Re: VMS patch set (extensive) - first try (almost)

>    Perhaps I exagerated a bit, but if the problem is with the progress
> message, then wouldn't a flush (and sync) before the message make more
> sense?

There is no need to sync at all.  All I care about is giving the data to the OS (instead of keeping
it in stdio's buffers), I don't need it written out to a physical device immediately.

Imagine if something like sendfile were used, or if the code used file descriptor IO directly: it
would be quite reasonable to write() out the chunk of data that has been read from the network. 
That is what fflush does -- nothing more and nothing less.