Changes:
- removed ugly PopupsContext::getInstance
- removed ugly PopupsContextTestWrapper helper
- defined all services inside ServiceWirings
- fixed unit tests to test classes directly or use MediawikiServices
Change-Id: Ie27e28bb07aebe01014848c290369b1b1f098e9b
Changes:
- introduced new UserPreferencesChangeHandler class that listens to
PreferencesFormPreSave hook
- introduced wrapper for EventLogging extension plus NullLogger when
EventLogging extension is not availalbe
- when user changes PagePreview to disabled system will trigger
disabled event
Bug: T167365
Change-Id: I63faecb0495eb30a9fc2763e7ddf3944baf7f55a
We had instrumentation for over 4 weeks and duplicate events rate
was very low. We want to keep stats so we check the duplicate events
rate but there is no need to filter those.
Bug: T167365
Change-Id: I72585beb21e9db589e45eeace657ef25f432abc9
Changes:
- introduced new event 'disabled', sent from settings popup
- added unit tests for 'disabled' event handling
Bug: T167365
Change-Id: I048b38122b8843199c86fd1ed9ec2ff21767e114
In order to debug the EventLogging instrumentation in production
environments, we want to be able to bucket ourselves at will.
When the debug flag (?debug=true) is passed send all events
for given page view.
Bug: T168847
Change-Id: Id1b13b0ecaa791b4f26be4d1151bdbbe5270b64d
Changes:
- when event doesn't have linkInteractionToken do not check for
duplicated tokens
- hygiene, move event duplication logic into separate functions
for better readability
Bug: T168449
Change-Id: I3ae197567ec9f67e104af109d4f1a1c1a6769d32
Following on from I4f653bba, since the schema and statsvInstrumentation
modules are similar, let's group/rename them:
schema -> instrumentation/eventLogging
statsvInstrumentation -> instrumentation/statsv
Change-Id: Ic59e0da7d4917f6733fd090f15d3c269af863f05
The following sniffs are failing and were disabled:
* MediaWiki.Commenting.FunctionComment.MissingParamComment
* MediaWiki.Commenting.FunctionComment.MissingParamTag
* MediaWiki.Commenting.FunctionComment.MissingReturn
* MediaWiki.Commenting.FunctionComment.ParamNameNoMatch
* MediaWiki.FunctionComment.Missing.Public
* MediaWiki.WhiteSpace.SpaceBeforeSingleLineComment.NewLineComment
Change-Id: Id9e98b7a4e87d00c63e7b509506c23f12cf0d380
Currently, the mw.eventLog.Schema class samples per pageview. However,
we expect that if a user is bucketed for a session, then all
EventLogging events logged during that session are in the sample.
Moreover, loading the class in the way that we did - asynchronously,
using mw.loader#using - introduced an issue where the eventLogging
change listener would subscribe in the next tick of the JavaScript VM's
event loop and miss the "pageLoaded" event being queued (see T167273).
Changes:
* Make the schema module follow the form of the statsvInstrumentation
module, i.e. make it expose the #isEnabled method, and add the
associated getEventLoggingTracker function.
* Update the eventLogging change listener accept the tracker returned by
getEventLoggingTracker.
* Update/fix related JSDoc documentation.
Bug: T167236
Bug: T167273
Change-Id: I4f653bbaf1bbc2c2f70327e338080e17cd3443d4
gateway/*/rest were copies of gateway/restProvider just passing
a different provider. Docs were the same, they were untested, and
looking at them they seemed like unnecessary abstraction.
This patch removes the plain vs html structure, and separates gateways
like before, by endpoint.
There is a light utility in gateway/restFormatters.js that adapts the
call from the rest gateway to use formatters.js functions. It needs
testing, that I'll add in the next patch.
The flow for creating a gateway ends up as follows:
1. index.js calls gateway/index#createGateway( mw.config )
2. createGateway chooses based on wgPopupsGateway and invokes
* mediawiki.js#createMediaWikiApiGateway or
* rest.js#createRESTBaseGateway w/ restFormatters.js#parsePlainTextResponse or
* rest.js#createRESTBaseGateway w/ restFormatters.js#parseHTMLResponse
Changes:
* Removed src/gateway/{plain,html}/rest.js
* Extracted formatter functions to src/gateway/restFormatters.js
* src/gateway/plain/mediawiki.js -> src/gateway/mediawiki.js
* tests/node-qunit/gateway/plain/mediawiki.test.js ->
tests/node-qunit/gateway/mediawiki.test.js
* gateway/restProvider{,.test}.js -> gateway/rest{,.test}.js
* Change gateway/index.js#createGateway to properly call the rest
gateways with the rest formatters
Bug: T165018
Change-Id: Ia75695dfc192aad5bc581a68882514bad6c29646
npm scripts are run under sh. globstar support is only available in bash
> 4 with globstar enabled, so **/*.js was actually being resolved to
*/*.js only running tests 2 directories deep.
This made the test mediawiki.test.js not run since it was moved to the
3rd level deep folder (and actually broke), resulting in:
208 tests
403 passed
With this change, it fixes getting the test files, and fixes the
mediawiki.test.js, resulting in:
215 tests
413 passed
instead.
This makes it work everywhere and as nested as it will get for now. Will
consider adding glob support to mw-node-qunit later for easier use in
other projects too.
Bug: T165018
Change-Id: Id0164b2673c8afe8a24fd0eb4aa255c134253862
And add tests, given it is growing in complexity.
Additional changes:
* Interface ext.popups.Gateway -> Gateway in docs
Bug: T165018
Change-Id: I8a12333ad9d14d6a7fbde11afc42f607881e8ea3
... from the statsvInstrumentation module so that the bucketing logic
can be shared with other instrumentation modules.
Change-Id: I5732fa539a3911939fa85fa88c102fa8dcfa5613
Refactor existing Restbase gateway and extract shared logic into
shared Restbase provider. Also introduced new createNullModel()
which defines an empty preview model.
Additionally improve naming in new gateways/formatter so function
names are more explicity.
* Htmlize() became formatPlainTextExtract() as it should be used
only with plain text extracts
* removeEllipsis() became removeTrailingEllipsis() as it removes
only trailing ellipsis.
* src/gateway/index.js defines gateways by configuration name stored
in extension.json
Bug: T165018
Change-Id: Ibe54dddfc1080e94814d1562d41e85cb6b43bfc1
Depends-On: I4f42c61b155a37c5dd42bc40034583865abd5d7a
This reduces the number of 301 Redirect responses when fetching previews
from RESTBase.
Bug: T167633
Change-Id: I830947ab79e72dcc023193412c8d5bcee986e23f
Page Previews should be able to consume HTML response generated by
MediaWiki. First we need to move out plain text crunching from
renderer.js and model.js. Mediawiki and Restbase gateways will have
to parse/htmlize plaintext into nice HTML by themselves.
Bug: T165018
Change-Id: I5d7e9f610bb809aa9fb035a4a9f96e9e8796c9d8
...instead of 1 event per link
Supporting changes:
* Delegate events on the container when booting up
* Check eligibility of title on event triggered
* Pass the title from the event handlers into the actions instead of
storing it in the dom
* Add title.fromElement as sugar over isvalid(gettitle())
* Not tested as it is sugar over the other 2 functions
* Fix action tests and integration tests
processLinks to be removed next
Bug: T165572
Change-Id: I4d9837706dc77ec64121ac94410c0d2da83692e4
... and update the RESTBASE_PROFILE constant to the latest "stable"
profile for the endpoint.
Prior to this change the Accept header sent by the rest gateway was
application/json; charset=utf-8profile="..."
This was discovered while responding to T166605.
Change-Id: I00f277e724c561634b26c9ab10bd35332c6dba91
When there's an interaction, then the "tapped settings cog" event should
have the same properties set as the other interaction-specific events.
This was discovered while QAing T164256.
Bug: T164256
Change-Id: I4749b52656203c7e0c42ae742556ee996eee322a
... and the previewType property as well.
Per the Popups schema [0], the "opened" action should have the
perceivedWait and previewType properties set.
Bug: T166323
Change-Id: I957d123434a6b750aff6f5279865321a08367382
Introduced PHPCS check in CI - using same configuration as in
MobileFrontend. Additionally fixed wrong code style.
Change-Id: I0c879553d355c2a277fcc4349a93e85c65eb2291
Given that interactions end up with an event logged, there shouldn't be
a reason to keep an interaction active after it's corresponding final
event has been logged. See Tbayer's state graph.
This patch removes the current interaction if an event with that token
is logged, effectively finalizing it and making it impossible for the
token to be reused from the state tree again.
Additional changes:
* Pass the logged event with the action EVENT_LOGGED so that the reducer
can determine if it needs to do anything else.
* Since the interaction is removed, when undefined, guard against
actions that use state.interaction freely. (Only allow BOOT,
LINK_PREVIEW, and EVENT_LOGGED)
Bug: T161769
Bug: T163198
Change-Id: I99fd5716dc17da32929b6e8ae4aa164f9d84c387
This fixes a bug in I8a63d82, where the pageTitleHover and
namespaceIdHover properties of EventLogging events weren't being set.
Bug: T164256
Change-Id: Ie2c2d253f6508b89d48129fd17a902e5ded7cad5
Action changes:
* Include the namespace ID in LINK_DWELL.
Reducer changes:
* Add the createEvent helper, which adds the pageTitleHover and
namespaceIdHover properties in the event.
* Create "dismissed", "dwelledButAbandoned", and "opened" events using
the createEvent helper.
Additional changes:
* Store the target page's associated mw.Title using jQuery.
* Update the eventLogging reducer's test cases so that all LINK_DWELL
representations include title and namespaceID attributes.
* Create the createStubTitle factory function, which returns a minimal
usable mw.Title stub, and use it in the actions and integration test
cases.
Bug: T164256
Change-Id: I8a63d82a65324680dff9176020a8ea97695428c4
I6a38a261 made the eventLogging change listener count and discard
duplicate events and count duplicate tokens.
While we isolate the issue(s) that lead to duplication (reuse, likely)
of tokens, make the eventLogging change listener discard events with
duplicate tokens as well.
Bug: T161769
Bug: T163198
Change-Id: I0dbb16c37814d39d7aec35c8fb7cc7309704c550
Binding the behavior has been left out as it requires some refactoring.
Ideally, that functionality should be tested via integration tests, which
is already done.
Bug: T133022
Change-Id: If2fd472847eb3557de97c7ec9619e8831e9bda6d
Per the Popups schema [0], the "dismissed" action expects the
perceivedWait property to be set.
Happily, the time until the preview was shown is already recorded and
used to determine whether or not to create a "dismissed" or
"dwelledButAbandoned" event.
Reducer changes:
* When creating the "dismissed" event, set the perceivedWait property to
the already accumulated timeToPreviewShow.
[0] https://meta.wikimedia.org/wiki/Schema:Popups
Bug: T164256
Change-Id: I3fd253f1c2ff5fa8e24c9225060d728ffd8dfd27
* Test for cases when a thumbnail won't be generated;
* Test thumbnail widths and heights;
* Test thumbnail coordinates.
Bug: T133022
Change-Id: I0f7d94816a2faab488b84f55b4462c8732d1031d
Since Ieea378c9 all QUnit tests are run in Node.js and not in the
browser. Tidy up references to QUnit inside of the codebase and tooling.
Changes:
* Don't exclude src/processLinks.js in Istanbul code coverage reports.
* Don't test for window.QUnit in createSchema. We no longer run the risk
of sending beacons while running the QUnit test suite as it's no
longer run in the browser.
Bug: T160406
Change-Id: Ifb6adb84b8019dd69231b50af00bf978b708fc60
The tests were removed in Iae0a78d0b8a13353de70794b67387f2c3bab44c6.
The ultimate goal is to refactor the renderer code and make it
testable, but before doing so we need to add tests to cover the existing
code. This will give us confidence that we won’t accidentally break
anything when we refactor.
Some of the tests have been removed as the functionality covered by
those were moved to model.js in I20f29657fcf94101a71ed13c0920508db71292ce.
Bug: T133022
Change-Id: I7b20324dd5fe8a428cdd96959b65bc82d44fb515
The ultimate goal is to refactor the renderer code and make it
testable, but before doing so we need to add tests to cover the existing
code. This will give us confidence that we won’t accidentally break
anything when we refactor.
Bug: T133022
Change-Id: I3ecbfb9bb3ac9c63fdd40df502796748c62949fe
The tests were removed in Iae0a78d0b8a13353de70794b67387f2c3bab44c6.
The ultimate goal is to refactor the renderer code and make it
testable, but before doing so we need to add tests to cover the existing
code. This will give us confidence that we won’t accidentally break
anything when we refactor.
Bug: T133022
Change-Id: I897276a1a953f6be62e4c2d4a24e0f22fc6ef141
With the final goal to remove the real mw stubs, move the processLinks
tests away from test cases, and split getTitle to be tested standalone.
Supporting changes:
* Move getTitle out to it's own module
* Will be tested separately in a followup commit
* Remove global declaration of mw.popups.processLinks (not needed any more)
Bug: T160406
Change-Id: Ieebd1257a2476081c67a318d3f05dffa1d3b9bdd
The eventLogging change listener is responsible for ensuring that the
internal state of Page Previews matches its external state (that
perceived by the user and UA). It does this by logging events with
ext.eventLog.Schema#log. This makes it the perfect place to track and
discard duplicate events enqueued by the Page Previews codebase observed
in T161769.
Make the change listener track events that it's logged by storing hashes
of the dynamic parts of them in memory. If the eventLogging change
listener sees the same event more than once, then it discards it and
increments a counter in StatsD.
This behaviour should be enabled for a matter of days as we should see
whether the duplicate events are being enqueued by the Page Previews
codebase immediately.
Bug: T163198
Change-Id: I6a38a2619d777a76dd45eb7300079e1f07b07b12
The statsv change listener depended on both the analytics tracking
function and whether it should log metrics to StatsD. We can simplify
the behaviour of the change listener by passing in a function which
doesn't log metrics to StatsD if such logging is disabled.
The change listener is now more isolated from other components.
Moreover, sharing the analytics tracking function with other components
is simpler as there's no repeated code.
Bug: T163198
Change-Id: Ibf4785fa4c27c1ad4739f02410f57412f56ff481
The setup and teardown hooks on QUnit.module are deprecated on the
latest versions.
See https://api.qunitjs.com/QUnit/module
mw-node-qunit supports them but we shouldn't be using them.
Bug: T160406
Change-Id: I32c07f22d01d16449a6e37f46ff20c577a1f14c6
Tests are basically unchanged, except for some stubs on beforeEach.
Supporting changes:
* Bring stubs from the mediawiki library for mw.Uri,
mw.Title.newFromText and mw.RegExp into stubs.js
* Remove hook onResourceLoaderTestModules given there are no resource
loader test modules after migrating processLinks.test.js
Why bring stubs from real source? This is not optimal. It could be the
case that the stubs would need to be updated at some point in the
future. That's why in the comment of each stub, it is specified where it
came from, and what was changed to make it work. It is not optimal but
it should help with a future update if necessary.
Also checked the history of the stubs and these three stubs are very
stable with a small commits per year, usually adding some extra
functionality (not breaking changes) (the rest of the commits are
docs/format stuff), so the core behavior that we rely on here shouldn't
change in a fundamental way. See the github links:
* https://github.com/wikimedia/mediawiki/commits/master/resources/src/mediawiki/mediawiki.Uri.js
* https://github.com/wikimedia/mediawiki/commits/master/resources/src/mediawiki/mediawiki.Title.js
* https://github.com/wikimedia/mediawiki/commits/master/resources/src/mediawiki/mediawiki.RegExp.js
Right now this stubs allow us to bring the test to run in isolation in
node.
The initial plan was to do change the test to be less test-case oriented
with dependencies on mediawiki.*.js and not to bring fake "real" stubs,
but after looking into it, given that:
1. the test cases in the test seem pretty informative showing the kind
of links that popups accepts
2. the stubs are acceptably easy to bring in, and are pretty stable
I decided to go with this approach initially to finish the migration
without changing the meaning of the tests.
If we want to remove the stubs and morph the test to verify stub calls
and move the test cases to documentation on the source, I'll tackle that
on a future commit.
Bug: T160406
Change-Id: Ieea378c9b7fec9116222b4a099c226d1f1131f65
If the user dwells on a link for long enough, then the gateway makes a
request, which is allowed to complete regardless of whether or not the
response is required.
If the user abandons the link but the request completes before the
abandon completes - currently, ABANDON_END_DELAY is 300 ms - then the
preview will be rendered temporarily.
Fix this by not rendering the preview if the user has started to abandon
the link.
Bug: T163350
Change-Id: I154dde4e3ccaed3d11cb023c85c44451fc0ad957
If the user CmdOrCtrl+Clicks on a link, then the link will remain
focussed. If a preview is shown, clicks another element on the page, and
then there'll be no token to include in ABANDON_* with.
Bug: T162924
Change-Id: Ie2237aa55ea9a11070498b66c73b8bf1898d8d44
I09d8776 introduced a bug in the eventLogging reducer where reducing
LINK_CLICK would remove any accumulated interaction state but not close
the interaction.
Update the associated tests so that they correctly test the reduction of
this action.
Bug: T162924
Change-Id: Ia03e719c228ee96f279c1fa89252dc6b6371a8e9
... the interaction.
Following on from Iccba3c4c, LINK_CLICK shouldn't be considered the end
of an interaction because the link or preview can still be abandoned by
the user. Unlike ABANDON_END and LINK_DWELL, therefore, LINK_CLICK
musn't destroy the interaction but indicate that no more events should
be enqueued for the interaction.
More concretely, if the user has clicked the link or the preview, then
a "dwelledButAbandoned" or "dismissed" event shouldn't be logged.
Changes:
* Distinguish between "finalizing" and "closing" an interaction, where
the latter is the current behavior of ABANDON_END, LINK_DWELL, and
LINK_CLICK, in the eventLogging reducer and associated tests.
* If the interaction is finalized, then either the "dwelledButAbandoned"
or "dismissed" events shouldn't be logged.
Bug: T162924
Change-Id: I09d8776da992053f89a77508e29a7cde3cfeeac6
MediaWikiGadgetsDefinition does some basic gadget name sanitization
and we have to do the same when checking "is gadget enabled for user"
Changes:
- sanitize gadget name same way as
MediaWikiGadgetsDefinitionRepo::newFromDefinition() does.
- add try{} catch() when loading gadget as getGadget might throw an
exception
Bug: T160081
Change-Id: Ia7a57e9dcfa3b25129d6d2bf75795372fad2b251
... in the eventLogging reducer.
Like ABANDON_END, the LINK_CLICK should be considered the end of the
interaction in the context of the EventLogging instrumentation, i.e. no
events should be logged after the user clicks the preview or the link.
See T159490#3172692 for additional context.
Bug: T159490
Change-Id: Iccba3c4c2b6121016ff7923c11b1622bc046ad6b
Like the FETCH_COMPLETE and ABANDON_END actions, the PREVIEW_SHOW action
was delayed but not conditionally reduced. As ABANDON_END is delayed,
there's a potential for a race and if ABANDON_END is reduced before
PREVIEW_SHOW, then there's no interaction to reduce the action into,
which causes an error, e.g. T159490#3165276 and T162373. Making
PREVIEW_SHOW require a token stops the error occurring in this scenario.
An alternative would be to clear the timeout created in
ext.popups.Preview#show in #hide. However, this would be inconsistent
with actions#fetch and actions#abandon.
Bug: T159490
Change-Id: Ibd2c0c6f45e4392582cc6ed08517f6ca1146d57a
Mirroring all other actions that are dispatched after some delay, add
the token to the PREVIEW_SHOW action.
Supporting changes:
* Pass the token to ext.popups.Preview#show so that it can be passed to
actions#previewShow.
Bug: T159490
Change-Id: I128fd56e770ed09d5d0dc55db73d11b013049c79
The ABANDON_START action test cases were (accidentally?) nested under
the FETCH_COMPLETE action test cases...
Bug: T162373
Change-Id: Ia7866178162512602dedd7f70d62c17995ee5076
... instead of using the active element.
In the case of the eventLogging reducer, this fixes scenario 4.1 from
T159490#3150331, which was caused by late FETCH_COMPLETE actions being
reduced regardless of whether interaction had been finalised or a new
interaction had started.
Bug: T159490
Change-Id: If9d718625b0302ea2f75a778005643b4eef62bde
Treat these responses not as an API failure. Show a generic
preview whenever the server responds with a 404.
Bug: T160744
Change-Id: Id6169d9d4c7493f5b6511cc78fe65d448cdadc03
Require that two promises are resolved (or one is rejected) before the
FETCH_COMPLETE action is dispatched. The first promise represents the
gateway request and the second represents an arbitrarily long delay. If
the first resolves before the second, then there'll be a delay until the
second resolves; whereas if the first rejects, then there's no delay.
Change-Id: I496fe317337745c593594efff26688c46d661bf3
The delay/timeout logic in actions#fetch will require more than one wait
call, for example.
Changes:
* Update the stub created in setupWait to store all deferred-promise
pairs and update the integration tests so that they don't use the stub
accidentally.
* Remove all references to the waitDeferred/waitPromise properties.
* Fix tests that relied on the waitDeferred/waitPromise properties being
available regardless of whether wait had been called.
* Update outdated or brittle - in the sense that it didn't reference
constants but their values - inline documentation.
Change-Id: I94345cdf4126b6c540d4fb8135a7a7e4d0507bed
Mixing in the delay was introduced in If3f1a06f so that the total RTT
for an API request could be calculated. Now that the FETCH_END action is
dispatched when the gateway request ends and not when the preview model
is resolved, this additional information (state) is redundant.
Change-Id: I7e6ffe0945ffedd9425525fa7da855e729d50b77
Ideally, the preview model is resolved after 500 ms, regardless of
whether the internal gateway takes 100 or 300 ms. Given this, there's an
important distinction to be made between the "fetch" ending and it
completing and their associated actions.
Changes:
* Dispatch the FETCH_COMPLETE action when the preview model is resolved.
* Update the reducers accordingly.
Change-Id: I62c9cb0430284b76338ea80bd170cac5af4be9d0
If the user has disabled PP via the settings dialog or they aren't in
the experimental condition, then link titles shouldn't be emptied.
Because this behavior has to respond to the user enabling/disabling PP
within the same page session, change the linkTitle change listener
rather than conditionally registering it.
Bug: T161277
Change-Id: I53c1a1d3e4436e2ffe08da27da388f394f4e8817
If the user abandons link A (or preview A) and immediately dwells on
link B, then log a "dismissed" or "dwelledButAbandoned" event.
In this context, "immediately" means before the ABANDON_END action is
dispatched, which, currently, is 300 ms after the ABANDON_START action
is dispatched.
Bug: T159490
Change-Id: I49f0f5dfb3e6c08844f1794fee8cb6170e93981b
Reducer changes:
* Add tests for ABANDON_END case.
* Extract the body of the ABANDON_END case into the createAbandonEvent
helper function.
Additional changes:
* totalInteractionProperty -> totalInteractionTime elsewhere in the same
file.
Bug: T159490
Change-Id: Ifff34271395f330b83cfe487e84800fe2d6f3811
Reducer changes:
* Make the eventLogging reducer queue a "tapped settings cog" event when
reducing the SETTINGS_SHOW action.
This was discovered while testing I6ce7d72b.
Bug: T159490
Change-Id: I6ce7d72b364d20c71b0e2cfed98e99f7997895e5
Like dwelling and abandoning, clicking on a preview is the same as
clicking on a link.
This fixes scenario 3 from T159490.
Bug: T159490
Change-Id: I6d9ff52b62bec93ebfcc9b6d267a46cf961852fb
For now, mirror the interaction modelling in the preview reducer in the
eventLogging reducer to handle the user either:
* Repeatedly dwelling on and abandoning a link.
* (Repeatedly) moving their mouse between the link and the preview.
This fixes scenarios 1, 2, 5, and the general issue from T159490.
Bug: T159490
Change-Id: Ia771f325e541c107348b16b47c5b786c97847652
For logging to work:
1. $wgWMEStatsdBaseUri needs to point to a valid statsv endpoint,
e.g. 'https://en.wikipedia.org/beacon/statsv'.
2. $wgPopupsStatsvSamplingRate needs to be set. Note that the codebase
already contains the EventLogging functionality, which is configured
separately. Separately configuring different logging mechanisms
allows us to avoid sampling mistakes that may arise while choosing
one or the other. For example, let's say we want to use EventLogging for
10% of users and statsv for 5%. We'd sample all users into two
buckets: 50/50. And then we'd have to set the sampling rates as
20% and 10% respectively, only because of the bucketing above. To avoid
this kind of complications, separate sampling rates are used for each
logging mechanism. This, of course, may result in situations where a
session is logged via both EventLogging and statsv.
3. The WikimediaEvents extension needs to be installed. The extension
adds the `ext.wikimediaEvents` module to the output page. The
logging functionality is delegated to this module.
Notable changes:
* The FETCH_START and FETCH_END actions are converted to a timed action.
* The experiments stub used in tests has been extracted to the stubs
file.
Logged data is visualized at
https://grafana.wikimedia.org/dashboard/db/reading-web-page-previews
Bug: T157111
Change-Id: If3f1a06f1f623e8e625b6c30a48b7f5aa9de24db
... thumbnails.
A good example of the difference in behaviour of the PageImages API is
here <T156800#3087507>. The API defers to File#transform, which scales
the largest dimension of an image, not always the width, e.g. if an
image is 1000px x 2000px and the request is for a thumbnail "of 1800px",
then the thumbnail will be 900px x 1800px.
Bug: T156800
Change-Id: I64bc2244ee78a594298d8637233b0da1083700eb
Keep all configuration-like values in one file.
Changes:
- moved EXTRACT_LENGTH to constants.js file
Change-Id: Ibe5ecfc60f2c09a30a9ecb3586bc5fb6a7365476
If the image isn't an SVG then it shouldn't be scaled past its original
dimensions. Handle the case where the requested thumbnail can't be
generated on the server as the original is too small ( < 320px,
currently [0]) in the same way.
Moreover, if the image happens to have the exact dimensions that we're
requesting (300px or 600px wide, currently [1]), then use the original
image to avoid unnecessary work/pressure on caches.
Supporting changes:
* Update the SVG_RESTBASE_RESPONSE fixture to use the extension returned
by RESTBase (and the PageImages extension) for the thumbnail:
.svg.png.
[0] https://github.com/wikimedia/restbase/blob/master/v1/summary.yaml#L121
[1] https://github.com/wikimedia/mediawiki-extensions-Popups/blob/master/src/constants.js#L2
Bug: T156800
Change-Id: I5d0aa161e80869e4b4f5425d906d1e753047a3a3
The same file is on preview/model.js which is the one actually used by
the application. For some reason the file is a duplicate of model.js,
and it is the file that is required on its QUnit tests.
This patch removes it and points the unit tests to the correct file,
preview/model.js
It was also required by popups.js which was removed in the previous
commit.
Change-Id: Id175a764d9b67fb8d0e8fbf4a4623a3420f76094
We used to query the MediaWiki API to only return non-free images.
This patch allows us to query the API for images with any license.
The RESTBase end point is already returning images with any license.
Bug: T158632
Change-Id: I9ac60b6f74a7f7eb2cb160ee522c2c3a26dd0858
Rather than manipulating the URL of the original image to get the URL of
the appropriately sized thumbnail, manipulate the URL of the thumbnail
image.
While we could manipulate either the thumbnail or original image URL,
there are subtle differences between the two, so manipulating the latter
makes the generateThumbnailData function as simple as possible, e.g. we
don't have to splice in "/thumb" after "/commons".
Also, ensure that the thumbnail's dimensions are scaled appropriately.
Bug: T156800
Change-Id: I6825bad14b1131dc81f23dcf120cf8ffb7d7b4f6
Supporting changes:
* Use mw.Map stub
* Use assert.expect instead of number of assertions in QUnit.test
(deprecated in newer QUnit)
* Don't specify assert.expect( 1 ), because it is the default (no
assertions will make the test fail).
Change-Id: I64a3e76917e75b8c6d496f20e5b5dcafb338a46a
* Logged in users bypass bucketing. They keep working as before.
* When Page Previews is configured as a beta feature, logged out users
won't see the feature.
* If an anonymous user has enabled/disabled the feature using
the settings cog then they will see or not see the feature
depending on the value of their setting.
* The other anonymous users are bucketed. By default 90% of these
users see the feature, the other 10% don't. These numbers can be
controlled by the config variable `wgPopupsAnonsEnabledSamplingRate`.
Bug: T157700
Change-Id: I5307587b10f4849c4e82d3b064ff759121c2de67
mw.storage#get doesn't take a default value to return if the underlying
storage is disabled or the key is missing. In the former case it'll
return false and in the latter it'll return null, i.e. in neither case
will it return undefined.
Bug: T157700
Change-Id: I3f653c11468e17b64765e85ebb3b8f03e311352a
Because of the globals mw.popups.wait usage and mocking in both actions
and integration, they need to be migrated in a single step, fixing them
both to require wait.js and mock using mock-require instead of the
global variable.
Additional changes:
* Fix FIXMEs about actions.js using the global mw.popups.wait instead of
the require one.
* Fix the unit tests to use require mocking for wait.js instead of
global variable mocking in both integration and actions tests
* Change tests that use deferreds and promises to be async qunit tests
(Deferreds are asynchronous with jQuery in node, apparently they
weren't in the browser)
* Change integration.test.js to use require on Redux and ReduxThunk
Change-Id: I8e3e87b158bd11c9620e77d0a73e611cf9e82183
The "checkin" part of the Popups schema was superseded by the
ReadingDepth schema, the implementation of which is tracked by T155639.
As well as removing all checkin-related code, update the Popups schema
to the latest version - the version that doesn't have the checkin
property.
Bug: T155639
Depends-On: I762ec3fc91decf3cffa869dbd783faf62f01329a
Change-Id: If764917b6e121e1f9db980a4efa30c0f7a166197
The render change listener is hard coupled to the renderer file, so in
order to migrate the test, instead of stubbing a global variable, we had
to either inject the renderer into the change listener factory as done
everywhere else, or mock the require call.
In order to do one thing per commit, we're mocking the require call
here to get the migration done, but added a FIXME to use dependency
injection instead in a future change.
Change-Id: I50f82cdc9664d34b8a8ccc1ff368f7209404159d
Given this is a jsdom environment assertions using jQuery's :visible
have been changed to check the display property for visibility.
See https://github.com/tmpvar/jsdom/issues/1048
Change-Id: Ifad8067c0b50053a94ac977ee1f1f5a3066bfa16
And on the way there actually put it in a changeListeners/ folder and
rename it to eventLogging.test.js
Change-Id: I60685021841b44f606f39b07bf7f5262344262f4
In order to run qunit tests on sources that use common.js modules, set
up infra to run qunit tests in the node cli when running:
npm run test:node
Changes:
* Add npm script test:node that runs the tests
* Run node tests on CI (npm test)
* Add a qunit node test runner: mw-node-qunit
* Migrate a test from the root hierarchy and another one from the nested
one to prove it works (globs fail otherwise)
* reducers/settings.test.js to node qunit to prove it works
* counts.test.js to node qunit to prove it works
Change-Id: I55d76b7db168f3745e0ac69852c152322ab385c3
This change resizes thumbnails to the appropriate width
based on the value of mw.popups.gateway.THUMBNAIL_SIZE
Tests cover
* When requested thumbnail is < than original size
* When requested thumbnail is > than original size
* When requested thumbnail is an svg and originalimage
smaller than requested thumb size
Bug: T156800
Change-Id: Ib375b97e2bc959e91de5177efc3df1f2ded54a5b
Sometimes we make choices on the users behalf if we don't have
sufficient information, so this name is more applicable.
If beta features is disabled and the user is anon they have not
explicitly opted in.
Change-Id: I5d816f569fc54f8bf74d6e5a06246b7fa7036e06
If the Popups code is enabled in beta feature mode only then
only add it to the page if the beta feature is enabled.
isEnabledByUser now returns false if the user is anonymous
and Popups is restricted to beta mode.
Bug: T156290
Change-Id: If152d2a67a079050173c6d642e0960b59730bc6d
* Create new namespace mw.popups.gateway to contain them
* Create folder resources/ext.popups/gateway/
* Split gateway{,.test}.js into gateway/{mediawiki,rest}{.test,}.js
* Extract stateless functions out of factory scope now that there are no
name collisions between the factories.
Bug: T123445
Change-Id: Ib256871c3e4cfe3f13361cb66d4e9a67e9823c7b
Instead of using constructor functions, use factory functions to
generate the gateway objects, because of:
* Consistency on the approach in the repository, no constructor
functions are used, the factory function pattern prevales.
* Real private data with closures
* No use of `this`, which is dynamic and unbound in JS and a source of
errors. In contrast, by using a factory function, the
function/methods-of-the-object are tightly bound to the internal data.
Additional changes:
* Specify more the type of createRESTBaseGateway's parameter to improve
clarity on the intent of the parameter
* From: @param {jQuery} api
* To: @param {Function} ajax function from jQuery for example
Supporting documentation:
* https://medium.com/javascript-scene/javascript-factory-functions-vs-constructor-functions-vs-classes-2f22ceddf33e
Change-Id: Iacbb098b646843a01f459b15343057e2e4851d35
* Introduce the new config variable `PopupsAPIUseRESTBase`.
* Create a gateway interface that exposes the getPageSummary
method. Both MediaWiki API gateway and RESTBase gateway
implement this interface.
Bug: T123445
Change-Id: I71a8a848f3143fa4a0dfd4ca182ee086903110bc
The existence of these files was causing me much confusion as I thought
they were still being used in the mpga branch.
If this is note the case they should be removed to make the repo easier
to understand.
Changes:
* Move the ext.popups.animation.less into ext.popups since that's where
its shipped.
* Remove files which are not being used.
* Since csslint now runs on ext.popups.animation.less fix the linting
issues (previously linting was not occuring)
Change-Id: Iae0a78d0b8a13353de70794b67387f2c3bab44c6
In I6bbff6a5 a preview was assigned the "extract" type if it had an
extract whereas it should've been assigned the "page" type.
Bug: T152004
Change-Id: I5e0cdc8f528647f1e20873753be29c6b85b3a384
Note this is a breaking change in the sense that it breaks the function
signature for createModel which no longer needs to be passed the last
modified time. Given this change is in the mpga branch at current time
and hasnt been exposed for public use I have deemed this okay.
The change removes tests relating to the last modified bar, removes
lines in the templates and updates existing tests and code to reflect
the new function signature for create model.
Settings icon is floated to right which will be flipped for RTL
users.
Bug: T137775
Change-Id: I7737e37d956c62f1f1c0694d7a25a58d91651f4d
The preview model is included in the FETCH_END action so, when reducing
that action, pluck the preview type from the model and add it to the
interaction state.
Change-Id: I0e51a85b4dada8f9a17ec4882e0c13b2c4382d24
Per T152004, if the preview has an empty extract, then it's a "generic"
preview; it's an "extract" preview otherwise.
Bug: T152004
Change-Id: I6bbff6a507fe3e4f03b3a44f64d875e7dbb512df
Following on from I20f29657, simplify the gateway by making it return
instances ofext.popups.PreviewModel.
Supporting changes:
* data -> model in mw.popups.renderer#render and the
create{,Empty}Preview helper functions.
Bug: T152004
Change-Id: I531f609056ae813ae5e6b1d53010d77be0799ec1
As noted in the review of I71a8a84, creating a new object to represent
the thumbnail is confusing. It's also somewhat wasteful.
Since both the MediaWiki API and RESTBase both agree on using "source"
as the key for the thumbnail's URL, it's reasonable to use it in the
gateway and renderer.
Bug: T152004
Change-Id: Ie1346d13b6e88004b817f5968448188ed83cfda4
After the gateway has made a request, it processes the page. This
processing both mangles and processes the MediaWiki API response. The
latter step contains two pieces of related business logic:
1. Deciding if a page has been modified recently.
2. Processing the extract/deciding if it's viable (see processExtract).
In order to support switching between the MediaWiki API and RESTBase and
to simplify the event logging reducer when it comes to instrumenting the
"type" of preview, extract this business logic into a model.
Bug: T152004
Change-Id: I20f29657fcf94101a71ed13c0920508db71292ce
I4c15296e introduced the wgPopupsConflictsWithNavPopupGadget client-side
config variable, an always-available flag that represents whether the
user has enabled the NavPopups gadget. This property has been missing
from the Popups instrumentation since I8a3f5835.
Changes:
* Add the value of wgPopupsConflictsWithNavPopupGadget to BOOT action
and reduce it into the base instrumentation data.
Bug: T151058
Change-Id: I09924d0cb10e96b0a846940506ebfeaa7086cd17
Since the footer link part of the settings system - clicking the link
opens the settings modal - it seems appropriate that it should be
controlled by state managed by the settings reducer.
Bug: T146889
Change-Id: I3a8549dbf1952cd0556f663496c55de91acaf2c0
As before, we need to send the Page Previews blob that logged out users
can enable/disable previews.
Changes:
* Unconditionally add the ext.popups RL module to the output.
* Send the value of Popups\PopupsContext#isEnabledByUser to the client
as $wgPopupsIsEnabledByUser.
* Make mw.popups#isEnabled return $wgPopupsIsEnabledByUser if the user
is logged in.
Bug: T146889
Change-Id: Id2ccfaa81a327e222fff400f4a5f22f96c99ea31
The behavior of the cog varies when:
* The user is logged in/out.
* Page Previews is enabled as a beta feature.
Since the behavior of the cog doesn't vary per-preview, it can be
determined at boot time and passed to the renderer. However, in order to
keep the renderer stateless, we pass the behavior to it when a preview
is added to the DOM.
Changes:
* Add the mw.popups.createPreviewBehavior factory function, which
returns an object that encapsulates how a preview responds to the user
dwelling on it, abandoning it, and clicking on the cog.
* Invoke mw.popups.createPreviewBehavior at boot time and pass it to the
mw.popups.changeListeners.render change listener, which then passes it
to mw.popups.Preview#show.
* Make mw.popups.Preview#show responsible for binding event handlers
and configuring the cog based on the behavior.
Bug: T146889
Change-Id: I39d7d0afd7b1fe896019a1b3a82ee907bfb20edd
The SchemaPopupsSamplingRate config variable is inconsistently named and
even forwarded to the client as wgPopupsSchemaPopupsSamplingRate.
Changes:
* SchemaPopupsSamplingRate -> PopupsSchemaSamplingRate
* Forward PopupsSchemaSamplingRate to the client as
$wgPopupsSchemaSamplingRate.
Bug: T146889
Bug: T146434
Change-Id: I80d6a0abccf462e2eb0fd96af6849b5e82b49c26
Per Id1339dc2, the LINK_ABANDON_* and PREVIEW_ABANDON_* actions can (and
should) be merged, as they are always reduced in the same way.
Change-Id: I71b30d4d2774deb4efea9e565f2ccd7383bf08c1
Initially, the link element was added to LINK_ABANDON_* to determine
when an interaction should be reset. However, this wasn't enough to
protect against timing-related bugs, e.g. T154923. Now that tokens are
used to determine when interactions should be reset, the link element
can safely be removed from the LINK_ABANON_* actions.
Supporting changes:
* Update the return type of mw.popups.actions.linkAbandon in its
DocBlock.
Change-Id: Iaaed7a4846af75f9c4051d7bc761022ea5b3f6cf
I2ecf575b introduced a regression where the linkTitleChangeListener
wouldn't destroy the active link's title if it had been dwelled on
within 300 ms of the previous active link being abandoned.
The LINK_ABANDON_END action is dispatched after 300 ms and ignored if
the active link has changed.
Change-Id: If0c16afd6ca1c44f0e7eed497f62f0190005a619
... instead of using the active elements, and check also the user
dwelling flag to only act if the user is not dwelling.
Preview and link abandon end and start share now test cases, meaning it
can be collapsed into one action creator in a followup commit.
Change-Id: Id1339dc23e8f1b1b7514c769cd09f8bdadc759b4
When dwelling back to a link, previously a new LINK_DWELL action would
be sent and the reducer would reset the interaction as if it was a new
one.
This shouldn't happen, since dwelling back to the active link (from
the preview, for example) is not a new interaction, and should not
create a new interaction or hide and show the preview.
With this change, the preview reducer has the notion of new or repeat
interactions, and only resets state on new ones, and only sets
isUserDwelling when dwelling on a the link of the current interaction.
Also when the interaction is repeat, we guard on the action creator and
don't trigger the wait or the fetch request.
Change-Id: I71cde81cbfe50b6f955e562e7e5b57d0f920fdb9
... after dwelling on preview.
Currently, the asynchronous PREVIEW_ABANDON_END action wouldn't be
ignored if the user were to have dwelled on the link that the preview is
about. The action is only ignored if the user is dwelling on the
preview, i.e. if preview.isUserDwelling is truthy.
Set preview.isUserDwelling to truthy if the user is dwelling either on
the link or on the preview.
Bug: T154923
Change-Id: I8e6989790da5884992ee9dda9e4895e2e58e6694
Unfortunately, QUnit doesn't complain that QUnit#module was being called
with a function and not an object.
Change-Id: I6d3b4941566fa4e4ecad2dbe6f7b5fdd62cdaf72
Changes:
- introduced PopupsGadgetsIntegration class handles all checks
- show help message on Preferences page when NavPopups is enabled
- unit test everything
Bug: T151058
Change-Id: Ia474b1b30378efe84dedf3ad47c1f833e88d69b5
Consider the user repeatedly dwelling on and abandoning a link. Testing
whether or not the element has changed before modifying state is not
sufficient. Fortunately, we already generate a unique token and include
it in the LINK_DWELL action for use in the eventLogging reducer - the
so-called "interaction token".
Change-Id: Iae9544055f7eccb4f7f071797eb360bc3456861a
Changes:
- removed onExtensionRegistration and related TODOs
- introduced onUserGetDefaultOptions hook
Change-Id: I6648802c19d90df36a211d6494033aca30c078ca
* To speed the time the tests spend running
* To make them reliably execute and not depend on magic numbers
Additional changes:
* Use mw.now instead of Date on checkin.js
* Split big setVisibleTimeout test into several smaller ones
* Extract helpers for the tests
Change-Id: I9d3233fccf6de0997f968d096e375df996e87786
While I1bf953b2 fixed the typo in FETCH_START_DELAY, it didn't do
anything to address the inconsistency in the time it takes to render a
preview (preview delay).
If the gateway result resolves before the target preview delay (500 ms)
is met, then delay dispatching the FETCH_END action until it is. If,
however, the gateway result resolves after the target preview delay,
then dispatch the FETCH_END action immediately.
Supporting changes:
* Expose the fetch action to ease testing it and to help tidy up the
mw.popups.actions.linkDwell tests.
Change-Id: Idf02343a0417cfbb33093a81d97ecebc5c1c7379
Test what mw.popups.haveCheckinActionsBeenSetup guards against so that
it doesn't have to be touched during the associated test.
Bug: T147314
Change-Id: I315e0782a0e3b1d04878f1ee6f538e54531f7df5
Use schema revision 16163887.
Add the 'checkin' action, which is accompanied by the 'checkin'
property. The action is logged at the following seconds
(Fibonacci numbers) after the page loads: 1, 2, 3, 5, 8, 13, 21,
34, 55, 89, 144, 233, 377, 610, 987, 1597, 2584, 4181, 6765.
The `checkin` property contains the values listed above.
Bug: T147314
Change-Id: Ib9ec7bd0e60aa34a04e32222b025347f6ee31794
Per #Design's mocks, the "Reading preferences" section should appear
immediately beneath the Skin section, if it's present.
Changes:
- onGetPreferences will try to inject PagePreviews option after
skin section or on the end of section when skin select is not available
- unit tests for new behaviour
Bug: T146889
Change-Id: Id14c0774aee917b7e949ab3ba5968a038b3ca933
Changes:
- added missing unit tests
- introduced PopupsContextTestWrapper, removed all unit-tests logic
from Hooks/Context classes
- removed getModule() from Popups.hooks.php, it's hooks responsibility
to keep single context instance
- removed class_exists calls, use ExtensionRegistry instead
- pass ExtensionRegistry as a dependency so it's easy to mock
- reach 100% code coverage
- introduce PoupsContext::isBetaOn() as it was used in many places
Change-Id: Id014efe72edf24628539c76968e53eb3e06709f4
Action changes:
* Update the BOOT action to include isEnabled and update the associated
tests.
Reducer changes:
* Make the preview and eventLogging reducers handle the BOOT action's
new shape.
Bug: T152687
Change-Id: I3fa2098269a32912eda99ceb8f13887688a14c15
The Page Previews A/B test was disabled in T144490. After the changes
for that task had been deployed we discovered that the old definition
for when previews are enabled was invalid.
Previews are enabled:
* If the beta feature is enabled and the user has enabled the beta
feature.
* The user hasn't disabled previews via the settings modal.
* And soon, if the beta feature is disabled and the user hasn't disabled
previews via their user preferences (see T146889).
Changes:
* mw.popups#createExperiment -> mw.popups#isEnabled
* Make mw.popups#isEnabled act like a factory to maintain backwards
compatibility.
* Update the associated tests.
Bug: T152687
Change-Id: I713a90dfb29866b27738e7d19e8a22518a12d417
When opening the settings dialog, the form didn't represent the current
enabled status of the application. This change makes sure that every
time the form is shown, the form state represents the current
application state.
If the application is enabled, then the 'simple' (Enabled) option is
selected on the form. If it is disabled then if there's Navigation
popups, it selects 'advanced' (Advanced popups), if there are not, it
selects 'off' (Disabled).
Changes:
* Change listeners
* Set the form state every time the form is going to be shown
* Update tests
* Supporting changes
* Add settingsDialog#setEnabled which updates the DOM form based on the
enabled flag and the navpops state
* Extract isNavPopupsEnabled function in settingsDialog.js to be used in the
form creation and also when setting the form enabled state
* Add test verifying changeListeners#settings shows and sets enabled state
properly when showing the dialog more than one time
Change-Id: Ic660f48d9a78e47c09a192ab86681196d2e01d61
Changes:
* Generalized previewCount change listener to syncUserSettings
* Added state.preview.enabled to be saved on change
Change-Id: I403a490fee9c8e125175996ba30c63c232b5598b
This changes introduces new option on Special:Preferences page that
allows users to enable/disable Page Previews feature.
By default feature is disabled. Temporarily option in Preferences
page is smoke and mirrors as switching logic is still WIP.
Bug: T146889
Change-Id: Ifdd17ce265d2d4c7583433ed4991443c563f1fe3
The save action has been implemented, and it is listened to by the
canonical enabled state in the previews reducer, and by the settings
reducer to perform UI changes.
The enabled state of the application has been kept in the preview
reducer as the canonical source of truth. See supporting changes for
documentation about the decision.
Actions:
* Introduce new action SETTINGS_CHANGE with the enabled status
* Trigger that action when clicking Save in the settings dialog
Reducers:
* Listen to SETTINGS_CHANGE in the preview reducer to update enabled
status
* On the settings reducer
* Handle the SETTINGS_CHANGE action
* Add showHelp flag to determine if the help should be showing
Change listeners:
* Switch to compare past vs present changes in the implementation
* Handle showing and hiding the help
Supporting changes:
* On the rendered settings dialog:
* Change #hide to actually just hide and remove legacy if statement
* Add #toggleHelp method to show or hide the help dialog
* Add doc/adr/0003-keep-enabled-state-only-in-preview-reducer.md to
support the decision of making the saveSettings action creator return
a Redux.Thunk and keeping the enabled state just in the preview
reducer.
* Add NB to actions#saveSettings explaining and linking to the
document
Follow commits soon:
* Persist the settings change to local storage when it changes, and
unify with the preview change listener
Change-Id: I80dc5f29fbe6286f2e3e3b50d909894bc5041ccd
Reducer changes:
* Break the mw.popups.reducers.rootReducer test into @@INIT tests
for each reducer as there's shouldn't be any need to test that a
framework works as documented.
Changes:
* Move the mw.popups.reducers#preview, #eventLogging, and #settings
reducers into their own files in resources/ext.popups/reducers/.
* Make the associated tests mirror the new layout.
* Move the initialization of the mw.popups.reducers namespace into its
own file.
* Remove the mw.popups.rootReducer property in favour of a simple
private factory function per the reducer change above.
Change-Id: I94229d9ef1985f3806eec44c2e8234a5bbddc94f
High-level acceptance tests shouldn't expose low-level concepts such as
the Resource Loader ("RL"). Merge the "And the RL module has loaded"
step into the "Given I am on the test page" step as they are only ever
used together.
Changes:
* Use the ArticlePage#wait_until_rl_module_ready helper method provided
by mediawiki_selenium@1.7.3.
* Update and include the Gemfile.lock.
Change-Id: Ic81eae5dc0caa8ec51c91b21aef544f52d8c49e1
Merge the "Given the test page has been created" and "Given I am on the
... page" as they are only ever used together to mean "Given I am on the
test page".
Supporting changes:
* samples/links.wikitext -> fixtures/test_page.wikitext as it's more
obvious what the file is.
Change-Id: Ide261a8e8c55133ab2a13a80c48e266252f61f5f
Changes:
* Rename the somewhat awkwardly named "popups_core" Cucumber feature and
its accompanying step definitions to "previews".
* Remove as many references to Popups and HoverCards as possible.
* Use language consistent with the QUnit unit and integration tests and
documentation, e.g. the user dwells on a link.
* Remove settings-related step definitions from the previews step
definitions.
Change-Id: Ibc0d07e8394aeb640f9efaabfe10be5317bf4b8c
* Port ext.popups.desktop/ext.popups.settings.js to settingsDialog.js
* Blank tests for now. Needs Qunit integration tests.
* Transform into a factory function for future testing.
* Saving functionality is commented out, will be removed when immplemented in
the actions
* Add new incomplete action saveSettings
* Will perform the saveSettings async tasks and then trigger enabling or
disabling popups.
* Rename action settingsDialogClosed to hideSettings for consistency
Change-Id: I3936d3a4bc476de16d76025139be09f1798796c4
Since the user can dwell on a link, abandon it while moving to dwell on
the preview and vice versa:
* The time at which the link/preview was abandoned is recorded and can
be updated.
* The event can only be queued when the link/preview has definitely been
abandoned.
Reducer changes:
* Make the eventLogging reducer:
* Calculate the time it took to show the preview.
* Mark the interaction as finished when the user abandons the link or
the preview.
* Queue a "dismissed" event if the preview hasn't been shown or a
"dwelledButAbandoned" event if it has when the abandon has
finalized.
Bug: T152225
Change-Id: I6a254136f615484fc26e440fe5125289e74688a6
Reducer changes:
* Make the eventLogging reducer queue an "opened" event with the
requisite supporting data.
Changes:
* Update the Popups EventLogging schema to the latest revision, which
defines the "opened" action.
Bug: T152225
Change-Id: I9d67daf83815f1c08b6479497c566dfa337de4bc
Action changes:
* Mix in timing information into actions that are likely to need it:
LINK_DWELL, {LINK,PREVIEW}_ABANDON_START, LINK_CLICK, and
PREVIEW_CLICK.
* Generate and include an unique token in the LINK_CLICK action.
Reducer changes:
* Make the eventLogging reducer:
* Handle the LINK_CLICK action's new shape.
* Start (create) and maintain a model of the user interacting with a
link.
Bug: T152225
Change-Id: I671b12432ba2f7a93bf81043adb57ac30a4c38c3
Move all of the change listeners into
resources/ext.popups/changeListeners and remove the ChangeListener
suffix. Also, do the above for the associated tests.
Bug: T152225
Change-Id: I90ada465ea291d601f8f1c5c6e775148a2100319
Action changes:
* Add the PREVIEW_SHOW action.
Reducer changes:
* Increment the user's preview count in the eventLogging reducer.
Changes:
* Dispatch the PREVIEW_SHOW action when the preview has been shown.
* Add the previewCount change listener, which is responsible for
persisting the user's preview count to storage when it changes.
* Call the change listener factories individually in
#registerChangeListeners as their signatures aren't consistent.
Bug: T152225
Change-Id: Ifb493c5bff66712a25614ebb905251e43375420a
Action changes:
* Include the user's preview count in the user property of the action.
Reducer changes:
* Make the eventLogging reducer add the bucketed user's preview count to
the state tree.
Changes:
* Extract mw.popups.UserSettings#getPreviewCount and #setPreviewCount
and associated tests from the ext.popups.core module.
Bug: T152225
Change-Id: I6b7afef31311be8fede685deb536f577845cb9cf
Action changes:
* Group user-related data in the user property, rather than repeating
"isUser" and "user" prefixes.
* Include user's edit count in the user property of the action.
Reducer changes:
* Make the preview reducer handle the action's new shape.
* Make the eventLogging reducer add the bucketed user's edit count to
the state tree.
Bug: T152225
Change-Id: I8fae9e2d0f6889ffdd30bb5192513d194f791967
Changes:
* Extract the various shapes of stub user object into the shared
mw.popups.tests.stubs#createStubUser factory method.
* Register the ext.popups.tests.stubs module in
PopupsHooks#onResourceLoaderTestModules and make the test suite depend
on it.
Bug: T152225
Change-Id: I893b11461d8484bcaabfd950c2fa0dc672454a9d
Action changes:
* Include whether the user is logged in/out and information about the
current page in the BOOT action.
* Add the EVENT_LOGGED action, which represents the eventLoggined change
listener logging the queued event.
Reducer changes:
* Move all tokens and timing information from the preview reducer to the
eventLogging reducer.
* Make the eventLogging reducer reset the state.
Changes:
* Add the mw.popups.createSchema function, which constructs an instance
of the mw.eventLog.Schema object that can be used to log Popups
events.
* Add the eventLogging change listener, which logs the queued event.
* Add hand-crafted, artisanal documentation.
Bug: T152225
Change-Id: I8a3f58358b211cc55417dcda7e796fe538e3d910
This reverts commit 047bccfac1.
It created merge conflicts with a bunch of patches that were days old and about to be merged. Instead of rebasing 10 patches, let's revert this one, finish merging the other ones and just rebase and fix this one. I'll re-introduce the patch in a bit.
Change-Id: Ib495755b0ab4bfa5fdf5590b1271792862a47d4b
Changes:
* Make the gateway handle missing pages, which are characterised by the
MediaWiki API response having both the missing property set to truthy
and the page not having any revisions.
* Add the preview-empty template and associated "mwe-popups-is-empty"
CSS class, which describe what an empty preview contains and how it
should look.
Supporting changes:
* Move the original preview template into the ext.popups module.
Bug: T151054
Change-Id: Ife75bf9c6bafdfe0a6cc3e20eea853b4ac8f951b
The preview reducer would treat the LINK_ABANDON_END and
PREVIEW_ABANDON_END actions as the PREVIEW_DWELL action. Fortunately,
this was a NOOP as they would only fall through if the preview was being
dwelled upon.
HT Joaquin Hernandez.
Change-Id: I01ff47de34ac41f1e6aa46aa5baccc2a27013f1b
As it is, that test is passing the same previous state and current state
because extend by default only copies one level deep, resulting in the
manual mutation mutating both prev and current state.
This passes because the change listener always sets the link visibility
from the current state regardless of the previous state (it is doing
redundant work, but not too much).
Change-Id: I8cf125cc4f6833fbce0fe14bdcb4ef550c7e8450
Like the next test also does, so that in case it fails we get some info
given the assert result doesn't match the test title.
Change-Id: I42d9e7c329c1a0b46133423e5303068b70b270ac
Action creator changes:
* Make the linkAbandon action creator asynchronous by splitting it into
two distinct actions, LINK_ABANDON_START and _END, the latter of which
is dispatched after a 300 ms delay.
* Introduce the previewDwell and previewAbandon action creators. The
latter is an asynchronous action that mirrors the linkAbandon action.
Reducer changes:
* Make the LINK_DWELL, LINK_ABANDON_END, and PREVIEW_ABANDON_END action
hide a preview, if one has been shown.
* Make the LINK_ABANDON_END action NOOP if:
* The user has interacted with another link, or
* The user is interacting with the preview.
Supporting changes:
* Update the mw.popups.reducers#preview and #renderer unit tests to use
an empty previous state so that the tests are more resilient to
modifications of the state tree.
Change-Id: I2ecf575bbb59bb64772f75da9b5a29c071b46a8d
If the user abandons the link after the API request delay (500 ms) but
before the it resolves (~10e3 ms), then the preview shouldn't be
rendered.
Changes:
* actions: Include the link in the FETCH_START, FETCH_FAILED, and
FETCH_END actions.
* reducers: If the active link has changed, then FETCH_END is a NOOP.
Supporting changes:
* reducers: Signal that a preview should be rendered and shown with
preview.shouldShow.
Change-Id: I3dd1c0c566ec63de515174c14845d7927583ce93
The QUnit test suite now completes in ~10e2 ms rather than ~10e3 ms.
Changes:
* Make mw.popups.actions#linkDwell to use mw.popups.wait.
* Make the tests for mw.popups.actions#linkDwell complete faster, but
still asynchronously, by stubbing mw.popups.wait.
Change-Id: I5cbef0ea69bc860f75cac27c1adea3d419c1ffad
Extract core rendering functionality from the mw.popups.renderer and
mw.popups.renderer.article objects.
For now, render and show the preview when the user dwells on and
abandons a link respectively.
Supporting changes:
* Add mw.popups.wait, which is sugar around window.setTimeout.
* action.response -> action.result in the FETCH_END case of the preview
reducer.
Change-Id: I14b437e7c2f55b988837fcb2800dd61a23c29a01
Supporting changes:
* Remove the preview.previousActiveLink property from the state tree as
it's unnecessary.
Change-Id: I657decf9425a7a9e2b27a798ed60b162569661d8
OO.copy doesn't copy Element instances, whereas $.extend does. However,
OO.copy does copy properties whose values are undefined or null, whereas
$.extend doesn't.
Since the state tree contains an Element instance - the
preview.activeLink property - we need to use $.extend.
Add the nextState helper function which copies the current state tree
with $.extend and mixes in all updates manually.
Change-Id: Ie8edd9fa0cc3a62a792ed60b49288f85b3ca73e9
If the user has abandoned the link or dwelled on a different link, then
don't make the API request.
Changes:
* Fix a bug in the linkDwell reducer where the activeLink was always
being set to undefined.
* Add the private fetch action creator.
* Make the linkDwell action dispatch the result of the fetch action
creator after 500ms.
Supporting changes:
* Make the link* action creators take DOM elements rather than jQuery
instances so that:
1. Testing whether the state tree has changed is an identity
comparison.
2. jQuery instances are constructed only when required.
* Document the ext.popups.Gateway type.
* Document the Redux.Store type.
* Rename the "previews-page-title" data attribute to
"page-previews-title" to avoid confusion.
Change-Id: I0b1cf3337a6f8d6450ad2bd127cb292ebb73af4f
Supporting changes:
* Add mw.popups.registerChangeListener, which registers a change
listener that will only be called when the state in the store has
changed.
Change-Id: Ibe6934058327c7f02f7d8092e74a667a5a1c600a
Changes:
* Add sessionToken and pageToken properties to the BOOT action and
update the preview reducer.
Supporting changes:
* Move the mw.popups.createActions to ext.popups/boot.js so that Redux
is used in one file and the actions can be tested in isolation more
easily.
Change-Id: Icd61bf1aeb466899e047432bf9798e2574652830
Reducers as a whole are a WIP, but this implements a baseline from which
to add more.
Changes:
* Create ext.popups.reducers to house all reducers
* Create reducers for preview and renderer state manipulation
* Create rootReducer by combining preview and renderer reducers
* Add QUnit tests for reducers
* Move action types into ext.popups.actionTypes
* Extract rootReducer from boot.js
Change-Id: I8a2296c6846cd4b0552a485e671af1d974944195
Supporting changes:
* Automatically register JavaScript files in the
tests/qunit/ext.popups/ directory.
Additional changes:
* Fix a grammatical error in the docblock for
mw.popups.createExperiment.
Change-Id: Ieb00709e353f0b960375fdaa0ca0dcdf950f2eb9
Changes:
* Make grunt:qunit run all QUnit tests in those modules whose names
being with "ext.popups".
* Add ext.popups/index.js, which initialises the mw.popups namespace.
* Add ext.popups/userSettings.js, which contains the code that deals
with interacting with the User Agent's storage.
* Add ext.popups/experiment.js, which contains the code that that
decides whether or not the user is in the experiment condition.
* Add tests for both units, converting existing tests where appropriate.
* Remove the associated code from ext.popups.core/ext.popups.core.js.
* Finally, dispatch the BOOT action against the store in
ext.popups/boot.js.
Change-Id: I697207677304bd49c7cfe1d37bb0a4af7810f387
You can use `grunt watch` now!
Changes:
* Re-enable QUnit (but ignore legacy tests)
* Add a sample QUnit test
* Add ability to run QUnit via `grunt watch` or `grunt test`
* Move linting to `grunt lint` task
* Add Redux to globals in linter
Change-Id: Ie4a65a8a66773d6472b3d73257267d18027ff3c3
There is one reference to removeTooltips in
ext.popups.targets.desktopTarget.
I suspect this got broken in a rebase. It seems Jenkins only runs
browser tests post merge. We should fix that.
Please resubmit the patch with this amended.
This reverts commit 0ff40a6532.
Change-Id: Idd8dffb853db760ebc5866190d008f173e3025ba
Tooltips are intended to be stripped upon `mouseenter` and `focus`
events and then restored during their corresponding `mouseleave` and
`blur` events. This was broken due to duplication of event registration
and no proper deregistration.
Changes:
* Rename `mw.popups.removeTooltips` to `mw.popups.removeTooltip` to
more accurately describe its effect
* Narrow the scope of `mw.popups.removeTooltip`
* Add `mw.popups.restoreTooltip`
* Add `onLinkBlur` to `desktopTarget.js` to handle tooltip restoration
* Update qunit test to reflect changes to functions
* Minor hygiene changes regarding event namespaces
Bug: T142723
Change-Id: I776a72e436ac823fdd6b68435d9a042a91c934e5
If a user has hovercards disabled, when they right click a link
this will trigger another hover event which will reset dwellStartTime.
This means when a dwelledButAbandoned event fires shortly afterwards
the totalInteractionTime will not be correct.
To remedy this, the calculation of totalInteractionTime and
perceivedWait is moved into ext.popups.schemaPopups
Now that we can trigger events without logging to the server and
checking the total interaction time duration in two places,
let's always run the event and only check it once.
A `hover` event triggers the setting of a dwell start time
A `display` event triggers the setting of the perceivedWait value
* Both are reset on a dwelledButAbandoned.
Since dwellStartTime is controlled inside a single place, getMassageData
no longer needs to clear it.
The test "returns false if dwelledButAbandoned event without a dwellStartTime"
is removed as this should no longer be possible.
Bug: T147846
Change-Id: Ie5917ca86f0d0ab27f4cf507e6dfa2c271433c03
This adds two new events "display" and "hover" which are not
recorded back to the server. The benefits of having these events
is that they are important events in the lifecycle of a hovercard.
This allows us to debug trackSubscribe and ensure we see the behaviour
we expect to see and in a future patchset will allow us to use these
events to drive the calculation of interaction time in one single location
(Sneak preview:
https://gerrit.wikimedia.org/r/316481 to get a feel for the why.)
Change-Id: I58eefc29444179fd245cfd722093dedea19455e8
The Hovercards checkbox on the beta preferences page doesn't have an ID
anymore. It has a name though, which is used from now on.
The regression was introduced in I8636f32330e23814ba3b4c0f5e22e55aaf77883e.
Bug: T148856
Change-Id: I7baa78b0d8c560ee29c3a550628bfd47cad1c30e
Given we currently have modules defined in extension.json and in hooks
it can be really confusing understanding how the code fits together.
This change hopefully makes this a little clearer by using folder names
that are named after the resource loader modules - this is also consistent
with how we do things in our other extensions.
A images folder is added to the route so that it is clearer that the images
are not used in ResourceLoader module definitions and are only used to illustrate
the beta feature.
Change-Id: Ia650ec03e3a6d3069165441ddfa069d390be4d10
ext.popups.experiment depends on .core as it initialized the mw.popups
namespace and .core depends on .experiments for
mw.popups#getEnabledState.
By merging the experiment module into core, we can eliminate any
circular dependencies.
Changes:
* Move ext.popups.experiment.js code into ext.popups.core.js
* Remove mw.popups.experiment module and any references to it
Note: ext.popups.experiment.test.js was left in its own file for cleaner
QUnit module setups and easier removal later. I'm not entirely happy
with doing it this way, but I'm not sure changing the mw.config within
the mw.popups.core QUnit module is worth merging the files.
Bug: T146035
Change-Id: I1f024567010acaa61c1d613c6e59c998198a5976
features/popups_settings.feature:33 seems to fail because of quick
interactions after disabling+enabling hovercards.
Setting a sleep 1 after disabling and enabling hovercards (to allow the
page to start refreshing) makes the tests more resilient
Bug: T145152
Change-Id: Iaf66aaf0e2c4f8e55419d22889f91a02891e49a1
Improving Hovercards' settings dialog to fulfill design specification.
Also fixing HTML structure by removing invalid `radiogroup` element and
adding missing `</div>` element, fixing related QUnit test.
Changes:
* Add close icon instead of using text
* Style header as a table for alignment
* Remove redundant "OK" button and have it replace "Save" instead
* Update text of "OK" button to "Done"
* Fix description for translation of "Done" button
* Fix qunit and selenium tests
* Remove unnecessary markup and less
* Add mediawiki-ui-button and mediawiki-ui-icon dependencies
* Shrink dialog width some per design spec
* Fix dialog horizontal position calculation to remove hard-coded value
Bug: T138612
Change-Id: I7395e3438836149becdd576942bdaf6f21b4163f
According to caniuse.com SVG support is available
from IE > 8, Firefox > 3, Safari > 3.1 and Android
> 2.3. Android 3-4.3 does not support masking.
Out of all these browsers, considering market share
and ResourceLoader support, none of these browsers
are of concern to us. In IE8 for example we do not
run JavaScript for our end users. Thus we should remove
this fallback support.
Changes
* Remove createImgThumbnail method and its test
* Groups duplicate CSS groups
* Refactor createThumbnail function
** Leave a FIXME on some curious code
Change-Id: I59ac2e320b2e07815bc4136d5942016fdc1d4340
Bug: T135554
Use the standardised MediaWiki storage system for the simple use
case of disabling/enabling Popups. The jStorage library is 12kb
uncompressed and cannot be justified for this usage (and if it becomes
the standardised library mw.storage will begin using it)
This means that browsers with full localStorage or no localStorage
support will not be able to disable Popups. Given the current ecosystem
of the web - localStorage is widely supported - and the fact that grade
A browsers enjoy localStorage support - this is not a problem.
See https://github.com/wikimedia/mediawiki/blob/REL1_27/resources/src/startup.js#L59
Changes:
* Stop using jStorage
* Cleanup and migrate previous values in jStorage with plan for removing this
in a month.
Bug: T136241
Change-Id: I6dac2911e84d6cb20731be34add01576cf407c46
As there are an unmanageable amount of synchronous checks of
mw.popups.enabled, convert mw.popups.experiment.isUserInCondition to a
synchronous method.
Follow on I4959749.
Bug: T132604
Change-Id: Ide07e62868c77bfcd78af58dcec7303a35a72157
Changes:
* Add the ext.popups.experiment module
* Add the mw.popups.experiment.isUserInCondition function, which
returns a promise that resolves with true if the user is in the
experimental condition, otherwise false
* If the experiment isn't configured, i.e. wgPopupsExperimentConfig is
null or undefined, then the user isn't entered into the experiment,
providing a kill switch
Bug: T132604
Change-Id: I49597494273e3862711a32e4951c8598e6c8bf59
This change is an intermediate step in our transition
to logging a variety of events easily.
Also:
* Some events may need sendBeacon support and may not be
logged if the navigator does not support sendBeacon.
* Do not load schema code if EventLogging is not available.
Bug: T131315
Change-Id: Iff939577f65f1c6c71701dd6967939445385fb70
Popups currently makes use of a render cache. Using a cache in this
way without protection causes problems (I will look at cleaning this up)
The hack below it
`mw.popups.$popup.html( mw.popups.$popup.html() );`
will lead to the html of the cached object being wiped out on IE9 meaning
future loads will show an empty popup.
Add a unit test (which previously would fail in IE9) to protect against this
in future.
Bug: T68496
Change-Id: Icef784fb389b0ab1856e2ba7464c9423ebcd62ab
Changes:
* Extract survey link element creation into
mw.popups.render.renderers.article#createSurveyLink
* Make #createSurveyLink throw an error if the URL doesn't start with
https or http
* Add unit tests that cover the new behaviour of #createSurveyLink
Bug: T129177
Change-Id: I8b61cb0e94ab4e30bc894c279bb05918ecc7719e
Changes:
* Quote the URL with double quotes when generating the background-image
rule
Bug: T129177
Change-Id: I74748c7efe67954272ce0a539455b0b00001a26a
Test pieces that make sense, everything else is already
covered in cucumber tests. See the following commit for more
info: I55f311b6b8845e6ebf4cc5698758afd1f9042a45.
Bug: T133025
Change-Id: I474c1569494601ae5865dcfba22ea728220ff8df
