Develop Reference

GEB: driver is not set as Browser.driver

I'm writing tests with GEB and Spock (I'm new to both).
Driver is declared in GebConfig (updated - the full config file added):
import geb.report.ReportState
import geb.report.Reporter
import geb.report.ReportingListener
import io.github.bonigarcia.wdm.WebDriverManager
import io.qameta.allure.Allure
import org.openqa.selenium.Dimension
import org.openqa.selenium.Point
import org.openqa.selenium.WebDriver
import org.openqa.selenium.chrome.ChromeDriver
import org.openqa.selenium.chrome.ChromeOptions
import org.openqa.selenium.firefox.FirefoxDriver
import org.openqa.selenium.firefox.FirefoxOptions
import org.openqa.selenium.firefox.FirefoxProfile
import org.slf4j.LoggerFactory
import utils.Configuration
def logger = LoggerFactory.getLogger(this.class)
baseUrl = "${Configuration.getStringProperty("BASE_URL")}/${Configuration.getStringProperty("CONTEXT_PATH")}"
baseNavigatorWaiting = true
autoClearCookies = false
cacheDriver = false
reportsDir = 'build/test-reports'
driver = {
WebDriver dr
switch (Configuration.getStringProperty("BROWSER_NAME", "chrome").trim().toLowerCase()) {
case "firefox":
case "ff":
dr = new FirefoxDriver(setUpFirefoxOptions())
break
case "google chrome":
case "chrome":
default:
dr = new ChromeDriver(setUpGoogleChromeOptions())
}
if (Configuration.getBooleanProperty("SET_DRIVER_POSITION", false)) {
dr.manage().window().setPosition(new Point(
Configuration.getIntProperty("BROWSER_X_POS", 0),
Configuration.getIntProperty("BROWSER_Y_POS", 0)))
dr.manage().window().setSize(new Dimension(
Configuration.getIntProperty("BROWSER_WIDTH", 1600),
Configuration.getIntProperty("BROWSER_HEIGHT", 900)));
} else {
dr.manage().window().maximize()
}
return dr
}
static ChromeOptions setUpGoogleChromeOptions() {
WebDriverManager.chromedriver().setup()
ChromeOptions options = new ChromeOptions()
String args = Configuration.getStringProperty("BROWSER_ARGS")
if (args) {
Arrays.stream(args.split("\\s")).each { options.addArguments(it) }
}
return options
}
static FirefoxOptions setUpFirefoxOptions() {
WebDriverManager.firefoxdriver().setup()
FirefoxOptions options = new FirefoxOptions()
FirefoxProfile profile = new FirefoxProfile()
profile.setPreference("network.automatic-ntlm-auth.trusted-uris", "http://,https://")
options.setProfile(profile).setLegacy(false)
return options
}
reportingListener = new ReportingListener() {
void onReport(Reporter reporter, ReportState reportState, List<File> reportFiles) {
def fileGroups = reportFiles.groupBy { it.name.split("\\.")[-1] }
fileGroups['png']?.each {
Allure.addAttachment(it.name, "image/png", new FileInputStream(it), "png")
}
}
}
Test example looks like (BaseTest code is added below):
class SimulationsRunningSpec extends BaseTest {
def "My great test"() {
println("test started")
setup:
to LoginPage
when:
println("when")
then:
println("then")
}
def cleanupSpec() {
browser.quit()
println "Clean up specification"
}
}
And I get the following log sequence:
test started
Created driver
when
then
Created driver
Clean up specification
So the driver gets created when to LoginPage is called.
Issue:
It is not set as Browser driver, so when the browser.quit() is called, a new instance is created and then closed (the first one still is opened).
Questions:
How to set the driver to browser properly to close it then via browser.quit()?
Am I right assuming that if I need to create a driver in setupSpec I can simply call to LoginPage there? Or what is the best way to init the driver in preconditions?
UPDATE:
After some debugging I found out that for some reason, browser gecomes null and is created again in cleanupSpec(). It doesn't matter whether the Spec extends Geb classes of custom base class. This reproduces my issue:
class TestSpec extends GebReportingSpec {
def setupSpec() {
to Page
println "setupSpec browser: $browser"
}
def setup(){
println "setup browser: $browser"
}
def "My first test"() {
println("test started")
when:
println ''
then:
println ''
}
def cleanup() {
println "cleanup browser: $browser"
}
def cleanupSpec() {
println "cleanupSpec browser: $browser"
}
}
This produces the following output:
setupSpec browser: geb.Browser#4beeb0e
setup browser: geb.Browser#4beeb0e
test started
cleanup browser: geb.Browser#4beeb0e
cleanupSpec browser: geb.Browser#5c73f672
The last two rows show that the browser object in cleanupSpec is different from created object in setupSpec.

I'm not sure, why the browser was closed before your cleanupSpec. Probably some other mechanism already took care of it.
The fact that you are getting a different instance in your cleanupSpec however is simply due to the fact, that getBrowser is implemented as a lazy getter. It creates a new instance if necessary, as you can see in the code.
Generally you don't need to call browser.quit using Geb. Geb takes care of that just fine.
Update
Here is what happens in GebSpec and YourSpec:
GebSpec.setupSpec is triggered ⇒ _browser is null
YourSpec.setupSpec is triggered ⇒ _browser is still null unless you use it here
GebSpec.setup is triggered ⇒ _browser is not changed
YourSpec.setup is triggered ⇒ _browser might be changed
YouSpec's first feature is triggered ⇒ _browser is used, so it won't be null anymore
YourSpec.cleanup is triggered ⇒ _browser is not changed
GebSpec.cleanup is triggered ⇒ _browser is set to null! As you can see in the code, resetBrowser is called unless YourSpec is #Stepwise and that sets _browser to null as you can see here.
YourSpec.cleanupSpec is triggered ⇒ _browser is null unless you use it, so it gets reinitialized
GebSpec.cleanupSpec is triggered ⇒ _browser is still null

This seems strange that you are seeing the browser reinitialise for the cleanup, what you have shown is correct.
For point 1: You are setting it correctly within the gebconfig.
For point 2: You don't need to initialise the browser within the setupSpec(), the config entry is all you need.
The browser should close automatically once all tests are run, UNLESS you have added the following to your gebconfig and set to false:
quitCachedDriverOnShutdown = false
setupSpec() is called after all methods within the spec have been run. Is what you have shown us the only code within your spec? Is your spec extending GebSpec or GebReportingSpec or a custom base class?
The only other thing i can think is that you have 2 tests in that spec, so you're seeing "Created driver" twice, and the cleanUpSpec() is called after all tests are run so you see that called at the end. If you had called cleanup() it would run between each test.

Add custom folders to classpath in bazel java tests

I'm trying to migrate a large codebase from maven to bazel and I've found that some of the tests write to target/classes and target/test-classes and the production code reads it as resources on the classpath. This is because maven surefire/failsafe run by default from the module directory and add target/classes and target/test-classes to the classpath.
For me to migrate this large codebase the only reasonable solution is to create target, target/classes and target/test-classes folders and add the last two to the classpath of the tests.
Any ideas on how this can be achieved?
Thanks

Another line of approach. Instead of generating a test suite, create a custom javaagent and a custom class loader. Use jvm_flags to setup and configure it.
The javaagent has a premain method. This sounds like a natural place to do things that happen before the regular main method, even if they don't have anything to do with class instrumentation, debugging, coverage gathering, or any other usual uses of javaagents.
The custom javaagent reads system property extra.dirs and creates directories specified there. It then reads property extra.link.path and creates the symbolic links as specified there, so I can place resources where the tests expect them, without having to copy them.
Classloader is needed so that we can amend the classpath at runtime without hacks. Great advantage is that this solution works on Java 10.
The custom classloader reads system property extra.class.path and (in effect) prepends it before what is in java.class.path.
Doing things this way means that standard bazel rules can be used.
BUILD
runtime_classgen_dirs = ":".join([
"target/classes",
"target/test-classes",
])
java_test(
...,
jvm_flags = [
# agent
"-javaagent:$(location //tools:test-agent_deploy.jar)",
"-Dextra.dirs=" + runtime_classgen_dirs,
# classloader
"-Djava.system.class.loader=ResourceJavaAgent",
"-Dextra.class.path=" + runtime_classgen_dirs,
],
,,,,
deps = [
# not runtime_deps, cause https://github.com/bazelbuild/bazel/issues/1566
"//tools:test-agent_deploy.jartest-agent_deploy.jar"
],
...,
)
tools/BUILD
java_binary(
name = "test-agent",
testonly = True,
srcs = ["ResourceJavaAgent.java"],
deploy_manifest_lines = ["Premain-Class: ResourceJavaAgent"],
main_class = "ResourceJavaAgent",
visibility = ["//visibility:public"],
)
tools/ResourceJavaAgent.java
import java.io.File;
import java.io.IOException;
import java.lang.instrument.Instrumentation;
import java.net.MalformedURLException;
import java.net.URL;
import java.net.URLClassLoader;
import java.nio.file.Files;
import java.nio.file.Path;
import java.nio.file.Paths;
import java.util.Arrays;
import java.util.LinkedList;
import java.util.List;
// https://stackoverflow.com/questions/60764/how-should-i-load-jars-dynamically-at-runtime
public class ResourceJavaAgent extends URLClassLoader {
private final ClassLoader parent;
public ResourceJavaAgent(ClassLoader parent) throws MalformedURLException {
super(buildClassPath(), null);
this.parent = parent; // I need the parent as backup for SecretKeyFactory.getInstance("PBKDF2WithHmacSHA1");
System.out.println("initializing url classloader");
}
private static URL[] buildClassPath() throws MalformedURLException {
final String JAVA_CLASS_PATH = "java.class.path";
final String EXTRA_CLASS_PATH = "extra.class.path";
List<String> paths = new LinkedList<>();
paths.addAll(Arrays.asList(System.getProperty(EXTRA_CLASS_PATH, "").split(File.pathSeparator)));
paths.addAll(Arrays.asList(System.getProperty(JAVA_CLASS_PATH, "").split(File.pathSeparator)));
URL[] urls = new URL[paths.size()];
for (int i = 0; i < paths.size(); i++) {
urls[i] = Paths.get(paths.get(i)).toUri().toURL(); // important only for resource url, really: this url must be absolute, to pass getClass().getResource("/users.properties").toURI()) with uri that isOpaque == false.
// System.out.println(urls[i]);
}
// this is for spawnVM functionality in tests
System.setProperty(JAVA_CLASS_PATH, System.getProperty(EXTRA_CLASS_PATH, "") + File.pathSeparator + System.getProperty(JAVA_CLASS_PATH));
return urls;
}
#Override
public Class<?> loadClass(String s) throws ClassNotFoundException {
try {
return super.loadClass(s);
} catch (ClassNotFoundException e) {
return parent.loadClass(s); // we search parent second, not first, as the default URLClassLoader would
}
}
private static void createRequestedDirs() {
for (String path : System.getProperty("extra.dirs", "").split(File.pathSeparator)) {
new File(path).mkdirs();
}
}
private static void createRequestedLinks() {
String linkPaths = System.getProperty("extra.link.path", null);
if (linkPaths == null) {
return;
}
for (String linkPath : linkPaths.split(",")) {
String[] fromTo = linkPath.split(":");
Path from = Paths.get(fromTo[0]);
Path to = Paths.get(fromTo[1]);
try {
Files.createSymbolicLink(from.toAbsolutePath(), to.toAbsolutePath());
} catch (IOException e) {
throw new IllegalArgumentException("Unable to create link " + linkPath, e);
}
}
}
public static void premain(String args, Instrumentation instrumentation) throws Exception {
createRequestedDirs();
createRequestedLinks();
}
}

If you could tell the tests where to write these files (in case target/classes and target/test-classes are hardcoded), and then turn the test run into a genrule, then you can specify the genrule's outputs as data for the production binary's *_binary rule.

I solved the first part, creating the directories. I still don't know how to add the latter two to classpath.
Starting from https://gerrit.googlesource.com/bazlets/+/master/tools/junit.bzl, I modified it to read
_OUTPUT = """import org.junit.runners.Suite;
import org.junit.runner.RunWith;
import org.junit.BeforeClass;
import java.io.File;
#RunWith(Suite.class)
#Suite.SuiteClasses({%s})
public class %s {
#BeforeClass
public static void setUp() throws Exception {
new File("./target").mkdir();
}
}
"""
_PREFIXES = ("org", "com", "edu")
# ...
I added the #BeforeClass setUp method.
I stored this as junit.bzl into third_party directory in my project.
Then in a BUILD file,
load("//third_party:junit.bzl", "junit_tests")
junit_tests(
name = "my_bundled_test",
srcs = glob(["src/test/java/**/*.java"]),
data = glob(["src/test/resources/**"]),
resources = glob(["src/test/resources/**"]),
tags = [
# ...
],
runtime_deps = [
# ...
],
],
deps = [
# ...
],
)
Now the test itself is wrapped with a setUp method which will create a directory for me. I am not deleting them afterwards, which is probably a sound idea to do.
The reason I need test resources in a directory (as opposed to in a jar file, which bazel gives by default) is that my test passes the URI to new FileInputStream(new File(uri)). If the file resides in a JAR, the URI will be file:/path/to/my.jar!/my.file and the rest of the test cannot work with such URI.

Remove folder from Java classpath at runtime

Is there a way to remove a folder from the classpath similar to adding a folder at runtime (Can a directory be added to the class path at runtime?)

Please find below a snippet as technical example to demonstrate adding / removing a path.
create following source files in any directory
import java.io.File;
import java.lang.reflect.Field;
import java.lang.reflect.Method;
import java.net.URL;
import java.net.URLClassLoader;
import java.util.Stack;
import sun.misc.URLClassPath;
public class EvilPathDemo {
public static void addPath(String path) throws Exception {
URL u = new File(path).toURI().toURL();
URLClassLoader urlClassLoader = (URLClassLoader)
ClassLoader.getSystemClassLoader();
Class<?> urlClass = URLClassLoader.class;
Method method = urlClass.getDeclaredMethod("addURL",
new Class[]{URL.class}
);
method.setAccessible(true);
method.invoke(urlClassLoader, new Object[]{u});
}
public static void removePath(String path) throws Exception {
URL url = new File(path).toURI().toURL();
URLClassLoader urlClassLoader = (URLClassLoader)
ClassLoader.getSystemClassLoader();
Class<?> urlClass = URLClassLoader.class;
Field ucpField = urlClass.getDeclaredField("ucp");
ucpField.setAccessible(true);
URLClassPath ucp = (URLClassPath) ucpField.get(urlClassLoader);
Class<?> ucpClass = URLClassPath.class;
Field urlsField = ucpClass.getDeclaredField("urls");
urlsField.setAccessible(true);
Stack urls = (Stack) urlsField.get(ucp);
urls.remove(url);
}
public static void main(String[] args) throws Exception {
String parm = args.length == 1 ? args[0] : "";
String evilPath = "/tmp";
String classpath = System.getProperty("java.class.path");
boolean isEvilPathSet = false;
for (String path : classpath.split(File.pathSeparator)) {
if (path.equalsIgnoreCase(evilPath)) {
System.out.printf("evil path '%s' in classpath%n", evilPath);
isEvilPathSet = true;
break;
}
}
if (isEvilPathSet && parm.equalsIgnoreCase("REMOVE")) {
System.out.printf("evil path '%s' will be removed%n", evilPath);
removePath(evilPath);
}
tryToLoad("Foo");
if (parm.equalsIgnoreCase("ADD")) {
System.out.printf("evil path '%s' will be added%n", evilPath);
addPath(evilPath);
}
tryToLoad("Bar");
}
private static void tryToLoad(String className) {
try {
Class<?> foo = Class.forName(className);
System.out.printf("class loaded: %s%n", foo.getName());
} catch (ClassNotFoundException ex) {
System.out.println(ex);
}
}
}
.
public class Foo {
static {
System.out.println("I'm foo...");
}
}
.
public class Bar {
static {
System.out.println("I'm bar...");
}
}
compile them as follow
javac EvilPathDemo.java
javac -d /tmp Foo.java Bar.java
During the test we will try to load the classes Foo and Bar.
without /tmp in the classpath
java -cp . EvilPathDemo
java.lang.ClassNotFoundException: Foo
java.lang.ClassNotFoundException: Bar
adding /tmp to the classpath
java -cp . EvilPathDemo add
java.lang.ClassNotFoundException: Foo
evil path '/tmp' will be added
I'm bar...
class loaded: Bar
with /tmp in the classpath
java -cp .:/tmp EvilPathDemo
evil path '/tmp' in the classpath
I'm foo...
class loaded: Foo
I'm bar...
class loaded: Bar
remove /tmp from the classpath
java -cp .:/tmp EvilPathDemo remove
evil path '/tmp' in the classpath
evil path '/tmp' will be removed
java.lang.ClassNotFoundException: Foo
java.lang.ClassNotFoundException: Bar
During the testing I found out that following cases are not working.
addPath(evilPath);
tryToLoad("Foo");
removePath(evilPath); // had not effect
tryToLoad("Bar");
removePath(evilPath);
tryToLoad("Foo");
addPath(evilPath); // had no effect
tryToLoad("Bar");
tryToLoad("Foo");
removePath(evilPath); // had no effect
tryToLoad("Bar");
I did not spent time to find out why. Because I don't see any practical use in it. If you really need/wish to play with the classpaths have a look how classloaders are working.

The removePath method from above did not work for me and my Weld Container, the url stack was always emtpy.
The following ugly smugly method worked:
public static void removeLastClasspathEntry() throws Exception {
URLClassLoader urlClassLoader = (URLClassLoader) ClassLoader.getSystemClassLoader();
Class<?> urlClass = URLClassLoader.class;
Field ucpField = urlClass.getDeclaredField("ucp");
ucpField.setAccessible(true);
URLClassPath ucp = (URLClassPath) ucpField.get(urlClassLoader);
Field loadersField = URLClassPath.class.getDeclaredField("loaders");
loadersField.setAccessible(true);
List jarEntries = (List) loadersField.get(ucp);
jarEntries.remove(jarEntries.size() - 1);
Field pathField = URLClassPath.class.getDeclaredField("path");
pathField.setAccessible(true);
List pathList = (List) pathField.get(ucp);
URL jarUrl = (URL) pathList.get(pathList.size() - 1);
String jarName = jarUrl.toString();
pathList.remove(pathList.size() - 1);
Field lmapField = URLClassPath.class.getDeclaredField("lmap");
lmapField.setAccessible(true);
Map lmapMap = (Map) lmapField.get(ucp);
lmapMap.remove(jarName.replaceFirst("file:/", "file:///"));
}

Class loaders can be nested so instead of modifying the system class loader which is the root of the tree of class loaders, it is better to simply create a nested classloader and use that to load classes.
The system classloader itself is immutable (for good reasons) but you can do whatever you want in nested class loaders, including destroying them to unload classes and resources. This is commonly used in e.g. osgi and application servers to load/unload e.g. plugins, applications, etc.
For nested class loaders you can completely customize how to load classes. The URLClassloader is probably a good starting point for what you want.

I dont think there is a straight forward way to do it. You can follow :
Get class path variables using : System.getenv("CLASSPATH"). It will return semi colon separated values.
String classPath = System.getenv("CLASSPATH")
Take the folder path as input and replace it with "" like :
String remainigPath = classPath.replace(inputpath,"");
Put the remaining paths in an array using split method.
String[] paths = remainigPath .split(";");
For adding classPath, You already have the code.

I had the same issue, so I tackled it by creating a library that works on every ClassLoader that uses a URLClassPath (so, currently, URLClassLoader).
The library has methods for:
Adding new entries in front
Appending new entries
Remove existing entries
Please note that, since this library accesses internal and proprietary APIs, it is no guaranteed to work in future versions of the JDK. It currently does for Java 7 and Java 8 (Oracle and OpenJDK).
Here is the GitHub page (contribution is appreciated), and here is the Maven Central artifact

groovy win cmd line class and script

I'm trying to run a groovy(2.4.3) script on windows that calls a goovy class xxxxx.groovy. I've tried a number of variations using classpath and various scripts, some examples below, always getting MultipleCompliationErrorsException.... unable to resolve class
classfile is firstclass.groovy
import org.apache.commons.io.FilenameUtils
class firstclassstart {
def wluid, wlpwd, wlserver, port
private wlconnection, connectString, jmxConnector, Filpath, Filpass, Filname, OSRPDpath, Passphrase
// object constructor
firstclassstart(wluid, wlpwd, wlserver, port) {
this.wluid = wluid
this.wlpwd = wlpwd
this.wlserver = wlserver
this.port = port
}
def isFile(Filpath) {
// Create a File object representing the folder 'A/B'
def folder = new File(Filpath)
if (!org.apache.commons.io.FilenameUtils.isExtension(Filpath, "txt")) {
println "bad extension"
return false
} else if (!folder.exists()) {
// Create all folders up-to and including B
println " path is wrong"
return false
} else
println "file found"
return true
}
}
cmd line script test.groovy
import firstclass
def sample = new firstclass.firstclassstart("weblogic", "Admin123", "x.com", "7002")
//def sample = new firstclassstart("weblogic", "Admin123", "x.com", "7002")
sample.isFile("./firstclass.groovy")
..\groovy -cp "firstclass.groovy;commons-io-1.3.2.jar" testfc.groovy
script test.groovy
GroovyShell shell = new GroovyShell()
def script = shell.parse(new File('mylib/firstclass.groovy'))
firstclass sample = new script.firstclass("uid", "pwd", "url", "port")
sample.getstatus()
c:>groovy test.groovy
script test.groovy v2 put firstclass.groovy in directory test below script
import test.firstclass
firstclass sample = new script.firstclass("uid", "pwd", "url", "port")
sample.getstatus()
c:>groovy test.groovy
just looking for a bullet proof, portable way to oranize my java classes, .groovy classess, etc. and scripts.
Thanks

I think that you can do using for example your first approach:
groovy -cp mylib/firstclass.groovy mylib/test.groovy
However I see some problems in your code which are probably causing MultipleCompliationErrorsException.
Since you're including firstclass.groovy in your classpath, you've to add the import firstclass in the test.groovy.
Why are you using script.firstclass in test.groovy? you're class is called simply firstclass.
In your firstclass.groovy you're using import org.apache.commons.io.FilenameUtils and probably other, however you're not including it in the classpath.
So finally I think that, you've to change your test.groovy for something like:
import firstclass
firstclass sample = new firstclass("uid", "pwd", "url", "port")
sample.getstatus()
And in your command add the remaining includes for apache Commons IO to the classpath.
groovy -cp "mylib/firstclass.groovy;commons-io-2.4.jar;" mylib/testexe.groovy
Hope this helps,
UPDATE BASED ON OP CHANGES:
After the changes you've some things wrong, I try to enumerate it:
If your file is called firstclass.groovy your class must be class firstclass not class firstclassstart.
In your test.groovy use new firstclass not new firstclass.firstclassstart.
So the thing is, your code must be:
class file firstclass.groovy:
import org.apache.commons.io.FilenameUtils
class firstclass {
def wluid, wlpwd, wlserver, port
private wlconnection, connectString, jmxConnector, Filpath, Filpass, Filname, OSRPDpath, Passphrase
// object constructor
firstclass(wluid, wlpwd, wlserver, port) {
this.wluid = wluid
this.wlpwd = wlpwd
this.wlserver = wlserver
this.port = port
}
def isFile(Filpath) {
// Create a File object representing the folder 'A/B'
def folder = new File(Filpath)
if (!org.apache.commons.io.FilenameUtils.isExtension(Filpath, "txt")) {
println "bad extension"
return false
} else if (!folder.exists()) {
// Create all folders up-to and including B
println " path is wrong"
return false
} else
println "file found"
return true
}
}
script test.groovy:
import firstclass
def sample = new firstclass("weblogic", "Admin123", "x.com", "7002")
sample.isFile("./firstclass.groovy")
Finally the command to execute it:
groovy -cp "firstclass.groovy;commons-io-1.3.2.jar" test.groovy
With this changes your code must works, I try it and works as expected.

value Engine is not a member of object android.speech.tts.TextToSpeech

In java, using android-19, this works fine:
import android.speech.tts.TextToSpeech.Engine;
But in scala:
import android.speech.tts.TextToSpeech.Engine
^
error: value Engine is not a member of object android.speech.tts.TextToSpeech
This is weird because I can import EngineInfo without problems, but it does not recognize the class Engine, although it is clearly in the android.jar
Any fix or workaround? Is this a scala issue? I'm developping on Eclipse 3.7 using AndroidProguardScala v51 and Scala IDE version 3.0.1
EDIT
The context in which I am using this is for example the following:
if(resultCode == TextToSpeech.Engine.CHECK_VOICE_DATA_PASS) {
mTts = new TextToSpeech(this, mTtsListener)
}

Unfortunately, TextToSpeech.Engine is declared non-static which means something like this:
public class TextToSpeech {
public class Engine {
// ...
}
// ...
}
So according to this, you can't access it. Here's my workaround using a wrapper:
object ConstantsWrapper {
private val cls = classOf[TextToSpeech#Engine]
def apply(fieldName: String) = cls.getDeclaredField(fieldName).get(null)
.toString
val CHECK_VOICE_DATA_PASS = apply("CHECK_VOICE_DATA_PASS")
// Insert more here
}
if (resultCode == ConstantsWrapper.CHECK_VOICE_DATA_PASS) {
mTts = new TextToSpeech(this, mTtsListener)
}