I think that found an error in the stdio module.
When reading the file consisting of entries in the coding UTF-8,
type:
.......
16235;Иванов;Петр;Петрович;17.09.1961
8765;Петров;Иван;Васильевич;25.12.1978
.......
total quantity of records about 30000
there are errors: invalid secuence Utf-8
Research of a problem showed that results failure from reading
from the buffer of the next character the FGETC function
(cast(_iobuf*)_f. _ p.handle) in the takeFront procedure. And,
over time failure happens only provided that characters return to
the (cast(_iobuf*)_f. _ p.handle) the ungetc() function. If it is
simple to realize reading characters, without returning them the
ungetc function(), this error isn't shown. For example it was
absent in the dmd 2.066 version where in stdio.d the ungetc
function wasn't used().
