hipl-core team mailing list archive

[Christof Mroz <christof.mroz@xxxxxxxxxxxxxx>]

On Sun, 29 Aug 2010 17:12:38 +0200, Diego Biurrun <diego@xxxxxxxxxx> wrote:

> This 8 characters idea is completely arbitrary and IMO not an improvement
> at all.  Creating more and arbitrary rules is not good.

I agree, but I saw some instances of code in HIPL using an indent of 8  
spaces so I assumed that was the right thing to do... well in this case,  
there's no need to change doc/HACKING I think.

> > @@ -553,6 +555,23 @@
> >
> > +static void blah(int a, int b) {
> > +    /* excessive nesting should be avoided but if it can't,
> > +     * remember the following rules */
> > +    if(a) {
> > +        if(b) {
> > +            struct opaque blah;
>
> *sigh* - This is not even in K&R style.  May I politely suggest that you
> go back to reading its definition?  (hint: spaces, keywords)

Alright

> > +            /* opening '(' on next line if no argument fits */
> > +            yet_another_function_with_a_long_name
> > +                    (a_function_with_a_long_name(NULL, &blah,  
> > 0xC0FFEE));
>
> This is not K&R style and completely counterproductive.  You destroy the
> visual cue that a string followed by an opening parenthesis gives you -
> in K&R style you immediately know it's a function call, now you don't
> anymore.

I prefer the "dangling" opening parenthesis too, in this case I was  
inspired by another mail of yours:

On Fri, 20 Aug 2010 16:56:49 +0200, Diego Biurrun <diego@xxxxxxxxxx> wrote:

> On Thu, Aug 19, 2010 at 09:34:25AM +0000, noreply@xxxxxxxxxxxxx wrote:
> > ------------------------------------------------------------
> > revno: 4898
> > committer: Artturi Karila <artturi.karila@xxxxxxxxxx>
> > branch nick: trunk
> > timestamp: Thu 2010-08-19 12:32:20 +0300
> > message:
> >   Fixed 'cast discards qualifiers' warnings in hip_get_next_param()
> >
> > --- firewall/conntrack.c	2010-08-18 16:21:16 +0000
> > +++ firewall/conntrack.c	2010-08-19 09:32:20 +0000
> > @@ -925,10 +925,12 @@
> >      // store the public key separately
> >      // store function pointer for verification
> >      if (hip_get_host_id_algo(tuple->hip_tuple->data->src_hi) ==  
> > HIP_HI_RSA) {
> > -        tuple->hip_tuple->data->src_pub_key =  
> > hip_key_rr_to_rsa((struct hip_host_id_priv *) host_id, 0);
> > +        tuple->hip_tuple->data->src_pub_key = hip_key_rr_to_rsa(
> > +                                 (const struct hip_host_id_priv *)  
> > host_id, 0);
>
> Breaking functions at the opening parenthesis looks ugly IMO.