Alex Shinn <xxxxxx@gmail.com> writes:

> On 8/19/05, Matthias Radestock <xxxxxx@sorted.org> wrote:

> I was using SISC's binary I/O procedures assuming that would make
> read-char etc. behave on octet values (Latin-1).  This was necessary
> to pass the integer tests, which failed before.  Is this the wrong approach?
> I don't see how the integer tests could have passed if read-char was
> assuming any other encoding.

Alex,

SISC will use whatever encoding Java thinks is the default encoding for
your platform, unless you override that. If your default encoding is
Latin-1 then read/write-char may indeed do what you expect, but for most
other encodings it will probably break.

Matthias.