don't sync() on an fstream at EOF

[kartik]

I open an fstream in read-only mode, read till the end, then try to
sync() before seeking to position 0 & reading again. But the sync
fails. I discovered that clear()ing the stream before the sync causes
the sync to work. Seeking to position 0 before the sync doesn't seem
to help.

If you look at sync() as simply invalidating the read buffer (isn't
that right?), isn't it illogical for it to fail in this situation?

Besides sync() doesn't flush the write buffer (on a read-write
stream), does it? Similarly, flush() doesn't invalidate the read
buffer, does it?

I'm using g++ 3.3.2 on Mandrake Linux 10; kernel 2.6.

-kartik

[Jack Klein]

On 29 Oct 2004 21:46:40 -0700, ka*************@yahoo.com (kartik)
wrote in comp.lang.c++:

I open an fstream in read-only mode, read till the end, then try to
sync() before seeking to position 0 & reading again. But the sync
fails. I discovered that clear()ing the stream before the sync causes
the sync to work. Seeking to position 0 before the sync doesn't seem
to help.

If you look at sync() as simply invalidating the read buffer (isn't
that right?), isn't it illogical for it to fail in this situation?

Besides sync() doesn't flush the write buffer (on a read-write
stream), does it? Similarly, flush() doesn't invalidate the read
buffer, does it?

I'm using g++ 3.3.2 on Mandrake Linux 10; kernel 2.6.

-kartik

Attempting to read past the end of file sets the fail bit on the
stream. All operations on the stream, including sync(), will fail
until clear() is used to remove the fail bit.

--
Jack Klein
Home: http://JK-Technology.Com
FAQs for
comp.lang.c http://www.eskimo.com/~scs/C-faq/top.html
comp.lang.c++ http://www.parashift.com/c++-faq-lite/
alt.comp.lang.learn.c-c++
http://www.contrib.andrew.cmu.edu/~a...FAQ-acllc.html