unicode and roller

Posted on Oct 1, 2004

Japanese is broken again at Planet Sun. The fix that I made previously was wrong. It made blogs.sun.com feeds that (supposedly) included unicode characters work properly, but broke some of those that didn’t. After a brief email exchange with Dave Johnson it turns out that the version of Roller running on blogs.sun.com is known to be broken in this respect. Each of the (say) three 8 bit characters that make up the UTF-8 representation of a Japanese character are included in the feeds, but they are entity encoded, which is wrong. This bug is supposed to be fixed in a newer version of Roller that Dave and co. are hoping to roll out on blogs.sun.com soon. Until then, sorry - articles that include unicode characters are just broken.