[Xcb] Proportional fonts and text extents

[Peter Harris]

Jake LeMaster wrote:
> 
> I threw this together to compare extents results for whatever font you
> want to specify on a test string:
> 
> http://ircage.org/spewns/shiz/extents.c

Just so you are aware, xcb_query_text_extents is the equivalent of
XQueryTextExtents.

There is currently no equivalent to XTextExtents in xcb. The equivalent
function would use the results of an xcb_query_font to do the
calculations without needing to communicate with the server. If you do
write this function, we would like to add it to xcb/util. Thanks!

> When using the 2 proportional fonts I have in there, I get results of...
> 
> xcb extents returns a width of:  [265]
> Xlib extents returns a width of: [249]
> 
> xcb extents returns a width of:  [371]
> Xlib extents returns a width of: [300]
> 
> The difference can be significant, as you see. When I use the 2
> monospace fonts though, I get the results I'd expect...
> 
> xcb extents returns a width of:  [371]
> Xlib extents returns a width of: [371]
> 
> xcb extents returns a width of:  [318]
> Xlib extents returns a width of: [318]
> 
> Thoughts?

There's a couple of bugs in your implementation of build_chars.

First, you compare individual character codepoints against the desired
length of the string. I'm sure that's a typo. Second, byte1 is the
high-byte, and byte2 is the low-byte.

So your function should probably read something more like:

xcb_char2b_t *build_chars(const char *str, size_t length)
{
  int i;
  xcb_char2b_t *ret = malloc(length * sizeof(xcb_char2b_t));
  if (!ret)
    return NULL;

  for (i = 0; i < length; i++) {
    ret[i].byte1 = 0;
    ret[i].byte2 = str[i];
  }

  return ret;
}

Peter Harris
-- 
               Open Text Connectivity Solutions Group
Peter Harris                    http://www.opentext.com/connectivity
Research and Development        Phone: +1 905 762 6001
pharris at opentext.com            Toll Free: 1 877 359 4866