Created attachment 6274 [details] [review]
fix byte-swapping for 32-bit

This patch

a) moves the byte-swapping functions up since we'll otherwise get warnings due
to implicit function declarations

b) replaces use of non-existent swap_8_bytes function by swap_bytes

Created attachment 6275 [details] [review]
make tools compile

This patch makes the tools compile with 32-bit. The patch is obviously wrong
and not meant to be applied. It just shows where the issues are. I'm not sure
what the correct fix would look like.

This patch looks correct.  If you don't have 64bit types then you are in a
special environment and can't accept 64 bit types over the bus.

Havoc do you have any objections to this patch?

The first patch looks right and should go in.

The second patch isn't quite right; DBUS_TYPE_INT64 should still be handled / 
assumed to exist if not DBUS_HAVE_INT64 - DBUS_HAVE_INT64 has to do with 
dbus_int64_t, not DBUS_TYPE_INT64. The right thing to do when encountering a 
DBUS_TYPE_INT64 on a platform without dbus_int64_t varies case-by-case.

BTW, what is the system you're using that lacks 64-bit?

In the end it turned out the "lack" of 64 bit types was a problem with whatever
autoconf version was used to build the package. Simply rerunning autoconf fixed
the build for me. So, in a way, I was trying to fix someone else's problems, I
guess... ;-)

The last comment indicates that this problem isn't a dbus problem and could be closed. Feel free to reopen the bug if required.