following_should_download_between_and_bytes step sometimes
fails because of server speed. Added extended timout
ensuring they wait enough before failing.
Also, not calling force download while running behat site
following_should_download_between_and_bytes step sometimes
fails because of server speed. Added extended timout
ensuring they wait enough before failing.
Also, not calling force download while running behat site
We already had the corresponding
I click on "Home" "link" confirming the dialogue
but I needed the opposite
I click on "Home" "link" dismissing the dialogue
Check that page load detection was correctly started before testing that a
new page was loaded.
Without this, it is possible to have mutliple subsequent cases of:
And a new page should have loaded since I started watching
Without first starting the page load detection.
The following improvements have been made to the page load watching:
* Improved the exceptions when a page load expectation fails.
* Added an exception if start watching happens twice without a page load.
* Improved the page load span and xpath to make it faster and less likely
to interfer in the future.
This fix is mostly based on what Colin Chambers found out. This commit
is a simplification of his work.
The problem is that the Chrome / Selenium 2 integration cannot swich to
a window with a blank name. The work-around applied here is, when we
switch away from an unnamed window, we set a name on it. Then we can use
that name to switch back.
These steps were created by sam marshall and Jes Ackland-Snow at the Open
University. I just altered them to make them a bit more generic, and
submitted them to Moodle core, at the point when I wanted to reuse them
in another plugin.
This commit introduces two new steps:
Then following "XXX" should download "YY" bytes
Then following "XXX" should download between "YY" and "ZZ" bytes
At the moment, this only works for links with href attributes.
Having a I click on we can use all our selectors rather
than having multiple steps for the same purpose. This
step allow us to manage JS native dialogues without
having to fail a scenario.
Credit for Mohamed Alsharaf.
The behat script is to test the batch action to lock and unlock submissions.
The new definitions are to prevent the "unexpected alert open" error in Selenium.
The step checks that an attribute exists, but doesn't contain the
specified text.
Example:
Then the "class" attribute of "#mydiv" "css_element" should
not contain "accesshide"
This commit creates a Behat step of the form:
Then the "title" attribute of "Toggle visibility" "button"
should contain "Show"
This is useful because unless I missed something there is currently
no step that checks attribute values; this seems like a simple
generic step. Behat is intended to check user-visible effects in
the resulting HTML, and sometimes these effects are implemented
using attribute values (two common examples are the alt= attribute,
which is visible to screenreader users, and the title= attribute,
which is visible to all users; there's also the src attribute of
images, and lots of other possibilities too).
Run all behat with medium (1024x768) screen size to
avoid failures with small screen size. Also, added
feature to let scenario chnage screen size if need be.
There are steps where we want to check that
"something" does not appear in the page
or does not exist. We still have to spin
but we don't need to spin for 6 seconds.
* Expanding fieldsets before setting the course format
* Ensure editors are loaded moved to expand_all_fieldsets
* Return true rather than false when JS is disabled and
non-allowed steps should continue without throwing an
exception.
* Submit create user form after editing it
* When looking for texts inside the page or inside
other containers we should wait until the elements
are visible.
* Same when expanding tree nodes.
* Normalizing loops to spin() function using
behat_base::TIMEOUT and behat_base::EXTENDED_TIMEOUT,
leaving TIMEOUT for DOM load processes and
EXTENDED_TIMEOUT for long processes that involves JS
too.
* Add page load waits between actions that involves
reloading the page.
In general aiming for compatibility with multiple browsers,
firefox, chrome and phantomjs to be more specific.
* Removing hardcoded waits
* Adding @_alert, @_switch_window and @_switch_frame tags,
to label actions that different drivers have problems with.
* Adding missing @_files_upload and @_only_local tags to features that
uploads files.
* Fixing a few wait for page ready what specified miliseconds.
* New methods to ensure elements (usual selectors), sections and editors
are ready to interact with
* Changing the select an option implementation to deal with the different
drivers implementations when listening to JS events.
1. Added new step definitions to check if an element exists or does not exist in another.
2. Added new step definitions to check if an element is readonly or is not readonly.
3. Automated testing for MDLQA-1812 and MDLQA-1813.
Added behat test for deleting groups/groupings with and without the
'moodle/course:changeidnumber' capability.
4. Automated MDLQA-2105.
Added behat test for creating groups/groupings without the
'moodle/course:changeidnumber' capability.
5. Automated MDLQA-2238 and MDLQA-2335.
Added behat test for updating groups/groupings with and without the
'moodle/course:changeidnumber' capability.
Using 'should see' outputs rather than an incorrect
xpath query. Also our custom selectors are not properly
used by WebAssert so switching to find xpath.
- Escaping steps arguments redirected to other steps
- Adding normalized-space() in all contains() assertions
- General xpaths review
- Convering provided xpath text strings to xpath literals
to avoid problems with arguments containing both single
quotes and double quotes
Is another of the exceptions thrown by
Selenium when it tries to interact with
an element that disappeared from the DOM,
in happens randomly when redirecting the
user to another page with a wait period to
show him/her a message.
Andrew Nicols