Re: [head tinderbox] failure on amd64/amd64

On Sat, 4 Feb 2006, Max Laier wrote:

> On Saturday 04 February 2006 00:07, FreeBSD Tinderbox wrote:
>> -finstrument-functions -Wno-inline /src/sys/security/audit/audit_arg.c
>> /src/sys/security/audit/audit_arg.c: In function `audit_arg_upath':
>> /src/sys/security/audit/audit_arg.c:676: warning: long long unsigned int
>> format, u_int64_t arg (arg 2) /src/sys/security/audit/audit_arg.c:678:
>> warning: long long unsigned int format, u_int64_t arg (arg 2) *** Error
>> code 1
>
> Once again, fixed by - for example - the attached patch.  Alternatively:
> 	.. "%ju", (uintmax_t)arg ..	// we are so C99
>
> I take this a chance to rant if we could remove this stupid 64bit 
> unportability.  It should be possible to have a CASSERT somewhere that shuts 
> this up if "sizeof(u_int64_t) == sizeof(unsigned long long)" or the like. 
> I hope somebody has more insight and how/where to fix it properly. 
> Otherwise we will run into this portability issue over and over again.

Thanks.  I just realized I did my amd64 test builds without INVARIANTS 
compiled in.  Patch committed.  No opinion on the format string thing, except 
that I keep shooting the feet of myself and others.

Robert N M Watson