[PATCH 10/11] argp: Fix shift bug

[Florian Weimer]

* Adhemerval Zanella:

>>From gnulib commits 06094e390b0 and 88033d3779362a.
> ---
>  argp/argp-parse.c | 15 +++++++++------
>  1 file changed, 9 insertions(+), 6 deletions(-)
>
> diff --git a/argp/argp-parse.c b/argp/argp-parse.c
> index 82c7b784de..99f8d9ecd4 100644
> --- a/argp/argp-parse.c
> +++ b/argp/argp-parse.c
> @@ -735,12 +735,15 @@ parser_parse_opt (struct parser *parser, int opt, char *val)
>  	    }
>      }
>    else
> -    /* A long option.  We use shifts instead of masking for extracting
> -       the user value in order to preserve the sign.  */
> -    err =
> -      group_parse (&parser->groups[group_key - 1], &parser->state,
> -		   (opt << GROUP_BITS) >> GROUP_BITS,
> -		   parser->opt_data.optarg);
> +    /* A long option.  Preserve the sign in the user key, without
> +       invoking undefined behavior.  Assume two's complement.  */
> +    {
> +      int user_key =
> +        ((opt & (1 << (USER_BITS - 1))) ? ~USER_MASK : 0) | (opt & USER_MASK);

Would this be clearer?

		   (int) ((unsigned int) opt << GROUP_BITS) >> GROUP_BITS,

Or does ubsan flag that as well?  Conversion to negative int is a GCC
extension:

| For conversion to a type of width N, the value is reduced modulo 2^N
| to be within range of the type; no signal is raised.

<https://gcc.gnu.org/onlinedocs/gcc-15.1.0/gcc/Integers-implementation.html>