Certain unicode characters get erroneously replaced by the replacement character if they are either sent or copied from the input line, for example U+26D3. Characters from U+269D up to and including U+2700 are affected. U+2600 up to and including U+269C do not appear to be affected, neither is U+2701. To reproduce, one can use the "Insert->Special Character" menu to insert an affected character. The character is then displayed correctly until the message is being sent out or the character is cut out and pasted back into the message.
QChar returns QChar::Unicode_Unassigned for them, because they are added in newer Unicode versions. Qt 4.8 only supports up to Unicode 5.0, but http://en.wikibooks.org/wiki/Unicode/Character_reference/2000-2FFF says they are added in 5.1 or later.
When a character of Unicode 5.1 or higher is copied into the input field, it is also replaced by the replacement character. Those two bugs doesn’t appear in other KDE or Qt-only applications. That means Konversation had a specific bug that prevent users to use characters added to Unicode in 2008 and afterwards.
Lacking Unicode support is still an issue in Konversation 1.6. Meanwhile, Unicode 8.0 is out.
OK, ignore what I just posted for now, it seems to be Konversation is not doing character substitution. That should be a different bug.
Seems to be working with Konversation 1.6 (based on KDE Frameworks 5.12).
Still works with Konversation 1.6 with KDE Frameworks 5.19.0). We should probably close this bug report.