https://bugs.documentfoundation.org/show_bug.cgi?id=172905

Julien Nabet <[email protected]> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |heiko.tietze@documentfounda
                   |                            |tion.org,
                   |                            |[email protected]

--- Comment #11 from Julien Nabet <[email protected]> ---
I've added some debug traces in SvStream::DetectEncoding in
tools/source/stream/stream.cxx
In the KO case, I've got:
getName=UTF-16BE confidence=10
getName=UTF-16LE confidence=10
getName=Shift_JIS confidence=10
getName=GB18030 confidence=10
getName=Big5 confidence=10
getName=ISO-8859-1 confidence=7
getName=ISO-8859-2 confidence=3
and since LO retrieves the first one, it'll use UTF-16BE

In the OK case (same file but truncated to have 10 lines):
getName=ISO-8859-1 confidence=30
getName=ISO-8859-2 confidence=14
getName=UTF-16BE confidence=10
getName=UTF-16LE confidence=10
getName=Shift_JIS confidence=10
getName=GB18030 confidence=10
getName=Big5 confidence=10
getName=ISO-8859-9 confidence=4
so LO will use ISO-8859-1

Since confidence is calculated by ICU lib, I'd say NOTOURBUG for this one.

Xisco/Heiko: any thoughts here?

-- 
You are receiving this mail because:
You are the assignee for the bug.

Reply via email to