简体繁体 English

如何在 java 中使用 utf-8 字符串而不分配新的字符串 object 而是作为字节数组的一部分？

[英]How to work with utf-8 strings in java without allocating a new String object but as part of byte array instead?

原文 2021-03-26 22:38:18 9 1 java/ scala/ utf-8

I have place of my code where I want to read from binary format which include utf-8 strings.我有我的代码的位置，我想从二进制格式中读取，其中包括 utf-8 字符串。

Also I don't want any allocations in this place because they summon GC which pauses the world which kinda bad for me.此外，我不希望在这个地方进行任何分配，因为它们会召唤 GC，这会暂停世界，这对我来说有点糟糕。

I can work perferctly well with most of my primitives and arrays except for strings because java is an "object oriented language" and emphasises on heavy usage of objects (= allocations).我可以完美地使用我的大多数原语和 arrays 除了字符串，因为 java 是一种“面向对象的语言”并且强调对象的大量使用（=分配）。 And it does not provide standard way of working with utf-8 strings without allocations as it has immutable object one.并且它不提供在没有分配的情况下使用 utf-8 字符串的标准方法，因为它具有不可变的 object 之一。 So, what i need from this stuff - validate, extract Char and not to create any other objects.所以，我需要从这些东西中得到什么——验证、提取Char而不是创建任何其他对象。 Ie i should be able to put this thing pool or other place, initialise with data: Array[Byte] and offset: Int and length: X , make no copy.即我应该能够把这个东西池或其他地方，用data: Array[Byte]和offset: Int和length: X ，不复制。 CharIterable thing with ability to refurbish this object to other string. CharIterable能够将此 object 翻新为其他字符串的东西。

So, should I do this by hand or someone already have done this?那么，我应该手动执行此操作还是有人已经执行此操作？

1 个解决方案

I guess you could try to directly call the low-level libraries that String uses internally, like CharsetDecoder which can decode from a ByteBuffer into a pre-allocated CharBuffer.我想您可以尝试直接调用 String 内部使用的低级库，例如CharsetDecoder ，它可以从 ByteBuffer 解码为预分配的 CharBuffer。

But you may be overdoing this, I'd measure if using String (and the associated object allocations) are really a problem first.但是你可能做得过火了，我会先衡量使用 String （以及相关的 object 分配）是否真的是一个问题。