Control: tags -1 +confirmed On Mon, Dec 17, 2007 at 01:58:14PM +0100, Thomas Viehmann wrote: > a few years ago, you submitted bug 108702[1] about searching > lists.debian.org in Japanese. > We have recently switched the search engine[2] and I think that part > of the problem is solved (if only by character handling getting > better in email clients and browsers).
Everything should be getting normalised to UTF-8 now, so differences in encodings used shouldn't be an issue. > I would much appreciate if > you could help me assess and ideally provide examples for what works > and what is still missing. In particular, I don't know whether the > notion of words the indexing uses fits well with how you would > expect it to work. The notion of what is a word is an issue still, as the tokenisation currently assumes words are separated by whitespace. So you can get some search results, but only where the word stands alone: https://lists.debian.org/cgi-bin/search?P=%E7%99%BA%E9%9F%B3 Cheers, Olly
