Reply to thread

Message

[QUOTE="Thomas Matthews, post: 2383238"]
No.  The point is that the index variable is modified
within the block of statements.  This was _not_ designed
as a forever loop.

Nope.  The point here is that floating point arithmetic
and comparison are not exact enough for this loop.  One
should not use floating point values for indices in a
loop.

This is another issue on unspecified behavior.
The loop causes an underflow at m == 1.  One cannot
subract the value of "2" from "1" when the domain
is restricted to all positive integers (including
zero).  This is undefined in terms of mathematics.

-- 
Thomas Matthews

C++ newsgroup welcome message:
          [URL]http://www.slack.net/~shiva/welcome.txt[/URL]
C++ Faq: [URL]http://www.parashift.com/c++-faq-lite[/URL]
C Faq:   [URL]http://www.eskimo.com/~scs/c-faq/top.html[/URL]
alt.comp.lang.learn.c-c++ faq:
          [URL]http://www.comeaucomputing.com/learn/faq/[/URL]
Other sites:
     [URL]http://www.josuttis.com[/URL]  -- C++ STL Library book
[/QUOTE]

Verification