792116 - Saving mail to Drafts corrupts Czech diacritics

Status: RESOLVED WORKSFORME

(MailNews Core :: MIME, defect)

Description

[karel]

User Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:15.0) Gecko/20100101 Firefox/15.0.1
Build ID: 20120905151427

Steps to reproduce:

Wrote an e-mail in Czech. Closed it to be autosaved to Drafts.
example mail text: "líbily nápady"


Actual results:

Mail saved to Drafts folders has corrupted diacritics.
example mail text as changed like this: "lĂ­bily nĂĄpady"
happened several times but does not happen every time. obviously some prerequsites must be met. did not discover them.



Expected results:

Should be saved with the diacritics (probably character encoding) used when writing the mail.

Comment 1

[WADA:World Anti-bad-Duping Agency]

What is your "default mail composing character set" setting?
What chars used in Subject? 7bits ascii char only?

If default charset != utf-8 and 7bits-ascii only in Subject:, problem is perhaps following.
  If 7bits-ascii only in Subject:, auto change to utf-8 is not invoked.
Do you see your problem with "Subject: líbily nápady"?

Comment 2

[karel]

Attachment: char encoding settings

Comment 3

[karel]

Thx for replying, WADA.
I have attached the snapshot of my settings.
The subject of both the original mail and my reply was simple "Re:"
I am not sure about what the difference between utf-8 and 7bits-ascii is.
Tried to close&save to drafts a mail containing "ěščř" in the body and first "escr" and then "ěščř" in the subject and they both saved correctly.
Hope it helps.
Karel

Comment 4

[:aceman]

Try to use ISO-8859-2 as that is the proper encoding for Czech.

But I can't see the problem even with ISO-8859-1 on TB18.

I think the characters í and á are included also in ISO-8859-1.
So when using only the "líbily nápady" string, this encoding is used and the message displays fine. As soon as I add ť character, the message is saved as utf-8.

Can you see what encoding is used in the save draft? Use View source (Ctrl-U) command.

Comment 5

[karel]

I am not fluent in character sets, but I think it is ISO-8859-2.

from the message source:
This is a multi-part message in MIME format.
--Boundary_(ID_qRmbUMr/OLMhLdrnoCddRQ)
Content-type: text/plain; charset=ISO-8859-2; format=flowed
Content-transfer-encoding: 8BIT
/Souhlas, hlavnÄ prvnĂ­ zĂĄbÄr na Zemi./

Comment 6

[:aceman]

Now that is strange. The charset is set to ISO-8859-2 but the bytes are as if encoded in UTF-8.

Could you temporarily try a TB18 nightly version from here:https://ftp.mozilla.org/pub/mozilla.org/thunderbird/nightly/latest-comm-central/

Use a zip version so that you do not need to uninstall your main TB installation. Just extract it anywhere and run the thunderbird.exe from the extracted folder.

Comment 7

[karel]

Sorry, no time to spare :-(

Comment 8

[Gingerbread Man]

(In reply to :aceman from comment #6)
> Could you temporarily try a TB18 nightly version

I couldn't reproduce this in the latest Daily, but it's important to note it has broken character encoding menus: the drop-down menus have no items.

(In reply to :aceman from comment #4)
> But I can't see the problem even with ISO-8859-1 on TB18.

I also tried the latest Earlybird. The default outgoing encoding is now UTF-8, so right off the bat this wouldn't be an issue anymore. Still, I tried it both with the default settings and with ISO-8859-1. No problems, whether "líbily nápady" was the subject line, the message body, or both.

It seems there's ample reason to mark this worksforme.