245693 - Unicode->EUC-KR converter loses data on encountering characters not representable in 2bytes

Status: RESOLVED FIXED

(Core :: Internationalization, defect)

Description

[Jungshik Shin]

EUC-KR is a Korean multibyte encoding in which a character can be represented in
1byte(US-ASCII), 2bytes(characters in KS X 1001) and 8bytes(8,822 Hangul
syllables not enumerated in KS X 1001). Mozilla has supported this encoding
rather well, but  it's recently found that Unicode->EUC-KR converter (encoder)
truncates the content on coming across a character not representable by either 1
or 2bytes (that is, a character that can be only represented in 8bytes in EUC-KR). 

This is a serious problem (data loss) for Korean users because obviously they
don't know what characters belong to KS X 1001 (representable in 1 or 2bytes)
and they'd be baffaled at realizing that their emails and the content of
submitted forms get truncated.

Comment 1

[Jungshik Shin]

Attachment: patch (one liner)

This is a trivial patch. I took this opportunity to document it better.

Comment 2

[Jungshik Shin]

Comment on attachment 150112 [details] [diff] [review]
patch (one liner)

asking for r/sr. i'll ask for a1.7 as soon as I get r/sr.

Comment 3

[Jungshik Shin]

FYI, this was a regression due to 

http://bonsai.mozilla.org/cvsview2.cgi?diff_mode=context&whitespace_mode=show&file=ugen.c&branch=&root=/cvsroot&subdir=mozilla/intl/uconv/src&command=DIFF_FRAMESET&rev1=1.19&rev2=1.20

(which I'm responsible for)

Comment 4

[Simon Montagu :smontagu]

Comment on attachment 150112 [details] [diff] [review]
patch (one liner)

r=smontagu

Comment 5

[Jungshik Shin]

Comment on attachment 150112 [details] [diff] [review]
patch (one liner)

asking for 1.7a
The risk is very low and the patch is a trivial one-liner. However, it fixes a
serious bug affecting a lot of Korean users.

Comment 6

[Jungshik Shin]

forgot to note that it's fixed on the trunk. thanks for quick r/sr.
I'm keeping this open until i get it fixed on 1.7branch and aviary-1.0

Comment 7

[(not reading, please use [e-mail] instead)]

Comment on attachment 150112 [details] [diff] [review]
patch (one liner)

a=sspitzer for 1.7

for aviary, you should email mscott directly

Comment 8

[Jungshik Shin]

thanks. it's also fixed on 1.7 branch