Re: fixes for -Wmissing-declarations and -Wwrite-strings warnings

[Pouech Eric DMI AEI CAEN]

returning a pointer to something on the stack is something I call an error
(how do you ensure that the buffer will not get overwritten by another function call ?)
A+
> Message du 02/06/05 15:25
> De : "Dimi Paun" 
> A : [e-mail], [e-mail]
> Copie à : 
> Objet : Re: fixes for -Wmissing-declarations and -Wwrite-strings warnings
>
> From: "Pouech Eric DMI AEI CAEN"
> > looking at the patch, I saw something in the code that made me sick:
> > char *audioAutoDetect(void)
> > {
> > [snip]
> > char *driversFound[10];
> > [snip]
> > return driversFound[0];
> > [snip]
> > }
> >
> > can someone fix that broken code ?
>
> Yeah, it's not pretty, but just by looking at this snippet there is no
> obvious error in the above AFAICT.

fixes for -Wmissing-declarations and -Wwrite-strings warnings

[Francois Gouget]

On Thu, 2 Jun 2005, Pouech Eric DMI AEI CAEN wrote:
> returning a pointer to something on the stack is something I call an error
> (how do you ensure that the buffer will not get overwritten by another function call ?)

It's not returning a pointer to something on the stack. It is returning 
driversFound[0] which is a pointer to a string which may have been 
malloc()'ed (you stripped that part of the code).
> > > char *audioAutoDetect(void)
> > > {
> > > [snip]
> > > char *driversFound[10];
> > > [snip]

maybe driversFound[0]=malloc(something);
> > > return driversFound[0];
> > > [snip]
> > > }

fixes for -Wmissing-declarations and -Wwrite-strings warnings

[Dimi Paun]

> returning a pointer to something on the stack is something I call an error

Right, but it's not doing that :)