I’ve been using Author-It v5.3.207.6063 to generate help in Japanese. This seemed to work fine, but the resultant chm file’s Search tab wouldn’t work with Japanese glyphs (only with English characters).

Further experimentation has revealed a way of getting the Search tab to work. I had to do the following:

  1. Update the regional settings (XP) to enable Japanese and set it as the default for non-Unicode applications
    1. On the Languages tab, tick “Install files for East Asian languages”
    2. On the Advanced tab, set it to use Japanese as the default for non-Unicode applications
    3. Reboot
  2. Publish to HTML Help format as normal
  3. Edit the HTML files (just the numbered ones directly relating to help topics) in the Publishing folder using Notepad:
    1. Change the charset in the text from UTF-8 to Shift_JIS
    2. Save As and select ANSI format
  4. Recompile the chm using hhc.exe from the command line

After these additional steps, the Search tab works as expected (i.e. searching for Japanese glyphs produces results).

I think the problem is that hhc.exe, being an old program, can only index ANSI-formatted files for its search (and these, in turn, need to be specified as Shift_JIS to display correctly). However, Author-It produces UTF-8 encoded Unicode html files and then runs hhc on these.

Unless I’ve misunderstood something, I think ANYONE publishing Japanese content to HTML Help format will produce chm files in which the Search tab does not work! Therefore, it would be good if Author-It could be changed so that it automatically produces ANSI format HTML files with the charset set to Shift_JIS when publishing HTML Help files in Japanese.

I’ve raised a support call about this, but thought I’d share the workaround here for anyone else experiencing the same problem.