Re: undefined-behavior obstack.c:139

Am Fr., 1. Dez. 2023 um 21:01 Uhr schrieb Paul Eggert <[email protected]>:

> On 2023-12-01 10:40, Bruno Haible wrote:
>
> > Indeed, this sentence appears to forbid ((char *) NULL) + something.
>
> Yes. However, Gnulib code can still use ((char *) NULL) + something)
> because the Gnulib portability guidelines allow it.
>
> The issue with clang false positives is covered here:
>
>
> https://www.gnu.org/software/gnulib/manual/html_node/Unsupported-Platforms.html
>
> which lists "clang -fsanitize=undefined" as an unsupported platform
> unless you also specify "-fno-sanitize=pointer-overflow".
>
> The obstack patch you installed is fine, as it's clearer and just as
> fast as the original. However, we needn't go through Gnulib and change
> other code merely because it runs afoul of this false alarm from clang.
>

It may not be a false alarm in future versions of the compiler.  At any
time, the compiler may decide to replace ((char *) NULL + 0) by
__builtin_unreachable.  It can make sense to find an ISO C-compliant
replacement of this idiom at some point.