Re: small patch allowing build with -stdlib=libc++

[Daniel Miller <bonsaiviking () gmail com>]

On 04/17/2014 02:23 PM, olli hauer wrote:
> Hi,
>
> is it possible to adopt the following patch to make the compiler happy and also allow usage of -stdlib=libc++
>
>
> Index: scan_engine.cc
> ===================================================================
> --- scan_engine.cc      (revision 32836)
> +++ scan_engine.cc      (working copy)
> @@ -3129,7 +3129,7 @@
>     }
>     if (o.spoofsource && !bind_failed) {
>       o.SourceSockAddr(&ss, &sslen);
> -    if (bind(sd, (struct sockaddr*)&ss, sslen) != 0) {
> +    if (::bind(sd, (struct sockaddr*)&ss, sslen) != 0) {
>         error("%s: Problem binding source address (%s), errno: %d", __func__, inet_socktop(&ss), socket_errno());
>         perror("bind");
>         bind_failed = 1;
> Index: traceroute.cc
> ===================================================================
> --- traceroute.cc       (revision 32836)
> +++ traceroute.cc       (working copy)
> @@ -221,8 +221,6 @@
>   /* A global random token used to distinguish this traceroute's probes from
>      those of other traceroutes possibly running on the same machine. */
>   static u16 global_id;
> -/* A global cache of known hops, indexed by TTL and address. */
> -static std::map<struct HopIdent, Hop *> hop_cache;
>   /* A list of timedout hops, which are not kept in hop_cache, so we can delete
>      all hops on occasion. */
>   static std::list<Hop *> timedout_hops;
> @@ -235,6 +233,9 @@
>   static struct timeval get_now(struct timeval *now = NULL);
>   static const char *ss_to_string(const struct sockaddr_storage *ss);
>
> +/* A global cache of known hops, indexed by TTL and address. */
> +static std::map<struct HopIdent, Hop *> hop_cache;
> +
>   struct Hop {
>     Hop *parent;
>     struct sockaddr_storage tag;
>
>
>
>
> _______________________________________________
> Sent through the dev mailing list
> http://nmap.org/mailman/listinfo/dev
> Archived at http://seclists.org/nmap-dev/
Olli,

I appreciate your reports improving Nmap's standards compliance. I think 
I understand qualifying ::bind to avoid collision with std::bind, but 
what is the purpose of the second part of the patch? What is the error 
that clang++ is giving? I don't have my clang build environment up and 
running at the moment, unfortunately.

Dan
_______________________________________________
Sent through the dev mailing list
http://nmap.org/mailman/listinfo/dev
Archived at http://seclists.org/nmap-dev/