Open Bug 271508 Opened 20 years ago Updated 3 years ago

thunderbird forces the FROM and the BODY charset to be the same

Categories

(MailNews Core :: Composition, enhancement)

x86
All
enhancement

Tracking

(Not tracked)

People

(Reporter: u162355, Unassigned)

References

Details

(Keywords: intl)

User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.5) Gecko/20041118 Firefox/1.0 Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.5) Gecko/20041118 Firefox/1.0 my "Your Name" is Gábor Farkas. when i send a mail, it usually becomes: From: =?UTF-8?B?R8OhYm9yIEZhcmthcw==?= <gabor@z10n.net> that's ok. but, when i want to send a japanese mail..or better to say a mail in ISO-2022-JP encoding, thunderbird refuses to send it, and complains that "The message you composed contains characters not found in the selected character encoding" in a way he is correct, because my from-name (Gábor Farkas) contains "á", which is not available in ISO-2022-JP. but that should not be a problem. the FROM header does not have to have the same encoding as the body. an exampel from rfc2047: " From: Nathaniel Borenstein <nsb@thumper.bellcore.com> (=?iso-8859-8?b?7eXs+SDv4SDp7Oj08A==?=) To: Greg Vaudreuil <gvaudre@NRI.Reston.VA.US>, Ned Freed <ned@innosoft.com>, Keith Moore <moore@cs.utk.edu> Subject: Test of new header generator MIME-Version: 1.0 Content-type: text/plain; charset=ISO-8859-1 " as you see, the FROM is in iso-8859-8, and the body is in iso-8859-1. thunderbird should just use for example UTF-8 for the FROM and whatever encoding was specified for the body. or at least allow me to have the FROM in a different encoding. btw. if i change my YourName to not have accented characters, then it works. Reproducible: Always Steps to Reproduce: 1. specify your name (the "Your Name" field in the settings) to have an "á" 2. compose a mail with only ascii characters (even an empty mail is fine) 3. set the character encoding for the mail to ISO-2022-JP 4. try to send it Actual Results: thunderbird refused to send it with a dialog saying: "The message you composed contains characters not found in the selected character encoding" Expected Results: either: simply use a suitable charset for the from-header, and use the specified charset for the body or: the same as the previous one, but first display a warning.
There's a very old related bug where Mozilla Mail/TB will not handle properly a message where the encoding of the subject is not the same as the one of the body (they will display properly, but get it wrong when you try to reply to the message). It's not great to use one encoding for the header and another for the body, because the recipient has to be able to decode both to handle properly the message => That makes all UTF-8 better. Of course, what makes the problem worse is the fact the dialog doesn't say what character is the culprit and the user might not suspect it's his from and not anything in the body (and I don't know if we have an open bug on that).
i understand that it's better to use utf-8 everywhere. and i do use utf-8 everywhere. ...where i can. this problem actually happened when posting to the sci.lang.japan newsgroup, where the only-allowed encoding is iso-2022-jp. so everytime when i want to post to that newsgroup i have to change my name from "Gábor Farkas" to "Gabor Farkas". very annoying :) hmmm..now that i'm thinking about it.. as a workaround i could set up my name directly in that header-encoded for. so i could set up my name as "=?UTF-8?B?R8OhYm9yIEZhcmthcw==?=". ugly as hell, but theoretically it could work. will have to check it out ;)
yes, it works ;) but as i wrote, it's ugly. is this thing so hard to be fixed? maybe the problem is that the thunderbird source code is written in such a way that it's only able to cope with one encoding per mail?
Jungshik can probably give a definitive answer here..
Reproduced with TB 1.0, Win2K.
Status: UNCONFIRMED → NEW
Ever confirmed: true
OS: Linux → All
Hm. With TB 1.0+0302, which has the new "Send Anyway" feature when you're warned about characters not matching, the From is *still* tagged (per RFC-2047) as the body's encoding. I have an old identity I used from testing another bug that has a real name of "bokförlaget". I tried using that identity send an ISO-2022-JP message (text copied from some spam). The From: was sent as: From: =?ISO-2022-JP?B?Ym9rZm8icmxhZ2V0?= <mcow@well.com> which converts to (as ASCII): From: bokfo"rlaget <mcow@well.com> (That double-quote shouldn't be a problem, but because Mozilla mishandles 2047 text, it results in the display of the name with an additional, closing double- quote character.)
The original report is not a bug per se. at most 'enhancement' request if not WONTFIX. Mike's comment seems to be about the display problem, which needs to be filed separately.
Severity: normal → enhancement
Component: Message Compose Window → MailNews: Composition
Keywords: intl
Product: Thunderbird → Core
Version: unspecified → Trunk
excuse me, but why isn't this a bug? there can be 2 cases: 1. thunderbird does not handle non-ascii FROM headers. -in this case thunderbird should DISALLOW the setting of non-ascii from-addresses 2. thunderbird does handle non-ascii FROM headers. -in this case, thunderbird does not implement it correctly. it is a bug. imho this is case #2.
(In reply to comment #8) > excuse me, but why isn't this a bug? > 2. thunderbird does handle non-ascii FROM headers. > -in this case, thunderbird does not implement it correctly. it is a bug. Name me just one mail program that allows you to use different MIME charsets in the message header and the message body when you compose a *fresh new* message. Anyway, what do we do incorrectly? Not allowing you to use different mime charsets for message header and message body is NOT doing it incorrectly. It's a feature that would be nice if we offer (which is why it's categorized as 'enhancement'), but its priority is very low.
(intro: just to make sure everyone understands..thunderbird-is-a-great-program, i-love-it-and-use-it, only-reporting-here-to-make-it-better and all these prefixes apply ;) >Name me just one mail program that allows you to use different MIME charsets >in the message header and the message body when you compose a *fresh new* >message. does this mean that if every other mail client sux, tbird should suck too? >Anyway, what do we do incorrectly? Not allowing you to use different >mime charsets for message header and message body is NOT doing it incorrectly. afaik we are talking about rfc2047 here.. if you check my original report, i showed an excerpt from the rfc: " From: Nathaniel Borenstein <nsb@thumper.bellcore.com> (=?iso-8859-8?b?7eXs+SDv4SDp7Oj08A==?=) To: Greg Vaudreuil <gvaudre@NRI.Reston.VA.US>, Ned Freed <ned@innosoft.com>, Keith Moore <moore@cs.utk.edu> Subject: Test of new header generator MIME-Version: 1.0 Content-type: text/plain; charset=ISO-8859-1 " as you see, the FROM is in iso-8859-8, and the body is in iso-8859-1. in my interpretation you do not support this part of the rfc. so your implementation is wrong. you are doing something wrong. i don't know if thunderbird is really unable to send a mail with differently encoded parts, or it's just a "gui" problem.. and let me please return to the first sentence.. >Name me just one mail program that allows you to use different MIME charsets >in the message header and the message body when you compose a *fresh new* >message. i think i clearly demonstrated why is it needed for me to send a *fresh new* mail with different charset in the body and in the headers... summary: 1. thunderbird does not behave the way he should (by rfc2047) 2. i demonstrated that it is not just a theoretical question, because i really need it when using thunderbird. your turn :)
> 1. thunderbird does not behave the way he should (by rfc2047) I know RFC 2047 from the top to the bottom. You don't.
>> 1. thunderbird does not behave the way he should (by rfc2047) >I know RFC 2047 from the top to the bottom. You don't. you're right. i did not read the whole rfc2047..well..maybe i did but i don't remember exactly.. so please tell me why is my interpretation (that it IS ALLOWED to have a different encoding in the FROM-header and in the body) wrong. i do not want to fight here.. i only want to be able to send mails... so please explain to me why do you think that thunderbird's behaviour is correct. thanks
Jungshik, what about the conversion in the header of the o-with-umlaut to ASCII-o plus double-quote (ö => o") when Send Anyway is selected? Is that considered correct? It seems more than just a display issue -- altho I agree, it's not *this* bug. Gábor, you've discovered a workaround already -- enter the 2047/UTF-8 string as the identity's name. But if the group you're posting to only allows 2022-JP, that implies that at least some of the people reading those messages will be using clients that cannot display the problem character anyway. What you want seems to be in conflict not with Mozilla but with the group's self-imposed refusal to use Unicode. How about simply using "G. Farkas" as your name for that group?
I'm sorry I shouldn't have pressed 'commit' button so rashily. I was annoyed a little bit too much by your assertion that we're not compliant to RFC 2047. Nonetheless, I should have given you more detailed explanation. Your interpretation that it is ALLOWED to .... is absolutely correct. (that's one of intents of RFC 2047 authors) Therefore, we handle those cases very well when rendering emails received. (MS OE doesn't as you already know.) However, your assertion that Mozilla-mail is NOT compliant to RFC 2047 by not allowing you to have different MIME charsets in the message header and the message body is a different story. It'd be nice to offer that feature/enhancement to advanced multilingual users like you. However, not doing so doesn't make us incompliant to RFC 2047. Given this, I don't regard this as a high priority item. It may not as complicated as I initially thought to implement this. Still, it's not high in my queue. Besides, as Mike pointed out, in your scenario, even if we offer that feature, your recipient is not likely to get your name rendered correctly. (if their mail client is so good at RFC 2047 handling as we're, it'd be able to handle UTF-8 as well.) As Mike found out, if you just go ahead with 'Send anyway', you get your name transliterated. That is, we have a second line of 'defense' (yes, Mike, that's what we intended). However, Mike also found some problems with '"' when letters like o-umlaut and u-umlaut are transliterated. I have to take a look at it perhaps in another bug.
*** Bug 310514 has been marked as a duplicate of this bug. ***
*** Bug 312540 has been marked as a duplicate of this bug. ***
Assignee: mscott → nobody
QA Contact: composition
Product: Core → MailNews Core
Severity: normal → S3
You need to log in before you can comment on or make changes to this bug.