Re: Build problem

[Ethan A M. <merritt@u.washington.edu>]

On Saturday 08 July 2006 12:22 pm, Hans-Bernhard Br=F6ker <broeker@physik.r=
wth-aachen.de> wrote:
>=20
> [gcc -Wuninitialized warnings]
>=20
> > I don't know of any way to make gcc more accurate in such cases.
>=20
> Well, just initialize them!  Worst that can happen is that the code=20
> gratuitously grows by 3 'load value into variable' operations.  *If*=20
> they're actually false alarms, that is.

Not quite.  The worst thing that can happen is that by initializing
to some spurious value, you trigger run-time errors that are no longer
accompanied by a relevant compiler warning.  Hiding a possible
problem is worse than leaving the possibly spurious error message.

The problematic constuct is code like the following:

  if (foo) {
    ...
    baz =3D something_calculated;
  }
  [lots of code]
  if (foo) {
    do_something_with(baz);
  }

gcc recognizes that baz is set only inside an if () statement,
but fails to notice that it is only dereferenced inside a=20
matching if () statment.

Yes, you could initialize at the head of the program:
  static double baz =3D 0.0;

But if somehow the pair of if conditions are not truly
equivalent, then you have a problem regardless of the fact
that baz has been initialized.  It is true that in many
cases an incorrect value of 0 is less likely to cause a
segfault than an incorrect value of NaN.  But it's incorrect
either way, and I see no particular virtue in silencing the
compiler's attempt to point out the source of a possible problem.
 =20

=2D-=20
Ethan A Merritt
Biomolecular Structure Center
University of Washington, Seattle 98195-7742