Fix pattern unlock method

[IgalKorb]

added the use of the last position (and now the actions are relative to the previous rather than relative to {0,0}

[jsf-clabot]

All committers have signed the CLA.

[mykola-mokhnach]

@vrunoa Can you please check the PR?

[mykola-mokhnach]

@IgalKorb Can you please update the unit test as well?

[IgalKorb]

Well, I'm not a js developer, I use appium with python API, so I'd say I prefer you pros to do a fix and unittest how you know best, I just noted a bug and a possible solution

…

On Sat, Jan 19, 2019, 11:38 Mykola Mokhnach ***@***.*** wrote: @IgalKorb <https://github.com/IgalKorb> Can you please update the unit test <https://travis-ci.org/appium/appium-android-driver/jobs/480438161> as well? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#489 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AKdwfDm5R2ZHRUYq4mtv3NR_ZkYj3W2Hks5vEuelgaJpZM4aDUjX> .

[imurchie]

No comment on the change itself, since I'm not sure about it. But the test can be fixed with the following:

diff --git a/test/unit/unlock-helper-specs.js b/test/unit/unlock-helper-specs.js
index 4deb609..1bbdb70 100644
--- a/test/unit/unlock-helper-specs.js
+++ b/test/unit/unlock-helper-specs.js
@@ -309,33 +309,34 @@ describe('Unlock Helpers', function () {
         }
       });
     });
     it('should verify pattern gestures moves to non consecutives pins', function () {
-      let keys = ['7', '2', '9', '8', '5', '6', '1', '4', '3'];
-      let actions = helpers.getPatternActions(keys, {x: 0, y: 0}, 1);
+      const keys = ['7', '2', '9', '8', '5', '6', '1', '4', '3'];
+      const actions = helpers.getPatternActions(keys, {x: 0, y: 0}, 1);
+
       // Move from pin 7 to pin 2
-      actions[1].options.x.should.equal(1);
-      actions[1].options.y.should.equal(-2);
+      actions[1].options.x.should.equal(2);
+      actions[1].options.y.should.equal(1);
       // Move from pin 2 to pin 9
-      actions[2].options.x.should.equal(1);
-      actions[2].options.y.should.equal(2);
+      actions[2].options.x.should.equal(3);
+      actions[2].options.y.should.equal(3);
       // Move from pin 9 to pin 8
-      actions[3].options.x.should.equal(-1);
-      actions[3].options.y.should.equal(0);
+      actions[3].options.x.should.equal(2);
+      actions[3].options.y.should.equal(3);
       // Move from pin 8 to pin 5
-      actions[4].options.x.should.equal(0);
-      actions[4].options.y.should.equal(-1);
+      actions[4].options.x.should.equal(2);
+      actions[4].options.y.should.equal(2);
       // Move from pin 5 to pin 6
-      actions[5].options.x.should.equal(1);
-      actions[5].options.y.should.equal(0);
+      actions[5].options.x.should.equal(3);
+      actions[5].options.y.should.equal(2);
       // Move from pin 6 to pin 1
-      actions[6].options.x.should.equal(-2);
-      actions[6].options.y.should.equal(-1);
+      actions[6].options.x.should.equal(1);
+      actions[6].options.y.should.equal(1);
       // Move from pin 1 to pin 4
-      actions[7].options.x.should.equal(0);
-      actions[7].options.y.should.equal(1);
+      actions[7].options.x.should.equal(1);
+      actions[7].options.y.should.equal(2);
       // Move from pin 4 to pin 3
-      actions[8].options.x.should.equal(2);
-      actions[8].options.y.should.equal(-1);
+      actions[8].options.x.should.equal(3);
+      actions[8].options.y.should.equal(1);
     });
   });
   describe('patternUnlock', withMocks({driver, helpers, adb, asyncbox}, (mocks) => {

This also should be rebased against master.

[vrunoa]

@IgalKorb I'll try to review it by the end of the weekend.

[dpgraham]

@vrunoa did you have a chance to look at this?

[IgalKorb]

@vrunoa @dpgraham Could you please review it? :D

[dpgraham]

@imurchie @mykola-mokhnach do either of you have any objections to just merging this with the failing test and then I can apply @imurchie 's patch immediately after?

[imurchie]

No objections from me.