Topic: WP def of underflow()


Author: peter@chinook.physics.utoronto.ca (Peter Berdeklis)
Date: 1995/06/22
Raw View

This is from [lib.streambuf.virt.get], under the definition
of underflow() - Effects:

--------
If eback() and gptr() are non-null then the function is not
constrained as to their contents, but the "usual backup condition"
is that either:

a) If the backup sequence contains at least gptr() - eback()
characters, then the gptr() - eback() characters starting at eback()
agree with the last gptr() - eback() characters of the backup
sequence.

b) Or the n characters starting at gptr() - n agree with the backup
sequence (where n is the length of the backup sequence)
-------


Am I misreading this, or are a) and b) exactly the same?

If n = gptr() - eback() then there is no difference between a) and b).
How could n != gptr() - eback()?

Of course, because of the disclaimer at the top, it doesn't really
matter, but I think its confusing to repeat the same thing and
suggest that there's a difference (because of the "either"/"Or").



Pete