Re: iphlpapi: Only call res_init() once per process

[Kai Blin]

On Wednesday 18 April 2007 20:39, Hans Leidekker wrote:
> +static int resolver_initialised;

Shouldn't that be

static int resolver_initilised = 0;

here? I thought that in C, ints aren't initialized to 0 automatically.
> +/* call res_init() just once because of a bug in Mac OSX 10.4 */
> +static void initialise_resolver( void )
> +{
> - if (!resolver_initialised)
> - {
> -
>    res_init();
> -
>    resolver_initialised = 1;
> - }
>
> +}

Cheers,
Kai

iphlpapi: Only call res_init() once per process

[H. Verbeet]

On 19/04/07, Kai Blin  wrote:
> On Wednesday 18 April 2007 20:39, Hans Leidekker wrote:
> > +static int resolver_initialised;
>
> Shouldn't that be
>
> static int resolver_initilised = 0;
>
> here? I thought that in C, ints aren't initialized to 0 automatically.

Afaik static variables are.

iphlpapi: Only call res_init() once per process

[James Liggett]

On Thu, 2007-04-19 at 07:49 +0200, Kai Blin wrote:
> On Wednesday 18 April 2007 20:39, Hans Leidekker wrote:
> > +static int resolver_initialised;
>
> Shouldn't that be
>
> static int resolver_initilised = 0;
>
> here? I thought that in C, ints aren't initialized to 0 automatically.

Actually I think in this case they would be. Most compilers do this by
default IIRC (if you'd like to see for yourself, try the attached test.)
Still, I think it would be prudent to put the 0 in there anyway, sine I
don't think there's a standard that says compilers have to do this so
there could potentially be portability problems.

HTH,
James

iphlpapi: Only call res_init() once per process

[Alexandre Julliard]

James Liggett  writes:
> Actually I think in this case they would be. Most compilers do this by
> default IIRC (if you'd like to see for yourself, try the attached test.)
> Still, I think it would be prudent to put the 0 in there anyway, sine I
> don't think there's a standard that says compilers have to do this so
> there could potentially be portability problems.

There are no potential problems at all, statics are _always_
initialized to 0 in C.

iphlpapi: Only call res_init() once per process

[Kai Blin]

On Thursday 19 April 2007 23:34, Alexandre Julliard wrote:
> James Liggett  writes:
> > Actually I think in this case they would be. Most compilers do this by
> > default IIRC (if you'd like to see for yourself, try the attached test.)
> > Still, I think it would be prudent to put the 0 in there anyway, sine I
> > don't think there's a standard that says compilers have to do this so
> > there could potentially be portability problems.
>
> There are no potential problems at all, statics are _always_
> initialized to 0 in C.

And another new thing learned about C. Thanks for the explanations.

Kai