java.util类和方法中嵌套的可读性

Question

I'm probably going to get 'Duplicated Post' but anyway.

Since I started programming, I learned that you should avoid nested loops, if/else, try/catch etc. because the code gets hard to read and understanding the context is quite difficult. There are a couple of things you can do to avoid them like ENUMs and Interfaces.

While looking for answers, I found posts saying, that nested loops are not always bad, as long as its readable.

Now while working on a project, I came across the java.util.Properties class with some very complicated methods with a lot of nesting like this:

 /*
 * Converts encoded \uxxxx to unicode chars
 * and changes special saved chars to their original forms
 */
private String loadConvert (char[] in, int off, int len, char[] convtBuf) {
    if (convtBuf.length < len) {
        int newLen = len * 2;
        if (newLen < 0) {
            newLen = Integer.MAX_VALUE;
        }
        convtBuf = new char[newLen];
    }
    char aChar;
    char[] out = convtBuf;
    int outLen = 0;
    int end = off + len;

    while (off < end) {
        aChar = in[off++];
        if (aChar == '\\') {
            aChar = in[off++];
            if(aChar == 'u') {
                // Read the xxxx
                int value=0;
                for (int i=0; i<4; i++) {
                    aChar = in[off++];
                    switch (aChar) {
                      case '0': case '1': case '2': case '3': case '4':
                      case '5': case '6': case '7': case '8': case '9':
                         value = (value << 4) + aChar - '0';
                         break;
                      case 'a': case 'b': case 'c':
                      case 'd': case 'e': case 'f':
                         value = (value << 4) + 10 + aChar - 'a';
                         break;
                      case 'A': case 'B': case 'C':
                      case 'D': case 'E': case 'F':
                         value = (value << 4) + 10 + aChar - 'A';
                         break;
                      default:
                          throw new IllegalArgumentException(
                                       "Malformed \\uxxxx encoding.");
                    }
                 }
                out[outLen++] = (char)value;
            } else {
                if (aChar == 't') aChar = '\t';
                else if (aChar == 'r') aChar = '\r';
                else if (aChar == 'n') aChar = '\n';
                else if (aChar == 'f') aChar = '\f';
                out[outLen++] = aChar;
            }
        } else {
            out[outLen++] = aChar;
        }
    }
    return new String (out, 0, outLen);
}

For me this is anything but readable. So I'm generally curious, if this is the best or the only way to make this. Because since this is as old as JDK1.0 shouldn't this have been 'cleaned' up or changed for readabilty? And there are a lot more methods with complicated and complex code.

thanks

Answer 1

Let's try to focus on facts , not opinions here:

• although "Clean Code advocates" don't like to hear that - but "clean code" rules are still "subjective. Some people think they are "holy grail", others do not agree to that.
• more importantly, the code you are showing here is probably 15+ years old. Most likely, it would be written differently nowadays.

And yes, you are definitely not alone in thinking "this code looks unreadable". To a certain degree, that is because we are not used to such code any more, and to a certain degree that is because this code is hard to read/grasp/understand.

Answer 2

It is low-level (like post-increments), written for speed / production usage, two nested loops, and aspects okay to me. Some spaces and maybe comment missing.

This low-level character is especially evident in not using other functions.

However I would not see how better style would be feasible, without calling other new methods (as function decompisition would slow things down).