#7681 closed task (fixed)
[meta] commit dijit robot tests
Reported by: | Douglas Hays | Owned by: | |
---|---|---|---|
Priority: | low | Milestone: | 1.10 |
Component: | Dijit | Version: | 1.2beta |
Keywords: | Cc: | haysmark, Becky Gibson | |
Blocked By: | Blocking: |
Description
Use this ticket to drop the initial wave of robot automated tests for dijit.
Attachments (1)
Change History (169)
comment:1 Changed 12 years ago by
Cc: | haysmark added |
---|
comment:2 Changed 12 years ago by
comment:3 Changed 12 years ago by
comment:4 Changed 12 years ago by
Summary: | commit dijit robot tests → [meta] commit dijit robot tests |
---|---|
Type: | defect → task |
comment:5 Changed 12 years ago by
comment:6 Changed 12 years ago by
comment:7 Changed 12 years ago by
comment:8 Changed 12 years ago by
(In [16667]) Test infrastructure updates, as discussed with Mark:
- Allow test page to listen to topics published by child page
- Use dojo.setContext() to point parent page to child page, so that dojo.global also gets redefined, in addition to dojo.doc... that lets test code easily access variables in the child by referencing dojo.global.foo, or perhaps more correctly by calling dojo.getObject()
Refs #7681
comment:9 Changed 12 years ago by
comment:10 Changed 12 years ago by
comment:11 Changed 12 years ago by
comment:12 Changed 12 years ago by
comment:13 Changed 12 years ago by
comment:14 Changed 12 years ago by
comment:15 follow-up: 16 Changed 12 years ago by
comment:16 Changed 12 years ago by
Cc: | Becky Gibson added |
---|
Replying to bill:
(In [16834]) Split slider tests into keyboard tests and mouse tests. Keyboard tests don't work on safari since we can't focus the handle of the slider, thus can't send keyboard commands to it. Refs #7681.
Just a note that keyboard tests *do* work for Safari 4 beta. (So far I've run the slider and tree robot tests).
comment:18 follow-up: 23 Changed 12 years ago by
comment:19 Changed 12 years ago by
comment:20 Changed 12 years ago by
comment:21 Changed 12 years ago by
comment:22 Changed 12 years ago by
comment:23 Changed 12 years ago by
Replying to bill:
(In [16840]) Fix robotTests.js to run the Menu test, and also to run a11y only on browsers that support them. Refs #7681.
While it makes sense to run keyboard a11y tests only in browsers that support them, the same is not true of the aria tests (roles, states, and properties).
The aria tests are checking that the dijit widget was properly constructed -- that it has the proper role, the required and supported states/properties. That's independent of the browser. The aria info will be set and manipulated by the object's code, even when run in browsers that do not support aria.
A solution is to parcel out the aria tests. For example, have a "Slider_mouse.html", a "Slider_keyboard.html", and a "Slider_aria.html".
comment:24 Changed 12 years ago by
comment:25 Changed 12 years ago by
comment:26 Changed 12 years ago by
comment:27 Changed 12 years ago by
(In [17637]) Initial robot tests for DateTextBox? and Calendar widgets. Working on IE6, FF3/PC, and Safari3/Mac, although failing on FF3/mac (but that's generally unreliable for robot tests).
Refs #7681 !strict.
comment:28 Changed 12 years ago by
comment:29 Changed 12 years ago by
comment:31 Changed 12 years ago by
comment:32 Changed 12 years ago by
comment:33 Changed 12 years ago by
comment:34 Changed 12 years ago by
(In [17970]) Some simple automated but non-robot tests for Tree, in particular testing Tree/dojo.data interaction. Perhaps should remove test_Tree_Programmatic.html and test_Tree_Notification_API_Support.html as Tree.html tests the same functionality.
Updating countries.json to separate id and name attributes just so I can test modifying the name of an item / label of a tree node.
Refs #7681.
comment:35 Changed 12 years ago by
comment:36 Changed 12 years ago by
comment:37 Changed 12 years ago by
comment:38 Changed 12 years ago by
comment:39 Changed 12 years ago by
comment:40 Changed 12 years ago by
comment:41 Changed 11 years ago by
comment:42 Changed 11 years ago by
(In [20153]) Safari and Chrome need more time for the click to take effect and toggle the ToggleButton?. Refs #7681.
comment:43 Changed 11 years ago by
(In [20441]) Get Spinner_mouse.html running on IE8, rather than hanging. Changed a bunch of tests that didn't (seem to) make sense, including:
- Multiple dojo.register("foo", ...) calls w/same "foo" name. Used array argument to dojo.register() instead.
- "0.5" is not an invalid value, changed test to type "0.a"
- "A11y" test group doesn't test a11y, that's what Spinner_a11y.html is for
- Various tests were mislabeld, setvaluetests and mousewheel tests were under the "arrowbutton" group, etc.
Refs #7681
comment:44 Changed 11 years ago by
comment:45 Changed 11 years ago by
comment:46 Changed 11 years ago by
comment:47 Changed 11 years ago by
comment:48 Changed 11 years ago by
comment:49 Changed 11 years ago by
comment:50 Changed 11 years ago by
comment:53 Changed 11 years ago by
(In [20725]) TitlePane? automated tests (refs #7681). Also some small updates to ColorPalette? tests.
comment:54 Changed 11 years ago by
comment:55 Changed 11 years ago by
comment:56 Changed 11 years ago by
comment:58 Changed 11 years ago by
comment:60 Changed 11 years ago by
comment:61 Changed 11 years ago by
comment:62 Changed 11 years ago by
comment:63 Changed 11 years ago by
comment:64 Changed 11 years ago by
comment:65 Changed 11 years ago by
comment:67 Changed 11 years ago by
comment:68 Changed 11 years ago by
comment:69 Changed 11 years ago by
comment:70 Changed 10 years ago by
(In [22970]) Added support for holding down the META key by itself. Forced mouseMove to stay within upper-left bound of window to prevent it from spuriously aborting tests.
Patch from Mark Hays (IBM, CCLA), thanks! Hope I compiled the jar correctly (it's working for me, and the Tree DnD tests are now running correctly on mac FF).
comment:71 Changed 10 years ago by
comment:72 Changed 10 years ago by
comment:73 Changed 10 years ago by
comment:74 Changed 10 years ago by
comment:75 Changed 10 years ago by
(In [23084]) dijit.dijit is not needed on robot control pages, but rather just on the controlled pages (i.e. the pages loaded through doh.robot.initRobot()).
Plus, requiring dijit.dijit causes "_destroyed is not an object" errors on page unload on IE, when dijit tries to destroy all widgets. dijit.findWidgets() is returning [null] rather than [] for some reason.
Refs #7681.
comment:76 Changed 10 years ago by
comment:77 Changed 10 years ago by
comment:78 Changed 10 years ago by
comment:79 Changed 10 years ago by
comment:80 Changed 10 years ago by
comment:81 Changed 10 years ago by
comment:82 Changed 10 years ago by
(In [23093]) Refs #7681. Replace some timing sensitive robot tests in InlineEditBox?.html with event-driven segments to eliminate seemingly random IE test failures (including Already Called exceptions).
comment:83 Changed 10 years ago by
comment:84 Changed 10 years ago by
comment:85 Changed 10 years ago by
comment:86 Changed 10 years ago by
comment:87 Changed 10 years ago by
comment:89 Changed 10 years ago by
comment:90 Changed 10 years ago by
comment:91 Changed 10 years ago by
(In [23140]) Refs #7681. Change ValidationMessages?.html delays from 150 to 250ms so they work on a really slow machine.
comment:92 Changed 10 years ago by
(In [23186]) Refs #7681. Add more events and less randomm delays to ValidationState?.html and ValidationMessages?.html to reduce both test size and time and to eliminate random failures.
comment:93 Changed 10 years ago by
comment:94 Changed 10 years ago by
comment:95 Changed 10 years ago by
(In [23258]) Refs #7681. Add some events to ProgressBar?.html to speed up the test and to reduce random failures.
comment:96 Changed 10 years ago by
comment:97 Changed 10 years ago by
(In [23262]) There's no reason for the ProgressBar DOH test to use robot because ProgressBar doesn't have any user interaction. Merging tests/test_ProgressBar.html and tests/robot/ProgressBar.html into a single tests/ProgressBar.html file.
Also, use dojo.hasClass() instead of custom code.
A lot of these tests are redundant; they aren't testing ProgressBar but merely the test page itself. I'll remove them in the next check in.
Refs #7681.
comment:99 Changed 10 years ago by
comment:100 Changed 10 years ago by
comment:101 Changed 10 years ago by
Changed 10 years ago by
Attachment: | timingfixes.patch added |
---|
Refs #7681. Reduce test timeout from 1 day to 30 minutes.
comment:102 Changed 10 years ago by
comment:103 Changed 10 years ago by
comment:104 Changed 10 years ago by
comment:105 Changed 10 years ago by
comment:106 Changed 10 years ago by
(In [23471]) Refs #7681. Workaround no keyup event sent for meta+RightArrow? in safari/mac.
comment:107 Changed 10 years ago by
comment:108 Changed 10 years ago by
(In [23474]) Intermediate check-in to combine href loading tests for StackContainer, TabContainer, and AccordionContainer, and to eliminate the timing issues in TabContainer-remote.html. !
ContentPane?.html still has timing issues. The problem is that getResponse.php?delay=1000 can take up to 5 seconds to return.
Refs #7681.
comment:109 Changed 10 years ago by
comment:110 Changed 10 years ago by
comment:111 Changed 10 years ago by
comment:112 Changed 10 years ago by
comment:113 Changed 10 years ago by
comment:114 Changed 10 years ago by
comment:115 Changed 10 years ago by
comment:116 Changed 10 years ago by
comment:117 Changed 10 years ago by
comment:118 Changed 10 years ago by
comment:120 Changed 10 years ago by
comment:121 Changed 10 years ago by
(In [24174]) Refs #7681. Slow InlineEditBox? test down by waiting for the editor's _onFocus instead of just postCreate.
comment:122 Changed 10 years ago by
comment:124 Changed 10 years ago by
comment:125 Changed 10 years ago by
comment:126 Changed 10 years ago by
(In [24187]) Refs #7681. Small timing change to InlineEditBox?.html test.
comment:127 Changed 10 years ago by
comment:128 Changed 10 years ago by
comment:156 Changed 8 years ago by
Priority: | high → low |
---|
comment:165 Changed 7 years ago by
Owner: | Douglas Hays deleted |
---|---|
Status: | new → assigned |
comment:166 Changed 7 years ago by
Status: | assigned → open |
---|
comment:167 Changed 7 years ago by
Milestone: | future → 1.10 |
---|---|
Resolution: | → fixed |
Status: | open → closed |
I don't think we need a meta-ticket for this anymore. If anything, in the future we'll switch to webdriver, and that surely deserves a new ticket.
(In [15280]) References #7681. The robot code should execute with the same BASE as the page being tested so that XHR gets work correctly (Safari 3.1 seems to require this).