Currently echo attempts to find a signature by looking for a series of
strings starting with what it thinks are the current aliases of NS_USER
and NS_USER_TALK. This has shown to be error prone, see the linked bug
for how a change to ru.wikipedia.org/wiki/Mediawiki:Signature broke
mention notifications.
Patch switches things arround to pull wikilinks out of the text and run
them through the Title class. The results of this parsing are checked
for NS_USER and NS_USER_TALK, giving a much stronger guarantee of finding
translated namespaces.
Bug: 71353
Change-Id: Ib0d0f4e068339d2fd28761087c05f5a1acb3c1fc
Has a performance impact, so this depends on deploying the job
that deletes older notifications.
Bug: 69919
Change-Id: Ia485c853d1b04c3c85e25e6a12f5060a046e9b11
The step formerly know as "I come back from grabbing a cup of coffee"
is replaced by a step that reloads the page a number of times, where
that number may be set by the user. If the notification has arrived,
the polling stops and the test continues.
In practice with a local browser pointed to beta labs, the
notification is arriving in plenty of time for the first page load
and I do not see the page reloading at all.
Change-Id: I95dd77d4f5525645e46244e40bfcd86e9124f3b1
Cucumber Scenario and Feature description are just plain text.
Instead of using comments, it is far better to describe in each
test step what should happen in each test step.
I believe I have described these test steps correctly. Please
advise if they need to be tweaked.
(The test itself seems to be fine, but feel free to review that
also)
Finally, since messages.feature uses the "I have no new notifications"
step, I updated those assertions for modern RSpec syntax
Change-Id: Ie884a85c1fc3f7045773a271a8d76fad6b0b3826
The max-height of the notifications list was hardcoded to
the height of the window minus 134 pixels. I don't know
where this value came from, but it doesn't seem to take
borders into account. The correct value appears to be
139 in Chrome and 140 in Firefox, so I changed it to 140.
Also, this really really should not be a hardcoded value.
This 140 value can be measured in JavaScript; that's how I
derived it. Even better, oojs-ui has a mixin called
OO.ui.ClippableElement that automatically does this for you
AND automatically adjusts the height when the window is resized
as well.
Change-Id: I17bc2c5333e5c3d5dd720e6bccf8cbbdbb4abe6c
"Then" steps should always contains RSpec assertions. In the
past this meant using some version of the word "should", but
the modern convention in RSpec is to use "expect()" as I
did here.
Change-Id: I1c438a310803e29c0985328a76d2ec08578e0cb0
* add @custom_browser tag so we don't create an extra browser instance
* remove @chrome tag because this only works in Firefox
* alphabetize within G/W/T buckets in commons_steps.rb
* alphabetize flyout_steps.rb
* bring env.rb up to date with the modern situation
* make assertions about the URL and heading on the final destination page
* update Ruby gems, especially mediawiki-selenium with a @custom_browser improvement
** use modern RSpec assert syntax
*** we're going to have to update all the repos for this soon
Change-Id: I6cb20126a662bd2a20e2b4bbebe91e8f3b04c7df
The "hover" pseudo-event was deprecated in jQuery 1.8 and removed in
1.9. This is logged as "JQMIGRATE: 'hover' pseudo-event is deprecated,
use 'mouseenter mouseleave'". This fix switches to using the .hover()
method which appears to be the original intent as two functions have
been used which .hover() supports, whereas .on() only accepts one
handler. Now the class change works as expected.
Change-Id: Ib28801293b72f8f344455b5f308876d185abc8bd
* use symbols not strings for element labels
* use double quotes except where single quotes required
Change-Id: I4eef50206c526611da535e529035c076d5bce266
I did `bundle update`.
Without this, browser tests on Firefox 32 fail with
`unable to obtain stable firefox connection in 60 seconds
(127.0.0.1:7055) (Selenium::WebDriver::Error::WebDriverError)`
Change I26190cca06ee gave Flow the same update, this Gemfile.lock is
identical except Echo has a "chunky_png" gem.
Change-Id: I0b62134d38f3262b734af83cc58b644040857645
so that we could skip parameters escaping when formatting email
subject. This is not a great solution but works for now.
Probably should re-think all the use case of echo formatting and
refactor the formatter
Change-Id: Ic1f6aa0f72d637558466db5cb2ee8d36038dd0b2
Some exception is not loggable, which won't be logged by
MWExceptionhandler::logException(), we should log it in
the debug log for debugging purpose
Change-Id: I709c45f7e47afacb98be3a869eb5b70660e18f48
