PB's implementation needs to quit being "updated" (simply recompiled against newer versions of OGRE3D) and actually be upgraded to support more of the features that OGRE3D has added over the many years. If this were to happen, PB could go neck to neck with DBPro and even B3D.