Any literature I seeked just had written that fact down without justifying it, as if it were so simply to see. And finally, I saw it: The confusion I had was that I misinterpreted as being only, i.e., disappearing at infinity (The closure of is compact). Here, it definitely was not clear that its values already disappear before infinity.
The rule of partial integration says, , where . If the integration should go over the whole space, then , and the same is with .
Now, if with being an interval, it follows that where already disappers, because it has compact (and therefore bounded) support; , where and are the left and right borders of . And this is why in the first question above.
For the sake of completeness, it is clear by the following what should mean for :