Re: Question about sc/source/core/tool/interpr2.cxx and lcl_GetArabicValue

On 30/12/2012 18:24, Markus Mohrhard wrote:
> Hey,
>
> > Cppcheck reported this:
> > [sc/source/core/tool/interpr2.cxx:2656]: (style) The scope of the variable
> > 'bIsDec2' can be reduced
> >
> > Here are the lines:
> >     2651         while( bValid&&  (nCharIndex<  nCharCount) )
> >     2652         {
> >     2653             sal_uInt16 nDigit1 = 0;
> >     2654             sal_uInt16 nDigit2 = 0;
> >     2655             bool bIsDec1 = false;
> >     2656             bool bIsDec2 = false;
> >     2657             bValid = lcl_GetArabicValue( aRoman.GetChar( nCharIndex
> > ), nDigit1, bIsDec1 );
> >     2658             if( bValid&&  (nCharIndex + 1<  nCharCount) )
> >     2659                 bValid = lcl_GetArabicValue( aRoman.GetChar(
> > nCharIndex + 1 ), nDigit2, bIsDec2 );
> >
> > But in fact, on the contrary of "bIsDec1" , "bIsDec2" isn't used in the rest
> > of the loop.
> > So either, bIsDec2 is useless and can just be replaced by "false", or
> > something is wrong here.
> bIsDec2 can't be removed since the method requires a non const
> reference. However cppcheck is right that the scope of bIsDec2 can be
> reduced. You can move it into the the same scope as the call to
> lcl_GetArabicValue.
Ok, I've followed your piece of advice, here's the commit 
http://cgit.freedesktop.org/libreoffice/core/commit/?id=2fbd64cef870568da1479a03f6ea778bab68a09b

Thank you Markus.

Julien