Re: Using std::cin.rdbuf()->in_avail()
Paul wrote:
if (std::cin >> ch) {
You realise the above skips whitespace?
Yes.
That means it reads an arbitrary number of characters, and thus may
block (e.g. if only whitespace has been entered).
std::cin.sync(); // flush remaining characters
cin.sync() doesn't do anything.
Surprisingly, it helped, although I accept that its behaviour is
unspecified.
That's interesting - I haven't tried it with recent compilers, though it
didn't work in the past.
Since I was reading only a character at a time (std::cin >>
ch), the new-line character remained unread and interfered with processing
later on whilst .sync() appeared to clear that. My logic at the time was
that since sync() is supposed to work for std::istream like .flush() for
std::ostream and since the latter flushes whatever we have to the real
destination, sync() would synchronise the internal buffer with the real
input source, i.e. clear the buffer. But I turned to using
std::cin.ignore(std::cin.rdbuf()->in_avail()).
That's definitely safer I think.
Tom