** Observed with 5/3/99 Win32 M5 candidate build ** There is a general editor bug for the Editor/Mail body text, Bug 5928. This is similar but the problem is in the Subject header field. Any high-bit characters input with the use of ALTGR+NumPad where the numbers come from (128-159) are sent without MIME encoding and also incorrectly. It seems to assign one same character for different characters, i.e. 0xC2 (Capitla A with Circumflex). Though the range (0x80 - 0x9F) is not proper iso-8859-1 range, we should be able to send these characters using this method, nonetheless. Maybe this problem will get resolved when the widget there is replaced by the Ender-based one. In that case, we can use this bug as a reminder to check on this problem.

I need to add that these characters are displayed correctly as you input them but are sent wrong.

Adding cata@netscape.com to cc. Cata, what is the expected result of the Latin1 encoder when we feed the unicode data not in Latin1 range?

As additional info, I found that when viewed with 4.6 Messenger, the thread pane shows only "Â" (Capital A w/ Circumflex), but viewed under 5.0 mail, the thread pane header shows: € (Capital A w/Circumflex + Euro symbol) However, the subject header in the View pane shows nothing, not even Â.

The widget needs to use Ender. Messenger is not getting unicode. For example, trade mark is 0x2122 in unicode while messenger getting 0x9900 (window-1252 encoded data put into nsString without conversion). This is a dup of 5061. *** This bug has been marked as a duplicate of 5061 ***

I'm going to mark this verified as a duplicate. However, when 5061 is resolved,we need to make sure that the issues described here are also taken care of.