UTF-8 encoding in track attributes

[aviezerl]

Currently, when setting attribute value to a non-ASCII convertible string (e.g., in Hebrew), the next call for emr_track.attr.get/set would fail with an error of "Invalid format of attributes file."

The problem is caused by the use of getc of BufferedFile.h, which doesn't deal with the variable length of UTF-8 encoded strings.

Example:

library(naryn)
load_test_db()
emr_track.attr.set("ph1", "test", "סבתא")
emr_track.attr.export()