cvs commit: ports/devel/dmucs Makefile ports/devel/dmucs/files patch-dmucs_db.cc patch-dmucs_db.h patch-dmucs_msg.cc patch-main.cc

[Max Khon]

Hello!

Pietro Cerutti wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA512
> 
> Max Khon wrote:
> | fjoe        2008-07-07 16:17:54 UTC
> |
> |   FreeBSD ports repository
> |
> |   Modified files:
> |     devel/dmucs          Makefile
> |   Added files:
> |     devel/dmucs/files    patch-dmucs_db.cc patch-dmucs_db.h
> |                          patch-dmucs_msg.cc patch-main.cc
> |   Log:
> |   Unbreak on 64-bit archs
> 
> I definitely do not agree. Please note that a pointer is not required to
> fit into a long, while it is required to fit into a size_t.
> 
> Your solution may work as soon as we are concerned only with LP64
> machines. As soon as a LLP64 arch steps in, you break things.

AFAIK it is not guaranteed that sizeof(void *) <= sizeof(size_t) too.
I used "long" because I did not want to cope with "[u]intptr_t" there.

Ok, I committed a modified patch that uses "const void *" instead of "long".

/fjoe