Re: [Qemu-devel] [6023] Use a hex value instead of possibly ambiguous 8 bit character

[Blue Swirl]

On 12/14/08, Johannes Schindelin <address@hidden> wrote:
> Hi,
>
>
>  On Sun, 14 Dec 2008, Blue Swirl wrote:
>
>  > Revision: 6023
>  >           http://svn.sv.gnu.org/viewvc/?view=rev&root=qemu&revision=6023
>  > Author:   blueswir1
>  > Date:     2008-12-14 09:30:41 +0000 (Sun, 14 Dec 2008)
>  >
>  > Log Message:
>  > -----------
>  > Use a hex value instead of possibly ambiguous 8 bit character
>
>
> /me is curious: how could buffer[j] = '\xb0' be ambiguous when buffer is
>  of type char *?  It's not as if C did UTF-8 conversion with chars.

The diff does not show it properly, there was a 8 bit character
between the apostrophes, not \xb0. One day some compiler might want to
parse the source text as UTF-8, then byte B0 and apostrophe after it
could decode to something different with mysterious side effects. 0xb0
will not ever cause these problems, '\xb0' could work too.

>  Besides...
>
>
>  > @@ -1249,7 +1249,7 @@
>  >       unsigned char* c=(unsigned char*)direntry;
>  >       int i;
>  >       for(i=1;i<11 && c[i] && c[i]!=0xff;i+=2)
>  > -#define ADD_CHAR(c) {buffer[j] = (c); if (buffer[j] < ' ') buffer[j] = 
> '\xB0'; j++;}
>  > +#define ADD_CHAR(c) {buffer[j] = (c); if (buffer[j] < ' ') buffer[j] = 
> 0xb0; j++;}
>
>
> in the meantime I think it would be more readable as
>
>  #define ADD_CHAR(c) buffer[j++] = (c) < ' ' ? '\xb0' : 'c';
>
>  Note that
>
>  - this code is only ever reached when DEBUG is defined, and
>
>  - this code still assumes that your terminal is ISO-8859-1, which is
>   typically wrong these days (UTF-8 is the de-facto standard).

Patches welcome :)