[Linux-ha-dev] HEAD + ANSI Flag
David Lee
t.d.lee at durham.ac.uk
Thu Sep 30 02:29:00 MDT 2004
On Wed, 29 Sep 2004, Lars Ellenberg wrote:
> / 2004-09-29 17:11:08 +0100
> \ David Lee:
> > On Wed, 29 Sep 2004, Andrew Beekhof wrote:
> >
> > > Does anyone know if HEAD compiles (or is close to compiling) with the
> > > ANSI flag turned on?
> >
> > Don't know about HEAD. (I have a local issue of not yet having the uuid
> > stuff.)
>
> I just did on linux:
>
> ConfigureMe configure --enable-crm --enable-ansi --disable-fatal-warnings
> make > make.out 2>&1
> grep "Warning\|error" < make.out > make.err
> sort -t: -k1,1 -k2,2n < make.err | uniq > make.err.sort-u
> [...]
Thanks. At least some of those errors (on Linux) are "sigset_t", which
also appear on Solaris.
Let's just backtrack a moment.
Is the following correct? The intented purpose ("service delivery", "end
of the day" sort of concept) of this "--enable-ansi" was to catch things
such as "//" comments that cause problems to various compilers and that
can be reasonably avoided. Correct? We weren't actually wishing to
enforce ANSI standard C which might cause trouble, as seems to be the
case, with certain ANSI v. POSIX v. ... environmental things (e.g.
"sigset_t"). Correct?
If so, is the "-ansi" option (gcc) the most appropriate vehicle for this?
Or are its side-effects (e.g. "sigset_t" etc.) highly undesirable for us?
Might there be a better way to trap these "//"-like programming errors?
For instance a quick look through "gcc -v --help" (on my 2.95.2 compiler)
shows "-pedantic" and "-std=<standard>" (e.g. "c89", "c9x", etc.). Might
one of these be a better vehicle than "-ansi" for achieving that intended
purpose?
--
: David Lee I.T. Service :
: Systems Programmer Computer Centre :
: University of Durham :
: http://www.dur.ac.uk/t.d.lee/ South Road :
: Durham :
: Phone: +44 191 334 2752 U.K. :
More information about the Linux-HA-Dev
mailing list