> From: l...@gnu.org (Ludovic Courtès)
> Cc: m...@netris.org, guile-user@gnu.org
> Date: Sun, 16 Jun 2013 17:04:15 +0200
>
> Eli Zaretskii <e...@gnu.org> skribis:
>
> > The changes below work around the lack of h_error on Windows:
> >
> > --- libguile/net_db.c~0 2013-03-25 23:44:46.000000000 +0200
> > +++ libguile/net_db.c 2013-06-13 13:30:20.448923200 +0300
> > @@ -55,6 +55,47 @@
> > #include "libguile/net_db.h"
> > #include "libguile/socket.h"
> >
> > +#ifdef __MINGW32__
> > +# define h_errno WSAGetLastError()
> > +# define HAVE_H_ERRNO 1
> > +# define HAVE_HSTRERROR 1
> > +# ifdef HAVE_DECL_HSTRERROR
> > +# undef HAVE_DECL_HSTRERROR
> > +# endif
> > +# define HAVE_DECL_HSTRERROR 1
> > +# define HOST_NOT_FOUND WSAHOST_NOT_FOUND
> > +# define TRY_AGAIN WSATRY_AGAIN
> > +# define NO_RECOVERY WSANO_RECOVERY
> > +# define NO_DATA WSANO_DATA
> > +
> > +char *hstrerror (int);
> > +
> > +char *
> > +hstrerror (int error_no)
> > +{
> > + static char buf[500];
>
> Rather allocate it dynamically, with ‘scm_gc_malloc_pointerless’.
>
> > + DWORD ret;
> > +
> > + if (error_no == 0)
> > + error_no = WSAGetLastError ();
> > +
> > + ret = FormatMessage (FORMAT_MESSAGE_FROM_SYSTEM |
> > + FORMAT_MESSAGE_IGNORE_INSERTS,
> > + NULL,
> > + error_no,
> > + 0, /* choose most suitable language */
> > + buf, sizeof (buf), NULL);
> > +
> > + while (ret > 0 && (buf[ret - 1] == '\n' ||
> > + buf[ret - 1] == '\r' ))
> > + --ret;
> > + buf[ret] = '\0';
> > + if (!ret)
> > + sprintf (buf, "Winsock error %u", error_no);
> > +
> > + return buf;
> > +}
> > +#endif /* __MINGW32__ */
>
> This looks good, but I feel that this ought to be in a Gnulib module,
> no? Would you like to propose such a module on bug-gnulib@?
I will try.
Thanks.
