h**********c
发帖数: 4120 | 1 鼓捣了两天jaxb，
marshal的时候，没明白设utf-8 和 utf-16有什么区别，
试验了一下utf-16，结果生成的文件全是乱码，
另外java 是怎么处理中文的，
知道的就给扫扫盲吧，也不想太深入，将来要用的时候有个底。 |
h**********c
发帖数: 4120 | 2 re
【在 h**********c 的大作中提到】
: 鼓捣了两天jaxb，
: marshal的时候，没明白设utf-8 和 utf-16有什么区别，
: 试验了一下utf-16，结果生成的文件全是乱码，
: 另外java 是怎么处理中文的，
: 知道的就给扫扫盲吧，也不想太深入，将来要用的时候有个底。
|
g*****g
发帖数: 34805 | 3 utf-8是可变长度，1-3个字节编码都有。utf-16是固定双字节编码，
java内部总是用utf-16
【在 h**********c 的大作中提到】
: 鼓捣了两天jaxb，
: marshal的时候，没明白设utf-8 和 utf-16有什么区别，
: 试验了一下utf-16，结果生成的文件全是乱码，
: 另外java 是怎么处理中文的，
: 知道的就给扫扫盲吧，也不想太深入，将来要用的时候有个底。
|
F****n
发帖数: 3271 | 4 Use UTF-8, not UTF-16. UTF-16 is based on the naive assumption that 2 bytes
are enough to encode all characters in the world. This has proved to be
wrong, and it has now become the source of a big mess in Java's internal
character representation, which is based on UTF-16.
【在 h**********c 的大作中提到】
: 鼓捣了两天jaxb，
: marshal的时候，没明白设utf-8 和 utf-16有什么区别，
: 试验了一下utf-16，结果生成的文件全是乱码，
: 另外java 是怎么处理中文的，
: 知道的就给扫扫盲吧，也不想太深入，将来要用的时候有个底。
|
c*****t
发帖数: 1879 | 5 Without the naive assumption, most string computations will be too complex
and inefficient.
It is merely a trade off people make.
bytes
【在 F****n 的大作中提到】
: Use UTF-8, not UTF-16. UTF-16 is based on the naive assumption that 2 bytes
: are enough to encode all characters in the world. This has proved to be
: wrong, and it has now become the source of a big mess in Java's internal
: character representation, which is based on UTF-16.
|
F****n
发帖数: 3271 | 6 Why?
【在 c*****t 的大作中提到】
: Without the naive assumption, most string computations will be too complex
: and inefficient.
: It is merely a trade off people make.
:
: bytes
|
p*a
发帖数: 592 | 7 utf16不是有surrogate pair吗，可以支持4个字节
bytes
【在 F****n 的大作中提到】
: Use UTF-8, not UTF-16. UTF-16 is based on the naive assumption that 2 bytes
: are enough to encode all characters in the world. This has proved to be
: wrong, and it has now become the source of a big mess in Java's internal
: character representation, which is based on UTF-16.
|
m****r
发帖数: 6639 | 8 我怀疑他看到乱码, 是因为他用notepad之类的去 开, 然后因为notepad不支持utf-16
所以是乱码. 没有道理说utf-8可以写, utf-16就写出来不对了.
【在 g*****g 的大作中提到】
: utf-8是可变长度，1-3个字节编码都有。utf-16是固定双字节编码，
: java内部总是用utf-16
|
F****n
发帖数: 3271 | 9 Then why UTF-16 in the first place. The fact is UTF-16 is a older and broken
system that people have to fix .
【在 p*a 的大作中提到】
: utf16不是有surrogate pair吗，可以支持4个字节
:
: bytes
|
F****n
发帖数: 3271 | 10 UTF-16 has a byte-ordering problem as many programs assume specific byte
ordering without specifying it in the output files. That's why UTF-16 files
frequently have this cross-program issue.
16
【在 m****r 的大作中提到】
: 我怀疑他看到乱码, 是因为他用notepad之类的去 开, 然后因为notepad不支持utf-16
: 所以是乱码. 没有道理说utf-8可以写, utf-16就写出来不对了.
|
h**********c
发帖数: 4120 | 11 thanks for the replies
This question is no longer my focus.
But I remember, not very sure, I did the following tests.
Marshal to UTF-16, then print with printwriter -> unreadable mess.
Marshal to UTF-8, java String has Chinese characters, writing to file, every
Chinese character becomes '?'
Win 7, Eclipse, English platform.
if using VC++ msxml, same PC, no such problem. But msxml has no option to
set UTF-16 or 8.
Thanks again, |
g*****g
发帖数: 34805 | 12 You should decode and encode using UTF-8 for Chinese characters.
Now once it's encoded to UTF-8, you need right edtor to open it,
like M$ word should do it.
every
【在 h**********c 的大作中提到】
: thanks for the replies
: This question is no longer my focus.
: But I remember, not very sure, I did the following tests.
: Marshal to UTF-16, then print with printwriter -> unreadable mess.
: Marshal to UTF-8, java String has Chinese characters, writing to file, every
: Chinese character becomes '?'
: Win 7, Eclipse, English platform.
: if using VC++ msxml, same PC, no such problem. But msxml has no option to
: set UTF-16 or 8.
: Thanks again,
|
