|
From: Scott M. <Sco...@sa...> - 2006-12-13 20:54:01
|
Is there a problem with 3.4 and ALTERNATE SHIFTED? I'm sorting with an = en_US locale, a STRENGTH of 4, and ALTERNATE SHIFTED and I'm getting = some strange results. I'm sorting a list of three strings: 2004 2010 23 and ICU 3.4 is returning the keys: 25 7E 25 82 01 06 01 06 01 26 25 7A 25 7A 26 00 25 7E 25 7C 01 06 01 06 01 26 25 7A 26 25 7A 00 25 7E 25 80 01 06 01 06 01 27 00 respectively and this results in an order of: 2010 23 2004 Without ALTERNATE SHIFTED, I get the following keys: 25 7E 25 7A 25 7A 25 82 01 08 01 08 01 26 00 25 7E 25 7A 25 7C 25 7A 01 08 01 08 01 26 00 25 7E 25 80 01 06 01 06 01 24 00 and this results in the order: 2004 2010 23 I checked against the demo page here: http://demo.icu-project.org/icu-bin/locexp?_=3Den_US&d_=3Den&x=3Dcol = =20 with and without the additional custom rule [alternate shifted] but this = is returning different keys, leading me to think that this demo page = might be using 3.6. Can anyone offer a quick sanity check here? It would be much = appreciated. Scott |