Re: mDNSResponder build failure on 64-bit platform (amd64)

[Elias Benali <elptr%users.sourceforge.net@localhost>]

On Wed, Sep 30, 2009 at 11:23 AM, Eric Haszlakiewicz 
<erh%nimenees.com@localhost> wrote:
> That doesn't seem to make much sense.  If there really is supposed to
> be a fixed size on-the-wire format as the comment implies, commenting
> out those checks probably isn't the right thing to do.
>
> eric
>

You're right, it was only supposed to make the code compile again.
The author made assumptions about the size of the structure and that should
be dealt with more appropriately. I had to hack it in the mean time.

On Wed, Sep 30, 2009 at 3:42 PM, Ty Sarna <ty%sarna.org@localhost> wrote:
> The dns_sd.h doesn't fail, (and seems extremely unlikely to anyway), so it's
> just commented out to avoid getting a warning in everything using the
> header.

That one is unlikely to fail true, but it still hides the flaw with
the union where the second
member (a pointer) has variable size, thus the on-wire format is
actually wrong there
with non 32-bit machines.

I'm not sure what kind of warning you're referring to if that code was
not ifdef'd out.

On Wed, Sep 30, 2009 at 3:42 PM, Ty Sarna <ty%sarna.org@localhost> wrote:
> Can you provide the whole error message? I started an amd64 cross build and
> I believe I'm already well past the part where this should happen. It was
> also reported to me that it builds on sparc64.

Unfortunately I don't have the error message anymore, but it failed
due to trying to
allocate a -1 sized array. I performed a cross build as well, on i386 Linux