[OE-core] [PATCH 08/17] connman: Fix builds to compile on musl

[Khem Raj]

> On Apr 13, 2015, at 9:50 AM, Burton, Ross <ross.burton at intel.com> wrote:
> 
> On 11 April 2015 at 02:20, Khem Raj <raj.khem at gmail.com <mailto:raj.khem at gmail.com>> wrote:
> ++#ifdef __GLIBC__
> + #include <execinfo.h>

yeah ideally, I could just use libunwind and get the backtrace API for musl too but I will do that when I need it.

> 
> As you're checking for this header at configure time wouldn't it be best to use #ifdef HAVE_EXECINFO_H, just in case another C library adds it later?
> 
> + /* from netinet/in.h */
> ++#if 0
> + struct in6_pktinfo {
> +       struct in6_addr ipi6_addr;  /* src/dst IPv6 address */
> +       unsigned int ipi6_ifindex;  /* send/recv interface index */
> + };
> ++#endif
> 
> #if 0?  Is this entirely redundant code (in which case, delete it), or is this disabling something?

it is redundant if you are including netinet/in.h

>  
> ++++ b/gweb/gresolv.c
> +@@ -876,8 +876,6 @@ GResolv *g_resolv_new(int index)
> +       resolv->index = index;
> +       resolv->nameserver_list = NULL;
> +
> +-      res_ninit(&resolv->res);
> +-
> +       return resolv;
> + }
> +
> +@@ -917,8 +915,6 @@ void g_resolv_unref(GResolv *resolv)
> +
> +       flush_nameservers(resolv);
> +
> +-      res_nclose(&resolv->res);
> +-
> +       g_free(resolv);
> + }
> 
> Surely these should be checked at configure time or wrapped in GLIBC tests, with the new codepath as a fallback.  Ideally we can upstream most of this.
>  
> Ross 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openembedded.org/pipermail/openembedded-core/attachments/20150413/93f2b846/attachment-0002.html>