fix the line handling of UTF-16le

[kmuto]

Hi, I noticed check-log doesn't correctly determnine UTF-16le (Little Endian).

The following problems occur with UTF-16le because log files are separated by "\n" based on bytes rather than based on encoding.

• The newline character 0x0a 0x00 in UTF-16le is delimited in the middle, so the second and subsequent lines contain 0x00 at the beginning.
• Similarly, line endings are broken when converted to UTF-8 because only 0x0a is used instead of 0x0a 0x00

As this does not happen with other encodings (unless you consider UTF-32), it seemed reasonable to deal only with UTF-16le.
I propose a patch here.

Test:

[Before patch]
$ go run main.go -p ☺️  -f testcode/t1-u16le.txt --encoding=UTF-16le --check-first -s . --return
LOG CRITICAL: 1 warnings, 1 criticals for pattern /☺️/.
[testcode/t1-u16le.txt]
☺️�

exit status 2

$ go run main.go -p ☺️  -f testcode/t1-u16le-bom.txt --encoding=UTF-16le --check-first -s . --return
LOG CRITICAL: 1 warnings, 1 criticals for pattern /☺️/.
[testcode/t1-u16le-bom.txt]
☺️�

exit status 2

$ go run main.go -p ☺️  -f testcode/t1-u16le-bom-crlf.txt --encoding=UTF-16le --check-first -s . --return
LOG CRITICAL: 1 warnings, 1 criticals for pattern /☺️/.
[testcode/t1-u16le-bom-crlf.txt]
�

exit status 2

[after patch]
$ go run main.go -p ☺️  -f testcode/t1-u16le.txt --encoding=UTF-16le --check-first -s . --return
LOG CRITICAL: 3 warnings, 3 criticals for pattern /☺️/.
[testcode/t1-u16le.txt]
☺️
☺️
☺️

exit status 2

$ go run main.go -p ☺️  -f testcode/t1-u16le-bom.txt --encoding=UTF-16le --check-first -s . --return
LOG CRITICAL: 3 warnings, 3 criticals for pattern /☺️/.
[testcode/t1-u16le-bom.txt]
☺️
☺️
☺️

exit status 2

$ go run main.go -p ☺️  -f testcode/t1-u16le-bom-crlf.txt --encoding=UTF-16le --check-first -s . --return
LOG CRITICAL: 3 warnings, 3 criticals for pattern /☺️/.
[testcode/t1-u16le-bom-crlf.txt]
☺️
☺️
☺️

exit status 2

testcodes.zip

[yseto]

seems good.