289892 - don't use UTF-8 as the default charset in vCard to improve compatibility with Outlook Express

Status: RESOLVED WONTFIX

(MailNews Core :: Backend, enhancement)

Description

[Gunnar Bolmsjö]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-GB; rv:1.7.6) Gecko/20050321 Firefox/1.0.2
Build Identifier: version 1.0 (20041206)

My name includes umlaut: (last name): Bolmsjö
(The last character is an "o" with two dots above it)
Now, the problem is that this character displays as follows in other mail
clients on PC platforms (I guess typically 'outlook'): Ã (a capital 'A' with a
tilde above it). 
It is of course correct in thunderbird, but that is not enough and most likely
an information about the encoding format of the characters is missing somewhere
in the system.

Reproducible: Always

Steps to Reproduce:
1. Fill in "Gunnar Bolmsjö" as the name of one reply card in thunderbird
2. send an e-mail 
3. open the mail in another type of mail client (outlook) 

Actual Results:  
The name gets coded wrong (the ö character)

Expected Results:  
typed it in the right way (as an ö)

a bit disturbing as I get postal mail written in this way.

Comment 1

[Jean-Marc Desperrier]

What do you call the 'reply card' exactly ? Is that the vCard ? I need to
understand where exactly you type 'Bolmsjö'.

What is your setting for 'use quotted printable MIME encoding' in
Options/Composition ?

Can you select one of the problematic mail in your 'sent' folder, save it as a
.eml file, and then put the file as an attachment to the bug ? (set the content
type to message/rfc822)

Comment 2

[Gunnar Bolmsjö]

Attachment: test mail containing my buisiness card

The problem is related to the v-card typed in from "account settings", and
"attach my vCard..." and editing this vCard. The box for 8 bit characters has
been tested as well (checked / unchecked).
If the vCard of the test mail is clicked on, outlook will open as default and
the "o" with an umlaut (two dots above) will display as "A" with a tilde above
and a paragraph sign after.

Comment 3

[Jean-Marc Desperrier]

OK, I did some test with Outlook Express. It's support of i18n in vCard is very
sloppy.

vCard are assumed to be in windows-1252, and the charset parametre is ignored.
It does not include a charset declaration in vCards it generates.
I don't know how it works in countries that use a different charset. I tried
changing the default charset, but it had no effect. I think it must depend on
CP_ACP, and I'd have to change it at the system level to modify the charset that
Outloook express assumes.

The net result for Thunderbird is that it's default setting of using UTF-8 for
vCard is problematic for interoperablity with Outlook Express (a quick check
shows that Outlook does not set the charset either in vCards). 
Even if it's really OE's fault for not been able to correctly support character
sets in vCard.

It would be better to use the charset that has been selected for the mail instead.

PS: Given the number of old vCard bug I find, it seem there is no ressource to
work on vCard support at the moment, don't hold your breath, reporter.

Comment 4

[Kena]

The problem lies in the fact that no vCard standard clearly defines what charset should be used when creating or reading vCards. There is no standard vCard directive to specify the charset: only the encoding (quoted-printable, base64 etc) can be specified.

Most webmail software and Outlook exchange vCards using iso-8859-1 or cp1252, which make them (sort of) cross-compatible. 

Thunderbird generates utf-8 inside the vcard text, regardless of the charset of the mail body, and attaches the vcard with a charset=utf-8 specification in the MIME *headers*. Unfortunately other software does not look at the MIME headers after the vcard has been saved to a file.

Besides there is no configurable option in Thunderbird to change the charset for vcards.

Comment 5

[Magnus Melin [:mkmelin]]

And what does outlook do if the name doesn't fit iso-8859-1?

Comment 6

[Óvári]

Can this enhancement bug be closed as “UTF-8 is now the only possible character set” based on “RFC 6350 vCard Format Specification”
A.1.  New Structure
   o  UTF-8 is now the only possible character set.
https://tools.ietf.org/html/rfc6350#appendix-A.1

Please note that RFC 6350 has been updated by RFC 6868
https://tools.ietf.org/html/rfc6868


A solution to your challenge is:
1. Export your Outlook Express Address Book contacts to a CSV file
http://www.topproducer.com/campus/helpfiles/8i/a_Tasks/Preferences/data_conversion/exporting_microsoft_outlook_express_6.0_data_to_a_.csv_file.htm
2. Open CSV file with LibreOffice with the iso-8859-1 character format
https://www.libreoffice.org/
3. Save CSV file using the UTF-8 character format
4. Import your contacts into the Thunderbird add-on called “CardBook”
https://addons.mozilla.org/thunderbird/addon/cardbook/


Thank you

Comment 7

[Magnus Melin [:mkmelin]]

Yes, OE doesn't exist anymore, and I don't think we want to change this in the wrong direction according to newest relevant RFC.