Wireshark-dev: [Wireshark-dev] Commit 54832 (packet-diameter.c HFILL initialization)
From: Joerg Mayer <jmayer@xxxxxxxxx>
Date: Sun, 19 Jan 2014 02:05:40 +0100
Sorry for this form of reply, but I accidentally deleted the original commit
mail...

> Author: Anders Broman <anders.broman@xxxxxxxxxxxx>
> Date:   Fri Jan 17 18:58:50 2014 +0000
>
>     Try to fix warning: initializer element is not computable at load time.
>
>     svn path=/trunk/; revision=54832
>
> diff --git a/epan/dissectors/packet-diameter.c b/epan/dissectors/packet-diameter.c
> index 9a23ab3..1878508 100644
> --- a/epan/dissectors/packet-diameter.c
> +++ b/epan/dissectors/packet-diameter.c
> @@ -1289,15 +1289,22 @@ reginfo(int *hf_ptr, const char *name, const char *abbr, const c
> har *desc,
>         enum ftenum ft, field_display_e base, const value_string_ext *vs_ext,
>         guint32 mask)
>  {
> -       hf_register_info hf = { hf_ptr, {
> -                               name,
> -                               abbr,
> -                               ft,
> -                               base,
> -                               NULL,
> -                               mask,
> -                               desc,
> -                               HFILL }};
> +       hf_register_info hf;
> +
> +       hf.p_id                                         = hf_ptr;
> +       hf.hfinfo.name                          = name;
> +       hf.hfinfo.abbrev                        = abbr;
> +       hf.hfinfo.type                          = ft;
> +       hf.hfinfo.display                       = base;
> +       hf.hfinfo.strings                       = NULL;
> +       hf.hfinfo.bitmask                       = mask;
> +       hf.hfinfo.blurb                         = desc;

> +       /* HFILL */
> +       hf.hfinfo.id                            = -1;
> +       hf.hfinfo.parent                        = 0;
> +       hf.hfinfo.ref_type                      = HF_REF_TYPE_NONE;
> +       hf.hfinfo.same_name_prev_id     = -1;
> +       hf.hfinfo.same_name_next        = NULL;

This is not good: We use HFILL for a reason: To make it trivial to change the
(private) elements of the hf structure. If we need to do this, we should create
another MACRO (e.g. HFILL_INIT) in the same place where we define HFILL and use
that instead of explicitely setting all the values inside a dissector file.

Ciao
      Jörg

--
Joerg Mayer                                           <jmayer@xxxxxxxxx>
We are stuck with technology when what we really want is just stuff that
works. Some say that should read Microsoft instead of technology.