Data types in samba4

[Anthony Liguori]

It may make sense to consider the C99 `bool' type too (defined in 
<stdbool.h>).  See 
http://www.gnu.org/software/autoconf/manual/autoconf-2.57/html_chapter/autoconf_5.html#SEC50 
for the recommended workaround if <stdbool.h> is unavailable.

Anthony Liguori
Linux/Active Directory Interoperability
Linux Technology Center (LTC) - IBM Austin
E-mail: aliguor at us.ibm.com
Phone: (512) 838-1208
Tie Line: 678-1208




"Green, Paul" <Paul.Green at stratus.com>
Sent by: samba-technical-bounces+aliguor=us.ibm.com at lists.samba.org
15/01/2004 12:14 PM

 
        To:     "'Christopher R. Hertel'" <crh at ubiqx.mn.org>, "Stefan (metze) Metzmacher" 
<metze at samba.org>
        cc:     samba-technical at samba.org
        Subject:        RE: Data types in samba4



I basically think it is sensible to define your own set of types. However,
it will certainly entail a large change in mindset.

I'd urge some careful consideration on the names that are chosen.  It 
would
be nice to make them obvious and extensible w/o colliding with potential
identifiers.

As far as I can see, reading the C-99 and POSIX-01 standards, uint_t is 
not
a standard type.  uintN_t, where N is an unsigned decimal integer, is
defined in C-99 (See paragraph 7.18).  The standard says such types are
optional.  (7.18.1.1).

The POSIX standard reserves all identifiers ending in "_t".  We should not
define any Samba-specific identifiers that use this suffix.

Since some of us are still on C-89 and POSIX-96, any attempt to use types 
in
more recent standards would need to have the usual conditional logic in
configure.  (And given the wide variation of implementations, such logic
would be a very good thing anyway).

Thanks
PG
--
Paul Green, Senior Technical Consultant, Stratus Technologies.
Voice: +1 978-461-7557; FAX: +1 978-461-3610; Video on request.