[Bug middle-end/119693] GCC assumes wrong bounds for strlen() return value, causing bogus bounds check elimination

[jannh at google dot com via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119693

--- Comment #6 from Jann Horn <jannh at google dot com> ---
@Andrew/Jakub: I am not convinced that that is theoretically in line with the
standard, but you've convinced me that it might be somewhat reasonable for the
compiler to assume in practice.

(Though regarding the mmap() part, I don't think I've seen any documentation of
that restriction in https://gcc.gnu.org/onlinedocs/gcc/C-Implementation.html or
so, or in the Linux mmap() manpage, or the POSIX spec for mmap.)

Anyway, I think the most important aspect here is the off-by-one.