SAX + Java 的奇怪字符（）行为

Question

在我的 XML 中，我有一个多行元素：

<tag id="sometag" ...>
    | first line
    |     second line
    |         third line
    |     fourth line
<tag ...>
....
<tag id="someothertag" ...>
    | ANOTHER FIRST LINE
    |     ANOTHER SECOND LINE
    |         ANOTHER THIRD LINE
    |     ANOTHER FORTH LINE
<tag ...>

然后在Java中我有必要的startElement、endElement和characters方法，但我发现characters有一些奇怪的行为：

public void characters(char[] ch, int start, int length){
    Log.d(TAG, "characters( "\"" + (new String(ch)).replaceAll("[\r\n]", "\\n") + "\", " + start + ", " + length + " )");
}

除此之外，我对角色什么都不做。我基本上是在创建解析器的两个实例。在一个实例中，我正在搜索sometag。如果我找到我要查找的内容并返回该元素，我会抛出异常。

D/MyProgram( 1565): STARTING document parsing...
D/MyProgram( 1565): characters( "n   ", 0, 1 )
D/MyProgram( 1565): characters( "        | first line", 0, 20 )
D/MyProgram( 1565): characters( "n       | first line", 0, 1 )
D/MyProgram( 1565): characters( "        |   second line", 0, 23 )
D/MyProgram( 1565): characters( "n       |   second line", 0, 1 )
D/MyProgram( 1565): characters( "        |       third line", 0, 26 )
D/MyProgram( 1565): characters( "n       |       third line", 0, 1 )
D/MyProgram( 1565): characters( "        |   fourth lineline", 0, 22 )
D/MyProgram( 1565): characters( "n       |   fourth lineline", 0, 1 )
D/MyProgram( 1565): characters( "        |   fourth lineline", 0, 4 )
D/MyProgram( 1565): Successfully found "sometag"!

...我正在寻找另一个全新的实例someothertag。我做的和以前一样。

D/MyProgram( 1565): STARTING document parsing...
D/MyProgram( 1565): characters( "n", 0, 1 )
D/MyProgram( 1565): characters( "    ", 0, 4 )
D/MyProgram( 1565): characters( "n   ", 0, 1 )
D/MyProgram( 1565): characters( "        | first line", 0, 20 )
D/MyProgram( 1565): characters( "n       | first line", 0, 1 )
D/MyProgram( 1565): characters( "        |   second line", 0, 23 )
D/MyProgram( 1565): characters( "n       |   second line", 0, 1 )
D/MyProgram( 1565): characters( "        |       third line", 0, 26 )
D/MyProgram( 1565): characters( "n       |       third line", 0, 1 )
D/MyProgram( 1565): characters( "        |   fourth lineline", 0, 22 )
D/MyProgram( 1565): characters( "n       |   fourth lineline", 0, 1 )
D/MyProgram( 1565): characters( "        |   fourth lineline", 0, 4 )
D/MyProgram( 1565): Successfully found "someothertag"!

我知道 XML 解析是基于流的（它解析块而不是整个字符串），但这是非常奇怪的行为。以下是我注意到的一些非常令人困惑的事情：

• 对于 characters() 的每次迭代，如果解析器确实完成了解析，则解析器不会从它停止的地方开始或完成字符：我什至得到 before 的字符第一个字符数组（'n'，它是换行符的替换）。
• ch 包含原本不存在的额外字符：“line”附加到“forth line”。
• 当我创建一个全新的解析器实例时，字符被“重新读取”​​。第二次执行应该是这样的：

..这...

D/MyProgram( 1565): characters( "n", 0, 1 )
D/MyProgram( 1565): characters( "    ", 0, 4 )
D/MyProgram( 1565): characters( "n   ", 0, 1 )
D/MyProgram( 1565): characters( "        | ANOTHER FIRST LINE", 0, 20 )
D/MyProgram( 1565): characters( "n       |     ANOTHER SECOND LINE", 0, 1 )

...等等。

知道我做错了什么吗？提前致谢。

Answer 1

正如 Margulies 所说，您没有在传递的字符数组中使用 start 和 length。

public void characters(char[] ch, int start, int length) {
    // use only the indicated segment.
    String str = new String( ch, start, length);  
    Log.d(TAG, "characters( "\"" + str.replaceAll("[\r\n]", "\\n") + "\", " + start + ", " + length + " )");
}