"Guillaume SELLIER" <g_sellier@...> wrote on Mon, 07 Mar 2005
> I get a problem with my ht://Dig 3.1.6 with single quote. The problem
> occurs when we search a string containing a single quote, like "
> l'alarme ". In fact, htsearch encode special characters and change this
> string in " l'alarme ". The characters # is in the
> valid_punctuation attribute so it's deleted. The character ; is
> considered as a end of word, then htsearch cut the string in a boolean
> expression " l39 and alarme". And on the numbered links to other pages,
> the & char is encoded in HTML, and give "l&39..." in the url, then
> the boolean expression become more and more crazy with the & character
> as we click on numbers.
> I worked with the file bad_words for the moment (put l39, etc), but the
> reminder of searched string in the others result pages is ugly.
> Does someone had same problem and know how to resolve it ? Can we force
> htsearch to keep the real searched string and not encode it ?
> For information I use the attribute locale: fr_FR, and I'm on a RH ES 3.
Vous avez trois problèmes differentes - pardon, je continuerai en Anglais:
1) I am not sure why "'" is getting converted to "&39;" - is it not part
of the standard character set for the locale FR_fr?
2) Indexing of &entities; - I just don't have this problem. Search
http://www.londonfreelance.org for "Société"...
3) The display of &entities; - the last time I checked, the ht://Dig is
still waiting for a volunteer to put a massive amount of work into making
the whole thing work with Unicode. It seems very, very unlikely that there
will be a temporary fix before then.