ZipInputStream brokenness

[Jeff Sturm]

On Mon, 2 Sep 2002, Per Bothner wrote:
> > Perhaps in bulk data copies?  I'd guess a raw InputStream should
> > perform slightly better than any buffered stream for large enough reads.
>
> That's application-levell buffering, I think.

True.  OK, I had misunderstood your point.  I claim that there are
applications wanting 100% control over buffering, but that isn't what
you were arguing.

Having access to raw I/O is demonstrably a good thing, but I'd agree
that the byte-at-a-time FileInputStream.read() method is quite useless.

Jeff