[guus@debian.org: Bug#290306: dbus-1-dev: silly typedefs]

[Havoc Pennington]

On Thu, 2005-01-13 at 23:41 +0100, Guus Sliepen wrote:
> On Thu, Jan 13, 2005 at 04:48:17PM -0500, Havoc Pennington wrote:
> 
> > > Sure, but as you can see, the current typedefs could well be the cause
> > > of a security hole on another architecture. The blind wrapping done in
> > > dbus-types.h is detrimental to security.
> [...]
> > It just isn't a problem in practice. I think glib and Qt (and a lot of
> > other libraries, for that matter) have shown that.
> 
> Of course it is not a problem for glib, because if you compile glib, it
> checks the size of (short/long/long long) int in the configure script
> and then actually uses the results to create the appropriate typedefs. I
> see that dbus's configure script checks for the sizes, but then does
> nothing with the result, except to define the 64 bit types.
> 

Exactly, that's what I meant in my first post where I said that when
porting to any platform where int isn't 32-bit, you would move
dbus_int32_t to dbus-arch-deps.h just like dbus_int64_t.

Havoc