dbus remaining windows port issues
fridrich.strba at bluewin.ch
Mon May 3 01:18:29 PDT 2010
OK, I modified some things according to the discussion on this list.
1) I incorporated the cmake stuff in the same patch
2) We include inttypes.h which is more portable according to those that
It is arguable whether one should include config.h in those two private
headers. It is eventually possible to make the same include block:
in every *.c file before including dbus-pipe.h or dbus-hash.h. But these
two are private headers and the [u]intptr_t appears in them as a
non-pointer type. I would really argue that the includes are better kept
in the two headers so that if another *.c file includes one of them, the
developer will not have to grep around for how his build-breakages are
fixed in other files. As usual, no religion about it though, as long as
the things work for me :)
So, check the attached consolidated patch
On Sun, 2010-05-02 at 23:59 +0200, Ralf Habacker wrote:
> Fridrich Strba schrieb:
> > On Thu, 2010-04-22 at 12:33 +0200, Ralf Habacker wrote:
> >> DBUS_USE_NONCE_TCP_DEFAULT_ADDRESS Use nonce tcp default address
> >> default=OFF
> > in configure.in, there is this fragment:
> > if test x$dbus_win = xyes; then
> > DBUS_SESSION_BUS_DEFAULT_ADDRESS="nonce-tcp:"
> > else
> > DBUS_SESSION_BUS_DEFAULT_ADDRESS="unix:tmpdir=$DBUS_SESSION_SOCKET_DIR"
> > fi
> > AC_SUBST(DBUS_SESSION_BUS_DEFAULT_ADDRESS)
> > Just wondering whether this does not mean a divergence between cmake and autofoo default.
> thanks for this pointer - this comes from the merge of the windbus and the dbus4win project.
> dbus mailing list
> dbus at lists.freedesktop.org
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 25572 bytes
Desc: not available
More information about the dbus