On Wed, Jul 24, 2024 at 11:50:17PM +0200, Stefano Brivio wrote:
> Even though we don't use : as delimiter for the port, making square
> brackets unneeded, RFC 3986, section 3.2.2, mandates them for IPv6
> literals. We want IPv6 addresses there, but some users might still
> specify them out of habit.
> 
> Same for IPv4 addresses: RFC 3986 doesn't specify square brackets for
> IPv4 literals, but I had reports of users actually trying to use them
> (they're accepted by many tools).
> 
> Allow square brackets for both IPv4 and IPv6 addresses, correct or
> not, they're harmless anyway.
> 
> Signed-off-by: Stefano Brivio <sbrivio@redhat.com>

Reviewed-by: David Gibson <david@gibson.dropbear.id.au>

> ---
>  conf.c | 24 +++++++++++++++++-------
>  1 file changed, 17 insertions(+), 7 deletions(-)
> 
> diff --git a/conf.c b/conf.c
> index 3cf9ed8..338742e 100644
> --- a/conf.c
> +++ b/conf.c
> @@ -209,14 +209,24 @@ static void conf_ports(const struct ctx *c, char optname, const char *optarg,
>  
>  		}
>  
> -		if (ifname == buf + 1)		/* Interface without address */
> +		if (ifname == buf + 1) {	/* Interface without address */
>  			addr = NULL;
> -		else if (inet_pton(AF_INET, buf, addr))
> -			af = AF_INET;
> -		else if (inet_pton(AF_INET6, buf, addr))
> -			af = AF_INET6;
> -		else
> -			goto bad;
> +		} else {
> +			p = buf;
> +
> +			/* Allow square brackets for IPv4 too for convenience */
> +			if (*p == '[' && p[strlen(p) - 1] == ']') {
> +				p[strlen(p) - 1] = 0;

Nit: I think = '\0' would make the intention here clearer.

> +				p++;
> +			}
> +
> +			if (inet_pton(AF_INET, p, addr))
> +				af = AF_INET;
> +			else if (inet_pton(AF_INET6, p, addr))
> +				af = AF_INET6;
> +			else
> +				goto bad;
> +		}
>  	} else {
>  		spec = buf;
>  

-- 
David Gibson (he or they)	| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you, not the other way
				| around.
http://www.ozlabs.org/~dgibson