Hello Mike,
Thank you for the patch.
I did check it and I have some proposal changes:
1) In epan/packet.c, epan/stream.c and wiretap/catapult_dct2000.c
Change GPOINTER_TO_INT to GPOINTER_TO_UINT
2) In gtk/dcerpc_stat.c
I did add a cast, (guint) to compile with a 32 bits linux system (see the
following patch)
(See attached file: dcerpc_stat.c.diff.gz)
3) In packet_h223.c
The hash key calculation has been changed.
<<
- guint hash_val = ((guint32)(key->call))^(((guint32)key->vc) << 16);
+ guint hash_val = ((guint32)key->vc)^(((guint32)key->vc) << 16);
>>
If we change only the syntax to compile without warning, we should have:
<<
- guint hash_val = ((guint32)(key->call))^(((guint32)key->vc) << 16);
+ guint hash_val = GPOINTER_TO_UINT(key->call)^(((guint32)key->vc) <<
16);
>>
4) in gtk/expert_comp_table.c and other expert info
I think, there is a faulty cast for the expert_data.goup.
The goup, defined the expert info structure is an integer, but is used
here, as a string.
<<
- if (strcmp((char *)expert_data.group, "Packet:")==0) {
+ if (strcmp(GINT_TO_POINTER(expert_data.group), "Packet:")==0) {
>>
I did make a change in the structure:
- to have the goup_value, as an integer
- and the group_column, used as a string.
So, we have not to make such cast.
(See attached file: expert_info.diff.gz)
Could you have a look to my comment, and synchronize with your patch
(because, I do not have currently a clean SVN view) ?
Regards
Florent
Attachment:
dcerpc_stat.c.diff.gz
Description: Binary data
Attachment:
expert_info.diff.gz
Description: Binary data