Consider an annotation processor that generates source files and errors in the AP round. E.g.:
---ProcessorImpl.java
import java.io.IOException;
import java.io.Writer;
import java.util.*;

import javax.annotation.processing.*;
import javax.lang.model.SourceVersion;
import javax.lang.model.element.*;
import javax.tools.Diagnostic.Kind;

@SupportedAnnotationTypes("*")
public class ProcessorImpl extends AbstractProcessor {
    int round = 0;
    @Override
    public boolean process(Set<? extends TypeElement> annotations, RoundEnvironment roundEnv) {
        if (round++ == 0) {
            try (Writer w = processingEnv.getFiler().createSourceFile("Extra").openWriter()) {
                w.write("public class Extra {}");
            } catch (IOException ex) {
                throw new IllegalStateException(ex);
            }
            processingEnv.getMessager().printMessage(Kind.ERROR, "processor error");
        }
        return false;
    }

    @Override
    public SourceVersion getSupportedSourceVersion() {
        return SourceVersion.latestSupported();
    }
    
}
---

And consider a source file being compiled with this processor, like:
---Test.java
public class Test extends Extra implements ExtraExtra {}
---

---
$ javac -processor ProcessorImpl /tmp/Test.java
Test.java:1: error: cannot find symbol
public class Test extends Extra implements ExtraExtra {}
                          ^
  symbol: class Extra
Test.java:1: error: cannot find symbol
public class Test extends Extra implements ExtraExtra {}
                                           ^
  symbol: class ExtraExtra
error: processor error
3 errors
---

This is inconvenient for two reasons:
-the "cannot find Extra" error is resolved by the file the AP generated
-the error reported by the AP is probably quite important, but is listed after the other errors, some of which are fixed by the AP