Re: jscript: Mark some fall-throughs in switch statements

[Alexandre Julliard]

Andrew Talbot <andrew.talbot(a)talbotville.com> writes:
> @@ -4037,6 +4037,7 @@ static HRESULT RegExpConstr_leftContext(script_ctx_t *ctx, vdisp_t *jsthis, WORD
>
> V_VT(retv) = VT_BSTR;
>          V_BSTR(retv) = ret;
> +        /* fall through */
>      }
>      case DISPATCH_PROPERTYPUT:
>          return S_OK;

I don't think that PROPERTYGET is supposed to fall through to
PROPERTYPUT.

-- 
Alexandre Julliard
julliard(a)winehq.org

jscript: Mark some fall-throughs in switch statements

[Bruno Jesus]

On Sat, Oct 8, 2011 at 14:58, Alexandre Julliard <julliard(a)winehq.org> wrote:
> Andrew Talbot <andrew.talbot(a)talbotville.com> writes:
> > @@ -4037,6 +4037,7 @@ static HRESULT RegExpConstr_leftContext(script_ctx_t *ctx, vdisp_t *jsthis, WORD
> >
> >          V_VT(retv) = VT_BSTR;
> >          V_BSTR(retv) = ret;
> > +        /* fall through */
> >      }
> >      case DISPATCH_PROPERTYPUT:
> >          return S_OK;
>
> I don't think that PROPERTYGET is supposed to fall through to
> PROPERTYPUT.

Isn't it that way just to save the use of an extra return S_OK?
Instead a break could be used too because that function returns S_OK
by default.

http://source.winehq.org/source/dlls/jscript/regexp.c#L4025

Best regards,
Bruno

jscript: Mark some fall-throughs in switch statements

[Andrew Talbot]

> Isn't it that way just to save the use of an extra return S_OK?
> Instead a break could be used too because that function returns S_OK
> by default.
>
> http://source.winehq.org/source/dlls/jscript/regexp.c#L4025
>
> Best regards,
> Bruno

A fall-through would indeed be lazy and not in the spirit of the fact that
DISPATCH_PROPERTYGET and DISPATCH_PROPERTYPUT are kind of the opposite of
each other. It either requires a break in both places (for consistency)
to the common "return S_OK" or two immediate such returns, instead.

-- 
Andy.