https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100417
--- Comment #3 from jkb at sanger dot ac.uk --- I'd definitely be in favour of John's rewording of the warning - "data pointed to by ...". This definitely led us up the garden path for a while. While I agree the code could have been written differently, such as not using a const char *, or having a length parameter instead of end pointer, I don't feel it is the compilers responsibility to prefer one style over another. That said, I concur that the minimal change to ignoring this check for zero sized objects would catch the more common idioms of bounds checking. Thank you for the response.
