User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:220.127.116.11) Gecko/20061201 Firefox/18.104.22.168 (Ubuntu-feisty) Build Identifier: Thunderbird 22.214.171.124 (20080724) Thunderbird displays an unicode replacement character in front of each e-mail-header which has charset iso-2022-kr. Reproducible: Always Steps to Reproduce: 1. Compose an email with following To-Header: =?iso-2022-kr?q?=1B=24=29C=0E "some korean characters" =0F?= <firstname.lastname@example.org> 2. View that mail in Thunderbird Actual Results: There will be an additional unicode replacement character in front of the korean text Expected Results: Don't show that character In intl/uconv/ucvko/nsISO2022KRToUnicode.cpp at lines 107 and 132 there is a check if mRunLength == 0 (Length of non-ASCII run). The unicode replacement character is added if the decoder finds the escape sequence (ESC "$" ")" "C") in front of a line (where it should be, as rfc1557 states).
Created attachment 333167 [details] [diff] [review] Patch
Comment on attachment 333167 [details] [diff] [review] Patch How to prevent a XSS attack like bug 381412? Indeed, the sequence ESC "$" ")" "C" can occur any number of times per RFC 1557, I'm hesitating to allow appearing the designator twice (or more). BTW I'm not a intl owner/peer. Is jshin inactive now?
Created attachment 333250 [details] [diff] [review] Improved patch What do you think of this approach, only allowing ESC $)C at the beginning of a line? I am a little hesitant because I don't have any examples of actual ISO-2022-KR mails to test against, and I don't know how well existing clients that send ISO-2022-KR mails conform to RFC 1557. Christoph, maybe you can attach some samples or forward them to me?
This would comply to RFC 1557 and works for me. I use gmime to create mails which uses iconv to convert from the internal utf-8 representation to a best fitting charset (my characters are japanese, but gmime encodes to iso-2022-kr). I don't have another sample and i don't know how wide spread this charset really is (this bug would have been found before).
Comment on attachment 333250 [details] [diff] [review] Improved patch Sorry for the late response. I think it's a reasonable compromise. ISO-2022-KR is rarely used in the real world.
Pushed as 17110:b17a3d22447e.
Comment on attachment 333250 [details] [diff] [review] Improved patch This is a regression from bug 381412, so we need the fix on the branches where that went in. The testcase includes a test for the regression and for issues similar to those in bug 381412.
Comment on attachment 333250 [details] [diff] [review] Improved patch Approved for 126.96.36.199 and 188.8.131.52, a=dveditz for release-drivers.