Re: warning: comparison is always false due to limited range of data type

[Jim Meyering]

Paul Eggert <address@hidden> wrote:
>> "Oskar Liljeblad" <address@hidden> wrote:
...
>>> quotearg.c: In function `quotearg_n_options':
>>> quotearg.c:586: warning: comparison is always false due to limited range of
>>> data type
...
> My own experience is that that particular warning is more trouble
> than it's worth.  Is there some easy way to ask GCC to not generate
> that warning?

GCC has no option to suppress that warning.

It's not pretty, but maybe as a last resort...
The following work-around suppresses the warning when
compiling with -Dlint.

Add this,

#ifdef lint
typedef size_t unsigned_int;
#else
typedef unsigned int unsigned_int;
#endif

and then change this:
      unsigned int n1 = n0 + 1;
to this:
      unsigned_int n1 = n0 + 1;

On 64-bit systems, that'd give warning-free builds with -Dlint,
at the expense of a small inefficiency.  Without -Dlint, you'd
still get the tighter generated code, along with the warning.
On 32-bit systems, it should make no difference.

But then due diligence would seem to require
a comment justifying this apparent micro-optimization.
If using `size_t' unconditionally means you don't have to make
excuses in the comments, then perhaps it's better after all.
The difference in text size (using gcc-4.0 -O3) is just 16 bytes.

Besides, who knows... compilers may eventually be able to optimize
away that xalloc_die call, in spite of the intermediate size_t.