The answer was given, but to make things more clear to anyone reading this:
1) if you wish to perform ONE transformation in your bitmap, you CAN use SET (setRotate, setScale etc).
But note that any call to a "set" method OVERRIDES other transformations. It's like a new matrix. That's why OP's rotation was not working. These calls are not performed line by line. It's like they are scheduled to be done at runtime by the GPU when the new bitmap is being drawn. It's like when resolving your matrix, GPU rotated it, but then, created a scaled new one, ignoring previous matrix.
2) if you wish to perform more then one transformation, then you MUST use "pre" or "post" methods.
And what is the difference between a postRotate and a preRotate, for example? Well, this matrix math stuff is not my strength, but what I know is that the graphic cards make these transformations using matrix multiplication. It seems to be way more efficient. And as far as I remember from school, when multiplicating matrices the order IS important. A X B != B X A. So, scale a matrix and then rotate it is different from rotate and then scale it.
BUUUUT, as far as the final result in the screen is the same, we high level programmers usually do not need to know these differences. The GPU does.
Well, in that rare cases when you are performing really complicated matrix operations, and results are not what you expected or the performance is terrible and you need to deeply understand these methods to fix your code, well, then android documentation can not be of much help anyway. Instead, a good Linear Algebra book would be your best friend. ;)
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…