1357198 - Provide ability to enter a unicode character into subject in message body by specifying the code point

Status: UNCONFIRMED

(Thunderbird :: Message Compose Window, enhancement)

Description

[tlhackque]

User Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:52.0) Gecko/20100101 Firefox/52.0
Build ID: 20170323105023

Steps to reproduce:

I received an e-mail which contained a properly encoded U1F30F in the subject line.

I wanted to generate such a subject.

Click "create new message".  Format->Text Encoding is Unicode.


Actual results:

The subject of the original message displays properly in the read window (and in the folder's message list).

But, I wanted to insert a Unicode symbol in the subject of a complaint that forwarded the e-mail.

Strangely, I can't find a way to:

a) Insert an arbitrary Unicode codepoint into the subject when composing a message

b) Insert an arbitrary Unicode codepoint into the body when composing a message

In the body, Insert->Character and Symbols has a few limited symbols, but no way to enter an arbitrary codepoint.  In the subject, Insert is grayed-out. (But MIME 2047 encoding IS supported for display...)

I am aware of the ALT+numpad mechanism for entering Windows codes; obviously, it doesn't work for Unicode (hex).

However, if I enter a properly-encoded string in the subject, it is correctly displayed in drafts.  E.g. 
=?utf-8?B?bGHFrQ==?= will produce an accented string in the message list of a folder.

But it appears encoded if I open the draft.  In a received message, the Subject in the header includes the accented string.

So, display in a message list or message reading window decodes RFC2047-encoded subjects, but a compose window does not.

This seems inconsistent.


Expected results:

It should be possible (and relatively easy) to do (a) and (b).

I realize that it may be impractical to have an exhaustive list  of all unicode characters embedded in a dropdown (!) or to have a screen keyboard.

At a minimum, a mechanism to insert by codepoint, e.g. Insert->Char/sym with numeric codepoint option.  In both the Subject and Body of a message.

Perhaps also capture codepoints used by a TB user in his profile, so Insert->Char/sym could have a dropdown with that subset.  (E.g. I'm not likely to need Unicode Klingon - but some line drawing characters.  On the other hand, some people (I'm not one) are in love with Emoj - and there are many more in Unicode than those on the Compose dropdown.)

Typing an RFC2047 encoding into the subject line of a compose window should escape it - it's just text.

Conversely, there should be a mechanism to generate one that doesn't require an engineer's understanding to execute...

And if a message's headers DO contain an encoded subject, the Subject line in the Compose window should show the Unicode - not the =?UTF8?B? gibberish (er, encoding).

Comment 1

[Jorg K (CEST = GMT+2)]

So enhancement request for:
a) Insert an arbitrary Unicode codepoint into the subject when composing a message
b) Insert an arbitrary Unicode codepoint into the body when composing a message

Currently the only facility is "paste" or let users with a suitable keyboard type it. As a guess, that would cover 99.5% of the use cases.

Which other application lets you do this, for example? Notepad++?

BTW, entering =?utf-8?B?bGHFrQ==?= into the subject results in =?UTF-8?Q?la=c5=ad?= in the draft and that shows correctly as "laŭ" when you edit the draft, at least in TB 55 Daily and the latest TB 52 ESR release. But I agree, let's not teach users to enter RFC2047-encoded subjects ;-)

Comment 2

[tlhackque]

Paste / suitable keyboard doesn't cover all the cases.  I think 99.5% is too optimistic.

E.g. I have a US keyboard, but want to type Greek letters/Math symbols. Or musical notes.  Or (heaven forbid), an Emoj.  Emoji's being the trendy thing - what keyboards (excluding mobile devices) have them?

c) If someone DOES manage to enter what appears to be an RFC2047 string in the subject, "quote" it such that it appears as the sequence =? ...?=.  Not what it decodes to.  WYSIWYG!

Other apps - here are a few:

M$ Word: https://support.office.com/en-us/article/Insert-a-symbol-or-special-character-81e64967-74c0-4fd9-814a-3aa867d4cfce#bm3

LibreOffice: https://ask.libreoffice.org/en/question/36653/how-do-i-insert-special-characters-in-libreoffice-on-ubuntu-if-alt188-or-other-key-combinations-dont-work/

Windows (sometimes) http://www.fileformat.info/tip/microsoft/enter_unicode.htm - except that in applications (like TB) that have menu alt/accelerator keys using a-f, the accelerator function supersedes entry of the codepoint.

I'm not sure why the subject didn't display decoded on re-opening when I entered the bug.  If I find a reproducer for that, I'll add it here.  (But I wasn't imagining things..)

As for teaching users RFC2047 encoding - surely we jest :-)  (If that effort is successful, let's ditch C, C++ & friends and go back to coding in binary - with mechanical switches... :-):-)

♫♪♫♪♫♪

Comment 3

[tlhackque]

BTW, Wikipedia has a summary of several methods used for Unicode input: https://en.wikipedia.org/wiki/Unicode_input

(I was surprised to find the (type hex)ALT-x mechanism exist in Windows applications using the RichEdit control.  Postfix notation in an Office UI - certainly creative, though I wonder what SIGCHI would say...)