Unicode characters destroyed by "ASCII trimming"

[anonymous]

Character semantics should be respected when trimming strings (e.g. feed title, post title), e.g. to enhance presentation. i.e. Avoid chopping a Unicode character in two.

This problem is most obvious with CJK feeds but affects many other languages.

Example:

On the index page, links presenting previous and next posts display trimmed item titles. The trailing Unicode character is often destroyed by byte-semantic trimming. The corrupted character is displayed as a dark diamond enclosing a question mark.

[anonymous]

Could you please provide a feed URL that exposes this problem?

When you do so, please add "dev-feedback" to the keywords list.

[anonymous]

Screenshot of Unicode corruption due to trimming

[anonymous]

Here's a Yahoo Taiwan news feed (UTF-8 Chinese) that should have a number of headlines of sufficient length to be trimmed. It may be necessary to browse through the per-feed item listing to see the effect.

http://tw.search.news.yahoo.com/rss?p=%AB%C8%BBy

[anonymous]

Changeset [1092] could fix this. Please let me know.

[sdcosta]

Haven't heard back from the poster. Please re-open if this is still a problem.