Re: dlls/cabinet: fix dead stores (llvm/clang)

[Dmitry Timoshkov]

"ricardo filipe" <ricardo_barbano(a)hotmail.com> wrote:
> --- a/dlls/cabinet/fci.c
> +++ b/dlls/cabinet/fci.c
> @@ -462,7 +462,7 @@ static cab_ULONG fci_get_checksum(const void *pv, UINT cb, CHECKSUM seed)
>      case 2:
>        ul |= (((ULONG)(*pb++)) <<  8);
>      case 1:
> -      ul |= *pb++;
> +      ul |= *pb;
>      default:
>        break;
>    }

Is there any reason that you ignored 2 similar cases above?

-- 
Dmitry.

dlls/cabinet: fix dead stores (llvm/clang)

[Marcus Meissner]

On Tue, Nov 18, 2008 at 03:11:52PM +0800, Dmitry Timoshkov wrote:
> "ricardo filipe" <ricardo_barbano(a)hotmail.com> wrote:
> > --- a/dlls/cabinet/fci.c
> > +++ b/dlls/cabinet/fci.c
> > @@ -462,7 +462,7 @@ static cab_ULONG fci_get_checksum(const void *pv, UINT cb, CHECKSUM seed)
> >      case 2:
> >        ul |= (((ULONG)(*pb++)) <<  8);
> >      case 1:
> > -      ul |= *pb++;
> > +      ul |= *pb;
> >      default:
> >        break;
> >    }
>
> Is there any reason that you ignored 2 similar cases above?

Yes, because it would be wrong ... Its a Duffs Device, falling through
to the last case. The other pb++ are necessary.

Ciao, Marcus

dlls/cabinet: fix dead stores (llvm/clang)

[ricardo filipe]

> Date: Tue, 18 Nov 2008 10:17:46 +0100> From: meissner(a)suse.de> To: dmitry(a)codeweavers.com> CC: ricardo_barbano(a)hotmail.com; wine-devel(a)winehq.org> Subject: Re: dlls/cabinet: fix dead stores (llvm/clang)> > On Tue, Nov 18, 2008 at 03:11:52PM +0800, Dmitry Timoshkov wrote:> > "ricardo filipe" <ricardo_barbano(a)hotmail.com> wrote:> > > > > --- a/dlls/cabinet/fci.c> > > +++ b/dlls/cabinet/fci.c> > > @@ -462,7 +462,7 @@ static cab_ULONG fci_get_checksum(const void *pv, UINT cb, CHECKSUM seed)> > > case 2:> > > ul |= (((ULONG)(*pb++)) << 8);> > > case 1:> > > - ul |= *pb++;> > > + ul |= *pb;> > > default:> > > break;> > > }> > > > Is there any reason that you ignored 2 similar cases above?> > Yes, because it would be wrong ... Its a Duffs Device, falling through> to the last case. The other pb++ are necessary.> > Ciao, Marcus

yes only the last one is dead. although i have to ask you guys, today i think i messed up on this one and it's a false positive, because at the beggining pb = pv.
so it will affect the pv and this shouldn't be applied right?
_________________________________________________________________
Conheça o Windows Live Spaces, a rede de relacionamentos do Messenger!
http://www.amigosdomessenger.com.br/