On Thu, Nov 28, 2013 at 01:40:31PM -0800, Guy Harris wrote: > > On Nov 28, 2013, at 1:39 PM, Jakub Zawadzki <darkjames...@darkjames.pl> wrote: > > > On Tue, Nov 19, 2013 at 07:15:54PM +0100, Jakub Zawadzki wrote: > >> htons(), htonl(), htonll() is kinda easier to write and looks prettier for > >> me than hton16, hton32, hton64(). > > > > It seems that such change would get us to have name conflict with > > <endian.h> (at least on Linux), > > phtoXX() take a pointer and handles unaligned data; renaming them won't cause > a conflict with the Linux #defines. > > htoXX() take an argument and just byte-swaps as necessary; they're currently > just wrappers for GLib macros: > > /* Turn host-byte-order values into little-endian values. */ > #define htoles(s) GUINT16_TO_LE(s) > #define htolel(l) GUINT32_TO_LE(l) > > so one option would simply be to get rid of htoles() and htolel() and just > use the GLib macros directly.
If we want to go this way (use glib macros directly), we can also kill BSWAP16, BSWAP32, BSWAP64 macros, and replace with: GUINT16_SWAP_LE_BE, GUINT32_SWAP_LE_BE, GUINT64_SWAP_LE_BE Ok? ___________________________________________________________________________ Sent via: Wireshark-dev mailing list <wireshark-dev@wireshark.org> Archives: http://www.wireshark.org/lists/wireshark-dev Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev mailto:wireshark-dev-requ...@wireshark.org?subject=unsubscribe