Use StringBuilder instead of StringBuffer

```java

private StringBuffer currentLine = new StringBuffer();

....

public String nextToken() {
while (st.hasMoreTokens()) {
String t = st.nextToken();
if (t.equals("\n")) {
++ line;
currentLine.setLength(0);
} else {
currentLine.append(t);
}
if (!isWhitespace(t)) {
return t;
}
}
throw new IllegalArgumentException("unexpected end of schema");
}
```

Use `StringBuilder` instead of `StringBuffer` as `StringBuffer` is synchronized (which is not required here).

**Reporter**: [David Mollitor](https://issues.apache.org/jira/secure/ViewProfile.jspa?name=belugabehr) / @belugabehr

_{**Note**: *This issue was originally created as [PARQUET-1921](https://issues.apache.org/jira/browse/PARQUET-1921). Please see the [migration documentation](https://issues.apache.org/jira/browse/PARQUET-2502) for further details.*}

Search the repository for the shown StringBuffer declaration and currentLine usage, then read the surrounding tokenizer code to confirm the buffer is not shared between threads. Replace the identified use with StringBuilder and run the repository's relevant Java build or test command to verify the tokenizer still behaves correctly.

Written by the indexing model from the issue text.