Re: Wininet Buffer Length Fixes

[Alexandre Julliard]

Robert Shearman <rob(a)codeweavers.com> writes:
> --- wine/dlls/wininet/http.c	4 Jul 2004 00:24:47 -0000	1.65
> +++ wine/dlls/wininet/http.c	13 Jul 2004 16:23:11 -0000
> @@ -1131,8 +1186,8 @@
>      if( result )
>      {
>          len = WideCharToMultiByte( CP_ACP,0, bufferW, len / sizeof(WCHAR),
> -                                     lpBuffer, *lpdwBufferLength, NULL, NULL );
> -        *lpdwBufferLength = len * sizeof(WCHAR);
> +                                     lpBuffer, *lpdwBufferLength+1, NULL, NULL );
> +        *lpdwBufferLength = (len-1) * sizeof(CHAR);

It seems to me this would potentially write beyond the end of the
buffer, that doesn't look right.

-- 
Alexandre Julliard
julliard(a)winehq.org

Wininet Buffer Length Fixes

[Robert Shearman]

Alexandre Julliard wrote:
> Robert Shearman <rob(a)codeweavers.com> writes:
> > --- wine/dlls/wininet/http.c	4 Jul 2004 00:24:47 -0000	1.65
> > +++ wine/dlls/wininet/http.c	13 Jul 2004 16:23:11 -0000
> > @@ -1131,8 +1186,8 @@
> >     if( result )
> >     {
> >         len = WideCharToMultiByte( CP_ACP,0, bufferW, len / sizeof(WCHAR),
> > -                                     lpBuffer, *lpdwBufferLength, NULL, NULL );
> > -        *lpdwBufferLength = len * sizeof(WCHAR);
> > +                                     lpBuffer, *lpdwBufferLength+1, NULL, NULL );
> > +        *lpdwBufferLength = (len-1) * sizeof(CHAR);
>
> It seems to me this would potentially write beyond the end of the
> buffer, that doesn't look right.

From:
http://msdn.microsoft.com/library/default.asp?url=/library/en-us/wininet/win...

In the case of a string, the byte count does not include the string's 
terminating null character.

Since WideCharToMultiByte returns the number of bytes written including 
the null terminator it is required to take one away from it, although 
the misleading *sizeof(CHAR) can be removed.

Rob

Wininet Buffer Length Fixes

[Alexandre Julliard]

Robert Shearman <rob(a)codeweavers.com> writes:
> In the case of a string, the byte count does not include the string's
> terminating null character.
>
> Since WideCharToMultiByte returns the number of bytes written
> including the null terminator it is required to take one away from it,
> although the misleading *sizeof(CHAR) can be removed.

I meant the line above, where you use *lpdwBufferLength+1, this will
overflow the buffer. Note that since len doesn't include the
terminating NULL, converting len/sizeof(WCHAR) characters won't get
you a terminating NULL even with a larger dest buffer.

-- 
Alexandre Julliard
julliard(a)winehq.org