Providing W3C compliance by TikhomirovSergey · Pull Request #829 · appium/java-client

TikhomirovSergey · 2018-02-11T00:12:03Z

Change list

added additional capability 'MobileCapabilityType.FORCE_MJSONWP'
added org.openqa.selenium.remote.NewSessionPayload
update of selenium to 3.9.1

Types of changes

No changes in production code.
Bugfix (non-breaking change which fixes an issue)
New feature (non-breaking change which adds functionality)
Breaking change (fix or feature that would cause existing functionality to not work as expected)

Details

I am not sure that this solution is 100% good but it seem the goal is achived.

@BeforeClasspublicstaticvoidbeforeClass() {
service = AppiumDriverLocalService.buildDefaultService();
service.start();
if (service == null || !service.isRunning()) {
thrownewAppiumServerHasNotBeenStartedLocallyException(
"An appium server node is not started!");
}
FileappDir = newFile("src/test/java/io/appium/java_client");
Fileapp = newFile(appDir, "ApiDemos-debug.apk");
DesiredCapabilitiescapabilities = newDesiredCapabilities();
capabilities.setCapability(MobileCapabilityType.DEVICE_NAME, "Android Emulator");
capabilities.setCapability(MobileCapabilityType.APP, app.getAbsolutePath());
capabilities.setCapability(MobileCapabilityType.FORCE_MJSONWP, true);
driver = newAndroidDriver<>(service.getUrl(), capabilities);
}

https://gist.github.com/TikhomirovSergey/0bcaf79e96cd1ef6644e5e8c2b975596

@BeforeClasspublicstaticvoidbeforeClass() {
service = AppiumDriverLocalService.buildDefaultService();
service.start();
if (service == null || !service.isRunning()) {
thrownewAppiumServerHasNotBeenStartedLocallyException(
"An appium server node is not started!");
}
FileappDir = newFile("src/test/java/io/appium/java_client");
Fileapp = newFile(appDir, "ApiDemos-debug.apk");
DesiredCapabilitiescapabilities = newDesiredCapabilities();
capabilities.setCapability(MobileCapabilityType.DEVICE_NAME, "Android Emulator");
capabilities.setCapability(MobileCapabilityType.APP, app.getAbsolutePath());
//capabilities.setCapability(MobileCapabilityType.FORCE_MJSONWP, true);driver = newAndroidDriver<>(service.getUrl(), capabilities);
}

https://gist.github.com/TikhomirovSergey/5987f5c74e37110427b8f5651030aa3b

There's a lot of overlap between the ProtocolHandshake and the NewSessionPayload. Time to deal with this overlap between the two. The main change is to make the NewSessionPayload capable of writing to an Appendable. It does so by streaming the Capabilities used to create the payload as the OSS capabilities and those used by old geckodrivers, and then streaming every capability through transforms to generate spec compliant capabilities. There appears to be a bug where we always generate synthetic capabilities, but we can deal with that later.

…ties

Using the magic of Buck's "maven-importer" and the following maven coordinates: ``` 'org.seleniumhq.selenium:htmlunit-driver:jar:2.28.5' \ 'junit:junit:jar:4.12' \ 'net.bytebuddy:byte-buddy:jar:1.7.9' \ 'com.google.code.gson:gson:jar:2.8.2' \ 'com.google.guava:guava:jar:23.6-jre' \ 'org.apache.commons:commons-exec:jar:1.3' \ 'org.eclipse.jetty:jetty-security:jar:9.4.8.v20171121' \ 'org.testng:testng:jar:6.13.1' \ 'org.pantsbuild:jarjar:jar:1.6.5' \ 'org.eclipse.jetty:jetty-util:jar:9.4.8.v20171121' \ 'org.eclipse.jetty:jetty-server:jar:9.4.8.v20171121' \ 'org.eclipse.jetty:jetty-servlet:jar:9.4.8.v20171121' \ 'org.hamcrest:hamcrest-library:jar:1.3' \ 'com.github.javaparser:javaparser-core:jar:3.5.7' \ 'org.eclipse.jetty:jetty-jmx:jar:9.4.8.v20171121' \ 'net.jcip:jcip-annotations:jar:1.0' \ 'org.yaml:snakeyaml:jar:1.19' \ 'org.mockito:mockito-core:jar:2.13.0' \ 'io.netty:netty-all:jar:4.1.19.Final' \ 'org.eclipse.mylyn.github:org.eclipse.egit.github.core:jar:2.1.5' \ 'org.littleshoot:littleproxy:jar:1.1.2' \ 'org.slf4j:slf4j-jdk14:jar:1.7.25' ``` The version of LittleProxy we use is a snapshot and is the old version from previously with dependencies updated.

jsf-clabot · 2018-02-11T00:12:04Z

Thank you for your submission, we really appreciate it. Like many open source projects, we ask that you all sign our Contributor License Agreement before we can accept your contribution.
1 out of 3 committers have signed the CLA.

✅ TikhomirovSergey
❌ shs96c
❌ barancev

mykola-mokhnach · 2018-02-11T07:33:57Z

+
+ /**
+ * This is the flag which forces server to switch to the mobile WSONWP.
+ * If {@code false} when it is switched to W3C mode.


mykola-mokhnach · 2018-02-11T07:36:57Z

+ private final FileBackedOutputStream backingStore;
+ private final ImmutableSet<Dialect> dialects;
+
+ private static List<String> getAppiumCapabilities(Class<?> capabilityList) {


List<String> or List<Object>?

@mykola-mokhnach
List<String>
privatestaticList<String> getAppiumCapabilities(Class<?> capabilityList) { returnArrays.stream(capabilityList.getDeclaredFields()).map(field -> { field.setAccessible(true); try { returnfield.get(capabilityList).toString(); //<-!!!! } catch (IllegalAccessExceptione) { thrownewIllegalArgumentException(e); } }).filter(s -> !FORCE_MJSONWP.equals(s)).collect(toList()); }

mykola-mokhnach · 2018-02-11T07:39:22Z

+ // We need to convert the capabilities into a new session payload. At this point we're dealing
+ // with references, so I'm Just Sure This Will Be Fine.
+ boolean forceMobileJSONWP = ofNullable(caps.getCapability(FORCE_MJSONWP))
+ .map(o -> {


you could use ternary operator here

mykola-mokhnach · 2018-02-11T07:39:52Z

+ }
+
+ public static NewSessionPayload create(Capabilities caps) throws IOException {
+ // We need to convert the capabilities into a new session payload. At this point we're dealing


mykola-mokhnach · 2018-02-11T07:42:03Z

+ return new NewSessionPayload(source, forceMobileJSONWP);
+ }
+
+ private NewSessionPayload(Reader source, boolean forceMobileJSONWP) throws IOException {


This copies a lot of stuff from the original selenium code. Is it somehow possible to avoid such duplication?

for example using reflection

mykola-mokhnach · 2018-02-11T07:43:16Z

+ // Opera: operaOptions
+ // SafariDriver: safari.options
+ //
+ // We can't use the constants defined in the classes because it would introduce circular


Do we need all the commented stuff above?

mykola-mokhnach · 2018-02-11T07:52:35Z

+ public String getKey() {
+ String key = stringObjectEntry.getKey();
+ if (APPIUM_CAPABILITIES.contains(key) && !forceMobileJSONWP) {
+ return "appium:" + key;


appium: might be a constant

mykola-mokhnach · 2018-02-11T07:53:12Z

+ .map(this::applyTransforms)
+ .map(map -> map.entrySet().stream()
+ .filter(entry -> {
+ if (forceMobileJSONWP) {


ternary operator might be more useful here

.filter(entry -> forceMobileJSONWP && ACCEPTED_W3C_PATTERNS.test(entry.getKey()) || !forceMobileJSONWP)

mykola-mokhnach · 2018-02-11T07:53:49Z

+ return Stream.concat(fromOss, fromW3c).distinct();
+ }
+
+ private Map<String, Object> convertOssToW3C(Map<String, Object> capabilities) {


mark as Nullable

mykola-mokhnach · 2018-02-11T07:54:32Z

+ name = input.nextName();
+ if ("alwaysMatch".equals(name)) {
+ return input.read(MAP_TYPE);
+ } else {


this else seems redundant

mykola-mokhnach · 2018-02-11T07:54:43Z

+ name = input.nextName();
+ if ("firstMatch".equals(name)) {
+ return input.read(LIST_OF_MAPS_TYPE);
+ } else {


mykola-mokhnach · 2018-02-11T07:54:56Z

+ return null;
+ }
+
+ private Collection<Map<String, Object>> getFirstMatches() throws IOException {


mykola-mokhnach · 2018-02-11T07:55:05Z

+ return toReturn;
+ }
+
+ private Map<String, Object> getAlwaysMatch() throws IOException {


mykola-mokhnach · 2018-02-11T07:56:04Z

+ backingStore.reset();
+ }
+
+ private Map<String, Object> getOss() throws IOException {


mykola-mokhnach · 2018-02-11T07:56:23Z

+ String name = input.nextName();
+ if ("desiredCapabilities".equals(name)) {
+ return input.read(MAP_TYPE);
+ } else {


else is redundant

mykola-mokhnach · 2018-02-11T12:09:21Z

+ "Illegal key values seen in w3c capabilities: " + illegalKeys);
+ }
+ })
+ .forEach(map -> {});


what is the purpose of this operator?

mykola-mokhnach · 2018-02-11T12:27:40Z

+ }
+
+ // Write the first capability we get as the desired capability.
+ json.name("desiredCapabilities");


I'd rather put all there magic strings into constants

- removal of redundant code from the original NewSessionPayload

TikhomirovSergey · 2018-02-11T22:19:24Z

@SrinivasanTarget Could you try it on iOS?

mykola-mokhnach · 2018-02-12T06:26:15Z

- /**
- * Which mobile OS platform to use.
- */
- String PLATFORM_NAME = "platformName";


Where this constant is present now?

in org.openqa.selenium.remote.CapabilityType
publicinterfaceMobileCapabilityTypeextendsCapabilityType

mykola-mokhnach · 2018-02-12T06:28:22Z

+
+ public static NewSessionPayload create(Capabilities caps) throws IOException {
+ boolean forceMobileJSONWP = ofNullable(caps.getCapability(FORCE_MJSONWP))
+ .map(o -> {


.map(o -> Boolean.class.isAssignableFrom(o.getClass()) && Boolean.class.cast(o))

mykola-mokhnach · 2018-02-12T06:31:20Z

+ String name = input.nextName();
+ if (DESIRED_CAPABILITIES.equals(name)) {
+ return input.read(MAP_TYPE);
+ } else {


else is redundant

TikhomirovSergey · 2018-02-12T23:00:41Z

Updated.
@SrinivasanTarget does it work fine on iOS?

SrinivasanTarget · 2018-02-13T03:49:54Z

+ return null;
+ }
+
+ private @Nullable Collection<Map<String, Object>> getFirstMatches() throws IOException {


getFirstMatch

mykola-mokhnach · 2018-02-13T06:17:14Z

+ name = input.nextName();
+ if (FIRST_MATCH.equals(name)) {
+ return input.read(LIST_OF_MAPS_TYPE);
+ } else {


else is redundant. same comment to other similar blocks

mykola-mokhnach

minor comments

SrinivasanTarget

Everything looks fine on iOS on both W3C and JSONWP

shs96cand others added 10 commits February 9, 2018 22:43

Better docs for NewSessionPayload.stream
3687731

Fix the Protocol Handshake
b67cfdc

Rookie mistake: only include the w3c capabilities in the w3c capabili…
9a34f41
…ties

Fix tests failing because of ProtocolHandshake
f860f33

Ensure that we forward metadata in the new session payload
ed5fba8

Introducing new string constant for "platformName"
67588ec

Use Capabilities rather than ImmutableCapabilities in mutators
fbb6e58

Update to selenium 3.9.1. Changes of NewSessionPayload.
1855291

TikhomirovSergey requested review from SrinivasanTarget and mykola-mokhnach February 11, 2018 00:12

mykola-mokhnach reviewed Feb 11, 2018
View reviewed changes

TikhomirovSergey added 2 commits February 12, 2018 00:55

fixing out:
a79c624
- removal of redundant code from the original NewSessionPayload

Removal of "magic" strings
c23e05b

appium deleted a comment Feb 12, 2018

mykola-mokhnach reviewed Feb 12, 2018
View reviewed changes

mykola-mokhnach mentioned this pull request Feb 12, 2018
iOS getAlertText not behaving as expected (XCUITestDriver + Java Client) appium/appium#10111
Closed

TikhomirovSergey added 2 commits February 13, 2018 01:52

Improving set of not so major issues
b694856

convertOssToW3C marked by Nullable
afa1039

appium deleted a comment Feb 13, 2018

SrinivasanTarget reviewed Feb 13, 2018
View reviewed changes

mykola-mokhnach reviewed Feb 13, 2018
View reviewed changes

mykola-mokhnach approved these changes Feb 13, 2018
View reviewed changes

SrinivasanTarget approved these changes Feb 13, 2018
View reviewed changes

Fixing issuies before the merging
6ffd2ea

appium deleted a comment Feb 13, 2018

TikhomirovSergey merged commit 409e6f3 into appium:masterFeb 13, 2018

Uh oh!

Uh oh!

Conversation

TikhomirovSergey commented Feb 11, 2018

Change list

Types of changes

Details

Uh oh!

jsf-clabot commented Feb 11, 2018• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mykola-mokhnachFeb 11, 2018• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

TikhomirovSergeyFeb 11, 2018• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

TikhomirovSergey commented Feb 11, 2018

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

TikhomirovSergey commented Feb 12, 2018

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mykola-mokhnach left a comment

Choose a reason for hiding this comment

Uh oh!

SrinivasanTarget left a comment

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

jsf-clabot commented Feb 11, 2018•
edited
Loading

mykola-mokhnachFeb 11, 2018•
edited
Loading

TikhomirovSergeyFeb 11, 2018•
edited
Loading