I am seeking for a way to suppress all possible warnings that i may get with Gcc with pragma directives. I had made some guard macros that help me silence 3rd party headers from warnings, and for now they work like charm for msvc and clang. I am still missing the correct way to use Gcc diagnostic pragmas in order to suppress every warning in a section. Let me give you some examples:
In msvc we can do this:
#pragma warning(push, 0)
// Code that produces warnings...
#pragma warning(pop)
And in clang we can do this:
#pragma clang diagnostic push
#pragma clang diagnostic ignored "-Wall"
#pragma clang diagnostic ignored "-Wextra"
// Code that produces warnings...
#pragma clang diagnostic pop
And the code that is in the middle is now being silenced from warnings for good.
And in Gcc we also have similar pragma directives with clang and i thought i could try something like this:
#pragma GCC diagnostic push
#pramga GCC diagnostic ignored "-Wall"
#pragma GCC diagnostic ignored "-Wextra"
// Code that produces warnings...
#pramga GCC diagnostic pop
But passing -Wall and -Wextra in diagnostic ignored pragma in GCC does not work like clang, and does not disable all the possible warnings. Instead of this passing a specific warning to disable works:
#pragma GCC diagnostic push
#pragma GCC diagnostic ignored "-Wunused-parameter"
void foo (int x) // No longer getting "unused parameter 'x'" warning
{
}
#pragma GCC diagnostic pop
So the only workaround i can think so far is to make a long list with all the GCC warning flags and use them like above. Is there a more elegant solution? If not where i can get the complete Gcc warning flag list (favorably in a plain list)?
-I
, use-isystem
instead. Problem solved. :) Then you can turn your warnings back on (plus-Werror
!) to catch bugs you introduce into your code in future. Your code may be warning free now but the entire purpose of warnings is to catch when you break it in the future... – Lightness Races in Orbit-w
(specifically lowercase w;-W
does something else) that disables all warnings, but it does not seem to be accepted as an argument to#pragma GCC diagnostic
, sorry. – zwol