[ros-diffs] [sir_richard] 45223

[Dmitry Gorbachev]

> Since __builtin_unreachable is only for GCC 4.5, we use
> exit(0) for now (it will never be reached) to trick GCC.

There is also __builtin_trap() …

[Ros Arm]

> > Since __builtin_unreachable is only for GCC 4.5, we use
> > exit(0) for now (it will never be reached) to trick GCC.
>
> There is also __builtin_trap() …
>
> Ros-dev mailing list
> [e-mail]
> http://www.reactos.org/mailman/listinfo/ros-dev

Aha, cool!

What seems to be the problem that caused the need for a revert? Can't we simply define NORETURN as attribute(noreturn) in nt+windef.h and then leave the built_intrap/unreachable hack in trap_x.h?

-r

[Dmitry Gorbachev]

It produces many warnings "warning: 'noreturn' function does return"
(which are treated as errors). Probably GCC 4.1.3 does not recognize
that exit() hack or __builtin_trap() never returns. Unfortunately, I
deleted old GCC a long time ago and now can't check how it behaves.

[Sylvain Petreolle]

We have to switch to gcc 4.4 and this will be resolved.
RosBE-Windows is ready and RosBE-Unix has a release candidate.
Is there still something missing to switch ?

Kind regards,
Sylvain Petreolle

----- Message d'origine ----
> De : Dmitry Gorbachev [e-mail]
> À : ReactOS Development List [e-mail]
> Envoyé le : Dim 24 Janvier 2010, 4 h 01 min 19 s
> Objet : Re: [ros-dev] [ros-diffs] [sir_richard] 45223
>
> It produces many warnings "warning: 'noreturn' function does return"
> (which are treated as errors). Probably GCC 4.1.3 does not recognize
> that exit() hack or __builtin_trap() never returns. Unfortunately, I
> deleted old GCC a long time ago and now can't check how it behaves.
>
> Ros-dev mailing list
> [e-mail]
> http://www.reactos.org/mailman/listinfo/ros-dev

[Daniel Reimer]

Yes, it seems like the windows build of GCC 4.4.X do not build with 
makex. Have to discuss that with Colin when he's back.

Sylvain Petreolle schrieb:
> We have to switch to gcc 4.4 and this will be resolved.
> RosBE-Windows is ready and RosBE-Unix has a release candidate.
> Is there still something missing to switch ?
>
> Kind regards,
> Sylvain Petreolle
>
> ----- Message d'origine ----
> > De : Dmitry [e-mail]
> > À : ReactOS Development [e-mail]
> > Envoyé le : Dim 24 Janvier 2010, 4 h 01 min 19 s
> > Objet : Re: [ros-dev] [ros-diffs] [sir_richard] 45223
> >
> > It produces many warnings "warning: 'noreturn' function does return"
> > (which are treated as errors). Probably GCC 4.1.3 does not recognize
> > that exit() hack or __builtin_trap() never returns. Unfortunately, I
> > deleted old GCC a long time ago and now can't check how it behaves.
> >
> > Ros-dev mailing list
> > [e-mail]
> > http://www.reactos.org/mailman/listinfo/ros-dev
>
> Ros-dev mailing list
> [e-mail]
> http://www.reactos.org/mailman/listinfo/ros-dev

[Dmitry Gorbachev]

Dmitry Gorbachev wrote:
> Probably GCC 4.1.3 does not recognize
> that exit() hack or __builtin_trap() never returns.

And this turned out to be a false and defamatory statement.