This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Wmissing-parameter-type (PR7651 Define -Wextra strictly in terms of other warning flags)
On 20/12/06, Joseph S. Myers <joseph@codesourcery.com> wrote:
On Wed, 19 Dec 2006, DJ Delorie wrote:
> It's not a new warning, but it is a new *option*. You'll need test
> cases that test various cases:
>
> 1. No option specified
> 2. Option on (warn), option off (no warn)
> 3. Option on, marked as an error, marked as a warning overriding -Werror
>
> etc.
-Wextra enabling warning, -Wextra -Wno-new-option disabling the warning.
Wpointer-sign*.c are examples where that sort of thing is tested.
Yes, I'd like such things as dg-pedwarn, dg-pedantic-pedwarn to avoid some
triples of tests I've added in the past (default, -pedantic,
-pedantic-errors). But for now we work with the testsuite infrastructure
we have.
I am just bootstrapping and testing the patch including the testcases
that you suggest. But still it makes me wonder why me and why now. Not
because I don't want to do it, but because almost no other option in
Wall or Wextra has such testcases, Wpointer-sign being one of the few
exceptions.
In the same manner, why aren't required similar testcases for
optimisation options enabled by -O1, -O2, etc? One testcase for the
default behaviour, other for -O1 enabling the optimisation, another
for -O1 -fno- disabling it, etc.
Again, I am all for this (will send the updated patch soon) but the
requirement still seems rather arbitrary.
Cheers,
Manuel.