Bug 151945

Version:            (using KDE KDE 3.5.8)
Installed from:    Ubuntu Packages
OS:                Linux

When typing japanese text into a multiline text field (such as Kmail composer, or a koqueror text field in a web page) with scim-anthy, the behavior is inconsitent.

At first, everything works fine. But after I go to a new line, the first word I try to convert goes wrong, but the following ones are ok again.

For instance, if I type "tesuto<space><enter><enter>tesuto<space>tesuto<space>
the expected result is

テスト
テストテスト

but I get

テスト
tてテストテスト

Also, if i delete the second line with backspace, when i remove the new line character and the cursor goes back to the first line, a テcharacter apears after the cursor, like this(the | marks the cursor) :
テスト|テ

Note that this problem does not happen in KWord (which probably uses its own input widget), nor in wengo (which uses qt, but is not a kde app).

My install of ubuntu is an update from 7.4 to 7.10.