From: Richard W. <ric...@nt...> - 2014-04-16 23:01:02
|
On Wed, 16 Apr 2014 10:02:03 -0700 "Steven R. Loomis" <sr...@ic...> wrote: > > Of course, for Pali in Thai script the features needed are something > > like: > > > > brkitr/pi.txt: > > > > pi{ > > ... > > boundaries{ > > line:process(dependency){"line_pi.brk"} > > word:process(dependency){"word_pi.brk"} > > } > > } > > > > line_pi.txt and word_pi.txt: > > # In the Thai script, words are separated by spaces! > > $dictionary = [$dictionary - [:script = Thai:]] And I presume add word_pi.txt and line_pi.txt to BRK_SOURCE in brkfiles.mk and pi.txt to BRK_RES_SOURCE. But that doesn't seem to be enough. What else should I do to include these files in my ICU build? I am working with ICU Version 53 if that is relevant. > So, I was mostly trying to address the question of ICU assuming > language from one character, and whether locale specific dictionaries > would even be useful, hypothetically. As it stands that would need to > be a new feature request in any event. At least most of the syntax is present. Richard. |