The problem with reading synchronously from /dev/stdin is that you can get a
spurious EOF when the input buffer is empty, even if more content is coming. Now
STDIN is read from a loop, and only stops polling when all input has been read.
This fixes#70#85 and other errors related to parsing large files on STDIN.
For now the implication is that UglifyJS will not leak a function
expression's name in the surrounding scope (IE < 9 does that).
(ref. mishoo/UglifyJS#485)
