InputStreamReader是否应与appendCodePoint一起使用？

Question

It is a common pattern in Java to read characters from a file with InputStreamReader and append them to a StringBuilder; the obvious way to do it is like:

int c = reader.read();
sb.append((char)c);

However, supposing the file (assuming we specified UTF-8 encoding if it makes a difference) were to contain a character (strictly speaking a code point) that doesn't fit in 16 bits. Would the reader return this as a single 32-bit code point instead of a pair of 16-bit chars?

If so, should the last line above actually read like:

sb.appendCodePoint(c);

Is there a known test case - a sequence of UTF-8 bytes - that would distinguish between the two options?

Answer 1

The Reader returns whatever it can make of the next piece of input, as a single character, as the Javadoc says. The only exception is the EOS indicator, which is -1 as an int. There is no basis for your suggestion.