>>>>> "Andre" == Andre Poenitz <[EMAIL PROTECTED]> writes:
Andre> On Mon, Nov 15, 2004 at 12:10:40PM +0100, Jean-Marc Lasgouttes
Andre> wrote:
>> So, we have two different definitions of what a word is, one being
>> by inclusion and the other by exclusion. I am not sure what is the
>> best way to define a word. Note that the isp_esc_char stuff should
>> move to ControlSpellchecker. Then we would be in postion to unify
>> these two definitions of a word. But how should we do it?
Andre> I think it doesn't matter much how it is done as long as it is
Andre> done uniformly. If in doubt I'd prefer the 'inclusion' method,
Andre> but that's most likely a personal preference about which I
Andre> don't care too much.
The inclusion method now does:
/// return true if a char is alphabetical (including accented chars)
inline
bool IsLetterChar(unsigned char c)
{
return (c >= 'A' && c <= 'Z')
|| (c >= 'a' && c <= 'z')
|| (c >= 192); // in iso-8859-x these are accented chars
}
We can consider that this is good enough for now... Of course, we
should be able to define word-chars for each charset. I suspect that
locale has all this information for us, but we are not sure that we
use the locale that corresponds to the document's language.
JMarc
