On Wed, 15 May 2013, I wrote:
> The patch is correct,
Thinking about it a bit more, I have come to the conclusion that perhaps
the type should really be pcre_int32 rather than just int, as otherwise
there may be issues in environments where int is just 16 bits wide.
However, are there any such environments still in use nowadays?
Also, I suspect that there are other int and unsigned int variables that
ideally should be pcre_int32 and pcre_uint32, so it probably is not
worth worrying about until there is the time and energy to go through
the whole source. (The same applies to checking all the signed/unsigned
variables. I'm afraid I wasn't very organized when I first wrote the
code.)
Philip
--
Philip Hazel
