To anyone who is looking at the GB 18030 xml mapping file in our charset repository at http://oss.software.ibm.com/cvs/icu/charset/data/xml/gb-18030-2000.xml We found - thanks to Carl Brown's questions - that we had checked in an incorrect file, different from the correct file that we had published elsewhere on our web site in February. The incorrect file was re-generated from our very ICU-specific .ucm mapping file, which means that that .xml file contained an incorrect, over-optimized validity table, and was missing the <range> elements that represent the part of the mappings that can be done algorithmically. The explicitly listed assignments were correct and complete. We have just fixed this. The file at the location mentioned above is updated. I am very sorry for this blunder. Sincerely, markus PS: The .ucm file itself is not incorrect, but does not contain all of the information. It is only complete with special GB 18030 code in the ICU library. For details please see http://www-106.ibm.com/developerworks/unicode/library/u-china.html
