On Mon, 2 Oct 2000, Poul-Henning Kamp wrote:
> Trying to compile the ircII port, it crashes with:
> ===> Building for ircii-4.4X
> cc -I. -I/usr/ports/irc/ircII/work/ircii-4.4X/include -O -pipe -DHAVE_CONFIG_H -c
> In file included from /usr/include/sys/queue.h:40,
> from /usr/include/netinet6/in6.h:76,
> from /usr/include/netinet/in.h:475,
> from /usr/ports/irc/ircII/work/ircii-4.4X/include/irc.h:79,
> from /usr/ports/irc/ircII/work/ircii-4.4X/source/alias.c:35:
> /usr/ports/irc/ircII/work/ircii-4.4X/include/struct.h:51: syntax error before `_'
> This is because there is a file called "struct.h" in the ircII distribution.
> The addition of #include <struct.h> to <sys/queue.h> looks a bit unsettling
> to me: A sys/* file shouldn't reference a /usr/include file I think ?
I corresponded with the author of this bug soon after it was committed, but
didn't follow up.
> Is the correct thing not to mv <struct.h> to <sys/struct.h>, create
> a <struct.h> which just pulls in <sys/struct.h> and have <sys/queue.h>
> #include <sys/struct.h> ?
Of course not. Just expand the one macro that is used (strbase()) in the
one place that it is used in <sys/queue.h>.
I think <struct.h> should have been removed when offsetof() became
standard. It was only used in one place in /usr/src (lastcomm). Only
the fldsiz() macro is used there. <struct.h> defines 2 other macros:
fldoff() and strbase(). fldoff() is a bad version of offsetof() (one
that takes a struct tag instead of a type, so it can't be used with
typedefed structs). strbase() is now used in <sys/queue.h>. It is
easy to implement directly using offsetof(). Unfortunately, if it is
implemented using offsetof(), then <sys/queue.h> will depend on
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message