There are 4 Bangla (AKA Bengali) letters who comes with a NUKTA (U+09BC) and
stands as a different meaning or pronunciation. They are Ra (U+09B0), Rra
(U+09DC), Rha (U+09DD) and Yya (U+09DF).
Ra 09B0 came from 09AC + 09BC
Rra 09DC came from 09A1 + 09BC
Rha 09DD came from 09A2 + 09BC
Yya 09DF came from 09AF + 09BC
This is what Unicode consortium says, because they didn't do any research on
Bangla and followed ISCII. Any ways, let's come to the point.
Wikipedia pages are behaving strangely after the correct input. If I write 09DC,
it becomes 09A1 + 09BC automatically after saving the data. Like that if I write
09DD, it becomes 09A2 + 09BC and 09DF becomes 09AF + 09BC. Fortunately 09B0 is
ok from this problem. 09B0 has no problem.
Now you have to sort the issue by reversing the rendering. If I type 09B0, it
should stay as it is and if I type 09AC + 09BC, it should become 09B0
automatically. As I told you that 09B0 has no problem, but you also need to
define 09AC + 09BC = 09B0. Like that 09DC will stay as it is and if anyone puts
09A1 + 09BC, it will become 09DC after saving. 09DD will stay as it is and if
anyone puts 09A2 + 09BC, it will become 09DD after saving and 09DF will stay as
it is and if anyone puts 09AF + 09BC, it will become 09DF after saving.
So please sort the issue ASAP.