311448 - Bugmail headers don't handle utf8 properly

Status: RESOLVED WORKSFORME

(Bugzilla :: Email Notifications, defect)

Description

[Joel Peshkin]

When there are 8-bit characters in a bugmail subject line, it should be passed
through Encode::MIME::Header

Comment 1

[Joel Peshkin]

Actually, it is not necessary to fix templetize first.   The following example
shows that it is possible to enocde only part of a subject line.  That means
this should be possible when performing the substitutions.

Subject: This is a subject with =?UTF-8?B?16nXnNed?=

The only complication is the rules for the header and body are differrent, so
this needs to be a bit more that just a tweak to the substs function.

Comment 2

[u197037]

Just 2 cents:
In order to suit needs of some [m.b. old] broken mailers (OE) the encoding of
headers should match encoding of message body.
At least that apply to situation, when subject is NOT base64 encoded but raw
8-bit (RFC violation, but commonly used)
To not violate RFC(some number), it is a good idea is to pass each header row
through encoder, not only Subject: one, since 8bit characters are not alowed in
header.

So, to ensure readability of both parts, encoding should match.
Also, Yahoo! WebMail had big problems with i8n, but it's not our fault :), just
be aware it handled that things not properly and probably should not be used to
test this feature.

Comment 3

[Max Kanat-Alexander]

We've done this for some time with encode_qp_words. It should work even better with bug 353711.