selenium* jobs are a new way of running Ruby based Selenium tests. The
jobs are triggered daily.
Bug: T128190
Change-Id: If0e6b61c9e3d52005d1851ca49ef65babd002919
Some of the code encountered accessed mw.config directly,
I cleaned that up by migrating to mmv.Config, which is an
abstraction added to avoid peeking at mw.config directly.
Bug: T132064
Change-Id: I2a95ba703e6c7f46296f8e435bceec769dceebf9
The browser test job uses 'www.mediawiki.org' to hit the API which is
then passed to cucumber for tag filter.
Update the scenario tag to match the canonical URL.
Bug: T128079
Change-Id: I68217b3e9867b768ccbba24bbffaa431c1c3035f
The job was failing with `301 Moved Permanently`. The problem was that
`mediawiki_url` was not `https` and was missing `www`.
Bug: T128079
Change-Id: I4f142c4ed17826b1e4051846395940e139e47689
We used to rely on the env variable BROWSER_USERAGENT for tests
requiring a customer user agent. 2c3ac55a6 got rid of it in favor of
loading the :browser_useragent from environments.yml where it has not
been set. That in turns caused tests to fail with:
"missing configuration for browser_useragent"
Set browser_useragent to 'test-user-agent' in the environments.yml
Also drop unused test2 env and add mediawiki-org
Bug: T128064
Change-Id: I720d590505dd5e30c6e9435de456563b8d36f9ea
Additional changes:
* Add a step to check JS has loaded
* Update selector to work on non-vagrant instances
Bug: T113620
Change-Id: Ib7579a43adc37b6e4add8b0ab75f99b6ab994bb6
Drop the title and site name. Use a shortened URL
created using the pageID for the curid query parameter.
Add pageID property to mmv.model.Image, using value already
fetched from an existing API request.
Bug: T119686
Depends-On: I0f00b986e6095bdb9b8d6af6fbc5b01995227e02
Change-Id: I9189f00bfbf7639974d66cf1921773855a77a865
On not-MMV -> not-MMV hash changes the bootstrap class has set
up and immediately torn down the overlay. Besides slowing things
down, this broke TOC navigation on Chrome in an ugly way.
Add a special case to skip the overlay loading when the new hash
does not contain #/media. This is a quick-and-dirty fix; the whole
loading and hash handling could use a rewrite.
Bug: T119854
Change-Id: I5494903dfe778e533773ff142fdf756475c2df32
Having many modules makes the startup module in the head
unnecessarily longer. Things should only be in their own
module if they're going to be loaded on demand on their
own or are a shared dependency of separate modules that
are.
This change brings down the amount of modules declared by
Media Viewer from 53 to 8. The ones that remain are
mostly things loaded on demand.
Bug: T103706
Change-Id: I5b0d29209599285b93217e96def818e641646c73
It decreases the likelihood of Varnish hits and is only 11% smaller on
average in file weight than the 800 bucket.
Bug: T102870
Change-Id: I8611791a192fa7b0e638ccfc886733b75bf30aed
These two tests fail on my local machine. I am making the two
assertions independent from
a) the order of HTML attributes (the test should not care about
the order, only if the attributes are there or not)
b) the message (both messages are possible and valid, the question
which message should appear is not part of that specific test).
Change-Id: I69b4793d365d3afebb869c787316e4242cb15b25
Should fix intermittent issues happening with the browser tests,
which can't currently tell the difference between the placeholder
and the final image.
Bug: T90589
Change-Id: Ibfdff4721ce5f37831863110ad1a2f9650b24d5e
In practice doing this quickly still works, but the browser test in IE
catches the menu at exactly the wrong time (when it appears briefly
before disappearing).
Bug: T89087
Change-Id: I4076943bf33a566ec882f69d17f6a3a9f1a28002
ImageData will now parse for restrictions (this part has not been
implemented in CommonsMetadata yet), and an orange trademark label
will be displayed next to the license label if there is a 'trademarked'
restriction.
Bug: T77717
Change-Id: Ib03f9708d1e4ff0b5befddc2688b274e2c7ce1f7
Added mmv.HtmlUtils.htmlToTextWithTags()
which is similar to htmlToTextWithLinks()
but allows <b> and <i>
Added test for mmv.HtmlUtils.htmlToTextWithTags()
Most text fields now use htmlToTextWithTags()
except Permission field which is not supposed to
have HTML
Bug: T69887
Change-Id: I16833f218e2f64586aa13925356fa2b8b7ec3100
Pass alt parameter from mmv.bootstrap.js to mmv.js and
set it as a parameter on the displayed lightbox image.
Include the alt text in the embed text.
Bug: T66519
Bug: T75923
Change-Id: I29503eb582ac2bc8cf89f737a3bcb787b660d918
MediaViewer now handles Template:Multiple_image. Instead of looking
for caption in whole thumbnail container, it tries to find the
closest one to the image.
Bug: T85354
Change-Id: I18d982a4bf245c4925213d83a3410274d499845e
metadataPanel overrides the grow() and shrink() methods in Permission
class instance, so the text is changed also when user clicks the
"view more" link inside the box.
Bug: T71233
Change-Id: I66fe57980c6f469d86e3d52b67d01e06a3a14270
- Adds attribution variable to Image model
- In mmv.ui.metadataPanel and mmv.EmbedFileFormatter, display
attribution line instead of author and credit when it is set
- Update junit tests for mmv.model.Image and mmv.EmbedFileFormatter
Bug: T67445
Change-Id: Idfe542a1542d28cf8d27c1720ab0bd54324b2f37
This should be pretty straightforward
* use RSpec3 expect() syntax
* prefer string to regex as argument to match() when appropriate
* use Selenium $browser.back instead of inline javascript
Change-Id: I2e214752224397859a00f11bf9acc5bdc7685464
Add back a sleep to make the test pass
* make a Then statement with no assertion into a When
* remove Then statements assertions from Given steps
* prefer single quotes per rubocop
* use strings instead of regexes where appropriate for match operator
* use appropriate wait_until syntax
* remove redundant steps that call a single other step
* sort invisible selection criteria from steps to Feature input
* add when_present polling because test is now much faster
Change-Id: I24435362836945833942e87d7caee49c9b34368d
String.replace interprets a plain string as an exact text to match
against, but String.match interprets it as a regexp. Javascript is
fun.
Change-Id: I93ac31079f1f87dffdf0d5e02df2fa38ae79a6cc
* use the same text for plain and HTML format
* use the same text whether or not a deed URL is set
Bug: T76030
Change-Id: Iea53766b2e2031206d707838551bd82dba29e8c4
This is complete, but it would be better if the HEAD request
was actually aborted by Varnish when the viewDuration parameter is
present, or if the hit pointed to a script that does that.
Change-Id: I66cafd97427756411e967de1901324af2215e3ae
Mingle: https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/1001
This will be needed by Erik Zachte for compiling per-file image view data.
Since Media Viewer does preloading, it skews the HTTP request-based
statistics. By marking image requests coming from Media Viewer,
it allows us to remove that bias.
Change-Id: Iac8e7770c1a379691547de4b6d47b7d54467f54d
Mingle: https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/1002
Also fixes popup texts which should have been updated in the parent commit
and the panel opening mechanism which did not work when the below-the-fold
description box was display:none-d.
Change-Id: I5fc37429a7dc7a8f1517fa42375cd140d7d11674
Mingle: https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/589
* when the caption is available, use it as the image title
* otherwise, fall back to description if available
* otherwise, fall back to filename
* if both caption and description are available, show caption as
title and show description below the fold; otherwise don't show anything
below the fold.
Change-Id: I9738ca46620d7cf3b30f6e46790a37966d4b9eaf
Mingle: https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/589
* make top of "below-the-fold" be actually above the fold
* put all above-the-fold content except title and more details button
into that area
* kep below-the-fold area in fullscreen mode when title/credit gets untruncated
* increase title length to 180 (140 before reducing font size) - works nicely on
my 1600x900 laptop
* remove old styles which were intended to make the button smaller when the
panel is close but were mostly broken anyway and only caused the button to
twitch
To avoid redoing all the size calculation logic, we cheat and use a negative
margin to pull the below-the-fold content above the fold.
Change-Id: I18d7bdb7dbbdfb8201c0d66257731febfac31263
Mingle: https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/833
* move license and terms link from title area to attribution area
* remove some ineffective styling for elements which have already been removed
in earlier changes
Change-Id: I5bd41e337a4c8133bfa27e4c4d03162b2d6e35bb
Mingle: https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/833
- Fixes the bug where the options menu wouldn't open if media
viewer was reopened
- Adds human-readable console messages for the options actions
- Makes the tooltip a tipsy one like all the other buttons
- Fixes the tooltip copy, it was referring to "previews", which we
said we wouldn't do anymore, in favour of "Media Viewer"
Change-Id: I1590f9501cc6c406cc1f466b414062d4d6435c1b
Mingle: https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/963
Fixed up feature indentation and clauses and moved global methods into
the page-object classes.
Added @custom-browser tag to avoid starting multiple browser sessions
and updated mediawiki_selenium dependency for fix to custom browser
instantiation with SauceLabs.
Change-Id: I9741afb8d13ee8729f72210eb7617d9446311a8e
Due to a jQuery bug, errors in local code (gadgets, user scripts)
can cause onready handlers to not be executed. For MMV this causes
catastrophic failure, with a black screen of death on exit.
This change makes sure that the setup code necessary for Media
Viewer to work is executed at latest when MV is invoked, even if some
onready handlers were skipped.
Opening MediaViewer via a hash-URL will still not work if the onready
handler fails, but that's hard to avoid and it is not a catastrophic
failure anymore. This change can be reverted when bug 70772 gets fixed.
Bug: 70756
Change-Id: Ida3b780791bc9dfec29303567d33e3aa4f44dd81
Mingle: https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/891
Also add tags to run all tests in beta and test2. Let's see
what happens and remove them later if necessary.
NOTE: one test had no tags and ostensibly should run only on mw.o
We may have to remove them later, but let's see what happens for
now
Change-Id: I9c4adc7803afc975d9643a35631801bbac459eb7
Experimental revert of commit
d6c142fd2b to see if it caused
the weirdness in the versus charts:
http://multimedia-metrics.wmflabs.org/dashboards/mmv#media_viewer_vs_file_page-graphs-tab
That commit affecting file page load times is extremely unlikely.
A change in the test environment which exactly coincided with it is
equally unlikely. This is an easy way to tell which is the case.
Change-Id: I3581abd36d25494ca294db892480e25f6f4c5a73
This works because the title doesn't exist if there's no caption and we
won't get to this logic branch if the thumbnail is an explicit |thumb|
with a caption already.
Refactored caption-fetching a bit.
Change-Id: If84c890e7b71880db640a0993f8e3d6cd59951b8
Mingle: https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/513
Using the "complete" property worked for the file page, but not
for media viewer because change the src of an img doesn't reset
the prop.
This introduces a custom event that cucumber can listen to in order
to know when the full resolution image had been loaded in media
viewer.
Change-Id: I40875166e70badbb35106c4a3536c706a7c815b4
Ensure that changing the preference via the quick link and
Special:Preferences is stored identically in the DB.
Change-Id: Ia37da1c6bfbb3edf0eba56f01105e4a5f3a5e4ba
mw.notify calls mw.loader.using now, which we mock out in some
tests. This change adds an exception to the mock.
Change-Id: I68d7293a4d45f904e949ea4b56ec409456624cbd
Changes document title (which is shown in history navigation)
to include the image name.
Bug: 67008
Change-Id: Id1b030f2b984571fb0877e35db2ca2ccc86f0130
Automatically reveal/hide full text as metadata panel is opened/closed.
Also makes metadata open/close a proper event.
Changes scrollTop calculation for the fully open metadatapanel so it is not
confused by the size of the panel changing.
Also rename MetadataPanel.$controlBar to $aboveFold (that field was missed
when the corresponding CSS class got renamed).
Change-Id: I7e66ca0f45c2188dab4b78508ad7f91154187de4
Mingle: https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/396
Reveal the full title + author + source when the user clicks one
of these, and make some related changes:
* expand the above-the-fold part of the metadata panel so they fit
* make the stripe buttons smaller and hide their text
* scroll the panel up if necessary
* modify tooltip texts when there is more text to show
Change-Id: I304297bc5e7be7b16e2fc4bde66ac19641b00029
Mingle: https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/396
When formatting the attribution as plain text, make sure that the
source parameter (which typically ends with an URL) is not followed
by a '.', which could break autolinking when e.g. the text is
copied into a mail.
Change-Id: I5406fe3a55bc4ef6802b32d5a3d533ce9f49b361
We should probably have a proper test image eventually, so that
we can use e.g. image dimensions to double-check the results, but
this way at least our tests won't get broken by unrelated changes
in other repos.
Change-Id: Iffffd94fde99c53da59217a5181ad0560fce314c
- rearrange DOM structure of above-fold part of the metadata panel:
- rename .mw-mmv-controls to .mw-mmv-above-fold
- the above-fold part is a single positioned div now, with height
explitcitly set
- less LESS gymnastics, above-fold height is a single variable
- add paddings to the p elements instead of the containers
- make all title elements align to baseline (except the logo which
would look horrible)
- discard some CSS which was superfluous
- overspecified sizes/positions
- some top/bottoms for staticly positioned elements
- get rid of the .mw-mmv-drag-affordance div, since a full-width bar
wouldn't really make sense on the bottom of the above-fold section
- flip the chevron and place it to the bottom of the above-fold part;
add colors etc. per spec
- fix stripe button horizontal spacing
Mingle: https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/706
Change-Id: Ic37b4150288055c3fae8d22919ed7b1249db1f09
Several people complained that "Learn more on Wikimedia Commons"
is not clear: it can be interpreted as a link to the Commons
main page, the text does not suggest that the link will be related
to this specific file. The new wording tries to address this.
Change-Id: Ia605cc30c4ca57598f7cebdde60061800a10b6e7
Instead of setting the parent's height as max height of the
<img> element, find the first parent which has a non-automatic
height (that would be .mw-mmv-image-wrapper).
With the old structure, the height of the parent element could
be determined by the height of the image, which would then be
written back into the max-height of the image, messing up the
aspect ratio. I did not see this in the wild, but it was easy
to reproduce by changing the timing of the resize handler (in
particular, I tried to call the resize handler before loading
the new resolution, to make the UI more responsive, and ran
into this problem). This cannot happen anymore now.
This also fix a bug on some browsers (IE 10, maybe iOS Safari)
where the size of the image could be slightly larger than the
available space, and the bottom of the image was obscured by
the metadata panel. I am still not sure how exactly that
happened, but it was related to the <img> parents with automatic
heights having incorrect height. After making sure the <img>
has a max-height derived from an element with non-automatic
height, I cannot reproduce the bug on IE 10 anymore.
Change-Id: I193aefc42e6d6072717643659a9e4c0c8b7c7e93
Mingle: https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/698
Bug: 66244
This includes images where the <img> element has that class
(achievable with [[File:Foo.png|class=noviewer]]) and also those
where some parent element has it.
Change-Id: I666be026828ea9ecb6e8c93d3f5ad1e3c190f81e
Mingle: https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/511
Previously 'up' brought the panel up, and 'down' brought it down,
which might conflict with expectations on scrolling. Up/down keys
now move the metadata panel to the opposite state, no matter what
the current state was.
Mingle: https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/697
Change-Id: I53079d81042afb86354bf44e9dfd662adf1576cb
The jQuery update broke onDomEvent('focus') in OOjs UI. This is a
workaround which fixes the issue by binding on the input/textarea
elements directly, instead of their parents.
This introduces the annoying side effect that the metadata panel jumps
a bit when the embed HTML text is selected. (Just for that one, yes.
Weird.) Still better for now than no selection at all.
Change-Id: Ifa4c0600d7b4c0c64487596cbcabd5b4f4a12a19
Mingle: https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/638
http://foo.com/bar was making a request to some random server
responding with HTTP 500 Internal Server Error.
Change-Id: I17f2e0908b849455db5ab1790b15c2344337c24b
Has issues:
- needs dependency injection
- needs to be DRY
- should not load mw.eventLogging when we are not in sample
Due to urgency this will be fixed in another commit.
Change-Id: I0df067a619109a7c945f82c8d33fa2e621217f0b
Mingle: https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/619
Per the design meeting, we are abandoning this feature for now - there is not enough time to ensure it is of acceptable quality.
This reverts commit 4329d453ec.
Change-Id: I27c113ffecb617d442557163722ea5181ed0b2f4
jQuery 1.9 changes how $.focus() calls are handled: instead of
directly calling the handlers, it just invokes the DOM element's
focus(), and leaves it to the browser's event handling to trigger
them. This can fail for several reasons (e.g. element is not
attached to document, element is already focused, browser bugs such
as http://bugs.jquery.com/ticket/13363 ), so we are using
triggerHandler('focus') instead, which calls the handlers directly
without simulating actual browser events. Since these are unit
tests verifying event handler attach/unattach behavior, not
acceptance tests verifying actual event handling behavior, that
should be okay.
Change-Id: I65ecda28ace4f380ad33d6212e12069e18001232
Just a link to the full-size file for now.
Since the link must be to a PNG/JPEG/GIF (so possibly a thumbnail),
and we want to cap the size, we might need to get the URL from the
API, but we need to open the new window right away to avoid popup
blockers, making this patch quite complicated.
Change-Id: I9ce9d2a2d27b75470eae2806d9f9ce2f95f4dac2
Mingle: https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/588
