28865 - Browser: Page Send - Main body charset is incorrectly set when charset of the page does not match default send charset

Status: VERIFIED DUPLICATE of bug 28908

(MailNews Core :: Internationalization, defect, P3)

Description

[Katsuhiko Momoi]

** Observed with 2/22/2000 Win32 build **

The above page has a meta tag = Windows-1251.
This is not in our send charset menu yet.

1. Set the Mail send charset to ISO-8859-1.
2. Go to the above page.
3. Check the View | Character Coding menu. There is a checkmark
   against Cyrillic (Windows-1251).
4. Now engage "File | Send Page" menu.
5. Check the charset menu -- it says "ISO-8859-1" instead of
   no check mark. And the mail goes out in ISO-8859-1, too.

This should be fixed for Beta 1 since we are not fixing Bug 28131
for Beta 1. There really is no way to satisfy Windows-1251 users
if the menu works this way.

If the encoding of the documented is not on the charset menu
but is one that we have support for, then we should set to
that encoding and not use the default send charset.
This is what we were going to tell the Cyrillic users to
get around the non-existence of Windows-1251 menu item.

I don't know who this should go to, but probably nhotta.

Comment 1

[Katsuhiko Momoi]

This functionality needs to be working for any encoding
not in the current send charset menu. This is the only way
a workaround for Bug 28131 would make sense.
Nominating for Beta 1.

Comment 2

[nhottanscp]

The menu issue to be resolved by 
25372 [FEATURE] Dynamic Charset Menu for Mail compose 

Those charset not listed in the menu to be added to the cached list (most 
recently used charsets).

Comment 3

[Katsuhiko Momoi]

I'm removing Beta 1 designation from this bug upon consultation
with nhotta. What we should deal with ij this bug is the following:

1. We should set the main body charset to match the charset of
   the page which is being sent by Browser's Send Page menu.
   This is because typically the title of browser pages are
   in the encoding of document. If we know the encoding via the
   meta tag or HTTP charset, we should use that. Otherwise
   the Subect which is normally the title of the page
   will be incorrectly encoded using the default charset.

I correcetd the summary to reflect the nature of this bug.
The menu issue will be resolved in the menu bugs.

Comment 4

[nhottanscp]

Adding rhp to cc.
We are currently using HTTP or META charset of attachment to create a charset 
label of the part but not main body. The change is to use it for the main body 
too only for send web page. Rich, is that possible?

Comment 5

[rhp (gone)]

Is this a good idea? What if the user enters text in the body and its not 
necessarily the same as the attachment.

- rhp

Comment 6

[nhottanscp]

The same issue for forward as attachment and send attachment which we do not 
change the main body charset (in fact, cannot if multiple attachments with 
multiple charsets).
What special about send web page is the page title is used as a subject. So I 
think this is the same category as reply and forward inline.

Comment 7

[Katsuhiko Momoi]

There are both good and bad in my request.
The bad is what you pointed out. We are sending all
new mail msgs with the default mail charset and probably
the Page send should not deviate from this.

The good is that we won't mangle the Subject header's
charset if we initialize the main body charset to the
page charset. 

Maybe there is a way to somehow deal with the subject charset
when 1) it contains anything other than ASCII and 2) when the
charset of the page does not match your send default. If there
is a good way to deal with this problem, we don't have to
do it the way this bug request suggests.

Comment 8

[nhottanscp]

We can use the web page's charset for subject header only without changing the 
main body charset. But this is confusing (we have no UI to change header and 
body charset separately) also some mailer won't understand it.

Comment 9

[Katsuhiko Momoi]

Well, the other way is a "warning" msg saying that you're about
to send out a header which contains characters not in the 
chosen encoding.

Comment 10

[nhottanscp]

Okay, I think it is not usual that the user to send a page which does not match 
with the default send charset. And the user has the chance to change it by menu. 
Plus there is going to be a feedback 
(20249 [FEATURE] feedback for mail charset when composing).
I mark this as WONTFIX.

Comment 11

[Katsuhiko Momoi]

OE5 does issue a warning when you include characters not coneverd
by the chosen encoding in the body text to 1) change the 
charset to something else or 2) suggest to send in UTF-8.

I would like to suggest that we do something similar for both headers and
body. Feedback UI has not been resolved and in any case we need
warnings in some cases. 
I propose to combine this with the case of other warning. 
Is there a bug for that one?

Comment 12

[Katsuhiko Momoi]

Since I filed a new bug to address teh warning issue,
I'm going to amrk this bug a duplicate of that bug instead of
WontFix.

Comment 13

[Katsuhiko Momoi]

Resolving this as a duplicate of Bug 28908.

*** This bug has been marked as a duplicate of 28908 ***

Comment 14

[Katsuhiko Momoi]

Verified as a duplicate.