This patch adds ve.track; it provides a means for VisualEditor code to log
various changes of state that are of potential analytic interest. This is done
without coupling VisualEditor to a particular analytics framework by providing
a method, ve.hook.registerHandler, by which event data may be routed to a
particular analytic framework for processing and dispatch.
ve.track uses a $.Callbacks-like object for tracking analytic events which can
remember an arbitrary number of past events. This is done by maintaining an
array containing the arguments of past calls and maintaining a counter for each
callback indicating its position in the queue (i.e., how many events it has
already received.) This ensures handlers are called for each event, including
events which were fired before the handler was registered. This allows the
load-order of VE and analytics components to remain unspecified.
Change-Id: I29740fa7a0ac403e484e0acee6dfcadaf6fc4566
See ApiVisualEditor.php#parsefragment.
Error code is also incorrectly about Parsoid, method
parseWikitextFragment does a FauxRequest to MediaWiki. Fixed
error.info. Keeping the same error.code for now for consistency.
Bug: 52483
Change-Id: Ic473ae4a5c1e9706140f6ec4cc8157fadd02c318
Objective:
* Make it possible to make a toolbar without a surface
Changes:
*.php
* Links to new file
ve.ui.Toolbar.js, ve.ui.SurfaceToolbar.js
* Split toolbar into generic and surface specific classes
*.js
* Update symbol names
Change-Id: Ice063a2fb67b5ce5155cdc96a0d47af49eee48cb
Follows-up ced2a8a which moved the tab layout to the server-side
and changed it to set up "source" tab and section links always
everywhere (even if VE would not be availabe in the namespace
or browser).
The JS logic (which continues to exist to take care of pages
cached before we moved it to PHP and/or to fix up pages cached
with a different configuration in the future) didn't do this yet
causing the "Edit source" tab to inconsistently appear on pages
for anonymous users viewing pages where VE is not available.
Change-Id: Ic575b3fcef17e636adaa338abc7748a4388ed9a9
This is bit of a hack, as leading whitespace could be
significant if styled with white-space:pre.
Long term VE shouldn't be editing the user's HTML, and
should just highlight potential formatting issues.
We avoid the stripping in preformatted elements as we
expect they will have that styling.
Bug: 51462
Change-Id: I654d98e17dd604cb2a192831ff3f3597f95b9962
Specifically: transclusions, formulae & references.
Captions also allowed to contain other images and reference lists
but we should probably not encourage that.
Change-Id: Ia1c99e51640a3f9ee53281a1e3ebd67dc13c8990
Because trying to work out how broken your test case is
by looking at two (potentially multi-page) serialisations
of the DOM summary is a pain in the arse. Diffing two
HTML strings may highlight the problem much more clearly.
TODO: Is it possible to defer the calculation of the HTML
infused summary object until we have determined the assertion
has failed. Otherwise we're slowing down our tests for no
reason.
Change-Id: I873bf2479ab81d15389792bd59d15580da63941a
Code speaks for itself, see also bug 52441.
Though not introduced by 14343c7bf7, that made the bug worse.
Bug: 52441
Change-Id: Ie2b80b22df03eb563de8812a47fb25152527e786
For configured wikis, show a dialog that welcomes the user to the
amazing and fantabulous world of VisualEditing, which is not only full of
wonderment and joy but also may lead to increased popularity and love.
The dialog only shows up once (uses a cookie).
Change-Id: I8e7c4dc2c63b36594378a543b9d66291395eebcf
* Generate the edit tabs and the section edit links in PHP, with a
fallback in JS for cases where we don't have them yet due to
caching. But only change things if VE is enabled, and have the JS
correct the state if the wrong cached HTML comes through.
* Make the order of the tabs/links and the messages to use as captions
configurable
* Make the edit tabs and section edit links always be present in the
page (regardless of namespace, user prefs, etc.) but be hidden and
have JS unhide them (using html.ve-available) if appropriate
* Add appendix messages so we can do a superscript "beta" even in places
where we can't use HTML in the message
VisualEditor.php:
* Add new hook registrations
* Remove edit link caption messages from the init init module because
they're now added dynamically in VisualEditor.hooks.php
* Add a noscript CSS module so we can hide some things in JS-less
environments
* Remove $wgVisualEditorTabLayout and replace it with
$wgVisualEditorPosition
* Add config vars for link captions, with null causing us to use
the default caption
* Add config vars for link caption appendices. Too many config vars
but we'll clean that up later
VisualEditor.hooks.php:
* Dynamically add tab messages to the init init module
* Remove unused globals in onBeforePageDisplay()
* Add noscript CSS module
* Add a SkinTemplateNavigation hook that changes and reorders the edit
tabs as appropriate
* Add a DoEditSectionLink hook that overwrites the edit section links
* Export the new config variables to JS
VisualEditor.i18n.php:
* Add beta appendix message
* Add a message for the default VE edit section link
ve.init.mw.ViewPageTarget.init.css:
* Remove the animation on the edit section links
* Darken the color of the brackets and the pipe from #ccc to #555
* Style the beta message to be superscript-like (but not real <sup> to
avoid moving the baseline)
ve.init.mw.ViewPageTarget.noscript.css:
* Hide the VE edit tab, the pipe and the VE edit section link initally
unless and until JS unhides
ve.init.mw.ViewPageTarget.init.js:
* Toggle .ve-not-available / .ve-available
* Edit tabs
** Only generate the the edit tabs if they're not already there from PHP
** Rewrite the edit tab generation to mirror what's being done in PHP
* Section edit links
** Same as for edit tabs
** Also add mw-visualeditor-expanded to pad the brackets
ve.init.mw.ViewPageTarget.js:
* #ca-ve-edit is now always the VE tab (and #ca-edit always the
edit source tab) so update the .selected behavior accordingly
Change-Id: Idcb15faea7fabe5fe7578b1508079969b27d2469
In some cases this would make VisualEditor be way too aggressive in
forcing you into the first matching page.
Bug: 52420
Change-Id: Ie0a793853d884ee0abf490a99c5214082e9dbf27
Added wgIsArticle to isViewPage checks, otherwise we were attempting
to load VE dynamically on Special:Move/Delete, which resulted in a
broken h1 title and odd behavior off the 'Read' tab.
Also added !wgIsArticle to pageExists. This is a bit of a hack because
we don't have any info on page existence in the JS on these pages
(I think?). But it's better to display 'Edit' for a page that doesn't
exist, than 'Create' for a page that does.
Bug: 49000
Change-Id: Ib47e5524d41e6066b362e0f5645750c769de5193
By testing against a regex of legal title characters we can determine
if the entered text is a valid internal link. If it isn't we should
prevent the link inpsector from creating/changing the annotation.
Bug: 33094
Change-Id: Ia1df602601e4e82fc351279e432c28c425f5157a
Not having a description yet is fine, but they should at least
be indexed as blocks so that they are searchable and listed
in the jsduck generated pages. jsduck defaults to @method + name
of prototype property. And it even guesses parameters sometimes.
Search: \n\n([a-zA-Z\.]+\.prototype\.[a-zA-Z]+)
Where: modules/ve,modules/ve-mw
Where-Not: modules/ve/test
Replace: \n\n/** */\n$1
Added @return in a few places where it was easy to add.
Change-Id: I830c94cc7dbc261bd7a077391f930cbfff165f9d
mw.ViewPageTarget is currently getting events from both the
platform target toolbar and context menu toolbar because the
event is emitted from within the toolbar to the surface.
Instead we're now emitting it on the toolbar itself and it is up
to the binder to access the correct one and listen to its events.
Bug: 52317
Change-Id: Ibd8053768e82b1df91081bd77a172628ea855db7
Follows-up 867ec44a9.
Because:
- #toolbarPosition is emitted more than once (so this
should at least have been a once() bind, not a connect)
- It is emitted for more than 1 toolbar (includes context menu
toolbar)
- Semantically incorrect (we want to know when the surface changes
not the toolbar)
- We want to get rid of that event entirely and this is the last
use of it.
Change-Id: Ica5ed04052f48fe84607abab72bcf65f97d689ed
Follows-up db3da5d. Also scanned code base for more matches but
looks like those have been cleaned up already.
Change-Id: I62d1c4367e1ed7b7162d7b688227dfa5dca562aa
Though this is already handled by Ie50b63ba5064e85d26 for the
server-side, and that should automatically reflect to the
client-side. Since we're dealing with cache conditions in
wmf-production where the user.options manifest does not yet
contain an entry for these relatively new preferences, lets
transfer the default value here as well. This does not affect
logged-in users (since their user options are always up-to-date
and embedded in the page).
Also made ve.support.es5 be casted to boolean. Previously it
would be a reference to JSON.stringify (if supported), or the
bottom value of whichever feature the browser didn't support.
Doesn't change any behaviour but should make things slightly
more performant when this value is evaluated.
Change-Id: I9ca430051ae6f4e603c2d89982e540e455055255
On browsers that implement the Navigation Timing API, performance.now()
provides values with microsecond precision that are guaranteed to be monotonic
(i.e., they are not subject to skew due to changes to the system clock).
This patch adds a `ve.now` utility function that will use this API when it is
available and fall back gracefully to `Date.now` when it is not.
Change-Id: I377025fcb23cb26399b9e437e33c8afa138916af
In target#setUpSurface, both target#setUpToolbar and
target.surface#initialize are called. #setUpToolbar does an
asynchronous animation.
After that animation is completed we call target.toolbar.initialize
and target.surface.context.update.
Right now ce.ProtectedNode needs to update the position of its
shields when the CE Surface changes position (which it does when
the UI Surface changes position because of the UI Toolbar changing
position), and does so by listening to toolbarPosition.
Adding this event to allow it to listen to that instead.
Change-Id: I826986794630c04c34cef6da36ccb15ff7dde49a
Formerly known as "The greatest commit in the history of the world*".
* Within a 3 block radius of Drayton Park and Auburt Park, starting
from July 30th at about 9pm or so.
Bugs:
* (bug 51404) Allow escaping out of the link inspector when in creation
mode (no text is selected, text will be inserted based on link target)
and the text input is empty
* (bug 51065 and bug 51415) Keep model and view in sync when changing the
link inspector's text input value and showing options in a menu
* (bug 51523) Either restore selection at the time of close to what it was
before opening the inspector (when using back) or to what it was before
closing (might be changed by transactions processed during the close
method) - this makes it simpler and more natural when clicking away from
the link inspector, even when there are changes that must be saved by
the link inspector on close
Bonus:
* Use only the light blue highlight color for menu widget items - the
checkmark already displays the selected item, the dark blue is just
masking the current highlight position and confusing the peoples
* Remove links when the user deletes everything from the link inspector's
text input and then closes the link inspector
* Replace select menu's evil "silent" selectItem/highlightItem argument
with a new method called initializeSelection which sets both selection
and highlighting to an item without emitting events - this is needed
when synchronizing the view with the model so the model isn't
immediately told to change to a value it already has
* Make the MWTitle lookup menu not flash like crazy as you type (this was
caused by a copy-paste oversight overriding
initializeLookupMenuSelection unnecessarily)
Bug: 51404
Bug: 51065
Bug: 51415
Bug: 51523
Change-Id: I339d9253ad472c2f42c3179edc84a83d27561270
The 'ext.visualEditor.genderSurvey' module was introduced in I2b4aba6a9
for the purpose of instrumenting a short-lived microsurvey that would be
concomitant with the deployment of VisualEditor. The survey has now run
its course; this patch removes its code.
Change-Id: I8be6198a66957d792757a5312e9e71b3c8cdd1e3
This code caught badtoken errors on load, but we can't get those
any more since the API module was split and we no longer send a token
on load. badtoken handling on save is done in mw.ViewPageTarget.
Bug: 51253
Change-Id: Idb172ee46b7142681d41e593eacd9600b8f11a9a
7557dd39ed make the badtoken handling code unreachable. Revert that
change, and fix the rest of the function to deal with the possibility
that editApi is undefined. Let handling of the read-only mode error
and any other errors fall through to the bottom of the function.
Change-Id: I0673f2bb629e5cc9449675c1074d283e3926e1d5
VisualEditor.php
* Add CSS file
ve.ce.MWMathNode.js
* Wrap the image in a span, so GenerateContentNode doesn't
try to nest an image inside an image
* Remove unnecessary attribute setting
* Only pass unwrapped image to deferred.resolve
* Retrigger MathJax rendering
ve.ce.Node.css
* Use inline-block for image wrapper
ve.dm.MWMathNode.js
* Mixin GeneratedContentNode and implement getHash
* Copy over functionality of MWTransclusionNode:
+ Just store data-mw for attributes
+ Store orignal(DomElement|MW|Index) for selser
ve.init.mw.ViewPageTarget.js
* Add mwMath to the toolbar
ve.ui.MWMathInspector.js
* Remove static.InputWidget, not required in this architecture
* Use multiline TextInputWidget
* Only update mw attribute
* Allow creation of new math nodes
ve.ui.MWInspector.css
* Set height of TextInputWidget
Change-Id: I520f8ccc9f89a2ce70aa1d9e02ed0c6cacbecc2f
It's totally unnecessary, and actually caused a bug where new references
got about="undefined", which caused Parsoid to about-group adjacent
new references together.
Bug: 52228
Change-Id: Id40d53c72a35412d612ae9441ae3d561622c1bec
When we encounter a ref tag inside the mw-data of a references
tag, we pass it off to the converter and store it as nested data.
In toDomElements we convert any nested children and write them
back to mw-data if changed.
As refs in references are invisible we exclude them when generating
the references list in ve.ce.MWReferenceListNode.
Bug: 51741
Change-Id: I31d06616849a00449df0fc77f3b33e46207cdc7f
op !== operation in all iterations of this loop except the first,
and using information from one operation to apply another one
doesn't work very well.
The fact that undefined cast to a number is NaN rather than 0
is very unhelpful :(
Also fixed some commas that should be semicolons.
Bug: 52238
Change-Id: I4138c023c955f2866881084506e24bb8b6db5a4d
Now Ctrl+\ (Cmd+\ on Mac) will trigger the 'clear annotations' button
on the current context. Ideally we'd also bond to the 'clear' keyboard
button (ASCII 12) but it does not seem possible to do that yet.
Bug: 51507
Change-Id: I300ec1ffa237e51418ec429be39001f820f053ae
== Renamed methods ==
* enableFloating -> enableFloatable
* disableFloating -> disableFloatable
* setPosition -> float
* resetPosition -> unfloat
== Scroll and resize event ==
Timeline for scroll event reduced from about half a dozen
"Recalculate style" and various forced "Paint" down to 0.
New timeline for scroll is clean (for me: from ~35 to ~59 fps):
* 1 Event (scroll)
* 1 Composite Layer
The composite layer action is the browser changing the viewport
to a different portion of the document drawing. Exactly the one
thing a simple scroll should do.
Timeline for resize event is still pretty crowded and low fps,
but it has improved. Further improvement would likely be around
using requestAnimation and going outside ve.ui.Toolbar.
== Changes ==
* New: ve.ui.Toolbar#initialize.
Similar to what surface has. Users of Toolbar should decide
whether to call enableFloatable, append it to the DOM at some
point and then call initialize() once.
* Don't compute offset() every time.
Eliminated by doing it once in #initialize. These 'top' and
'left' offsets do not change.
* Don't compute outerWidth() and $window.width() every time.
Reduced by doing it once in #initialize to compute the 'right'
offset. Updating it only on resize.
* Don't set 'top' every time.
This is already in the stylesheet. It was never set to anything
else so the abstraction for it in #float has been removed.
This also made it obvious that code for "ve-ui-toolbar-bottom"
was unused and left behind. Tha class was only ever being
removed from something (never added).
The one addClass call for it was in a condition that is always
false ("if top > 0").
* Don't set 'left' every time.
Eliminated by doing it once in #float.
* Don't set 'right' every time.
Reduced by no longer doing it on scroll. Done once in #float,
and on resize after computing the new value for it.
* Remove no-op style operations.
Wrapped methods in if-floatable, if-floated etc. to reduce a
fair amount of easily avoided re-paint overhead.
* Avoid double re-paint in mw.ViewPageTarget.
Though we prevent a lot of redundant re-paints now, whenever
we do repaint we want to do it in 1 repaint instead of 2.
ve.ui.Toolbar emits #toolbarPosition, which tells
mw.ViewPageTarget to update toolbarTracker which would read
the new $bar style properties and copy them over to the
$toolbarTracker. However, this read operation forces the browser
to do an immediate re-paint half-way just for $bar.
Browsers only repaint when style properties are changed and
JS has yielded. The exception to this is JS reading style
properties: in that case the browser is forced to do those
deferred repaints directly and reflect the new values.
We can avoid this double repaint by passing the updated values
as data instead of requiring the receiver to read the DOM (and
thus a keep the deferred repaint). Now toolbarTracker can use
them directly whilst the browser hasn't even repainted $bar yet.
== Clean up ==
* Redundant "border-radius: 0". This would reset something, but
it never does. None of the things it inherits from set a
border-radius. There is one subclass where toolbar is used
with a border-radius (".ve-ui-surfaceWidget .ve-ui-toolbar-bar"
sets a border-radius) which overrides this on purpose, so the
default of 0 is redundant.
* Pattern "if ( .. ) addClass() else removeClass()" changed to:
"toggleClass( , .. )"
Bug: 52014
Change-Id: I9be855148962eee068a77fe83e98eb20bbdcfeec
These have been pointing to the same method for a while now,
we can safely remove these obsolete aliases and just use it
as generic copy.
* Each file touched by my editor had its new line at EOF fixed
where absent
* Don't copy an otherwise unused empty object
(ve.dm.Converter)
* Use common ve#copy syntax instead to create a link
(ve.dm.Document, ve.dm.example)
* Remove redundant conditionals for isArray/copyArray/copyObject
(ve.dm.example)
Change-Id: If560e658dc1fb59bf01f702c97e3e82a50a8a255
The core changes to postedit in I778b18b that this depends on were
deployed to the cluster as part of 1.22wmf11.
Bug: 39632
Change-Id: Id4a8bc22c09a552ef79670b0d4fc4a70df07ec33
Opera 12 seems to work well enough, but I'm not confident enough to
whitelist it just yet.
Opera 15 is basically Chrome with a different interface, so it should
work perfectly, but it's barely out of beta and untested right now.
Bug: 36000
Change-Id: Ia80a6f53f8c128ef52d0bfde1828fdc132046afb
This is the language inspector UI engine with ULS core.
The Language Inspector works alongside ULS to choose and change language
blocks in text. The inspector was based on ve.ui.TextInputWidget and
now changed to inherit ve.ui.Widget and display details in a table
instead of an input textbox.
Added jQuery.ULS module:
* Repository: https://github.com/wikimedia/jquery.uls
* Latest Commit 728f112ffc90b03b50c0109487886a2647f12020
* Taken 'src' / 'images' and 'css' folders into modules/jquery.uls
Bug: 47759
Change-Id: I3c9fd6c135c05a54f6c7fc28c5962fc0a6677806
Annotations' attributes might contain DOM elements, which cause infinite
recursion in ve.compare(). Annotation classes can protect against this
by overriding getHashObject() to summarize DOM nodes, but that doesn't
help if that's not respected everywhere.
Instead, compare the hash objects, those are safe. This does not appear
to be a problem in practice, currently, because the nowiki annotation
is experimental, oo.compare() now short-circuits if a === b, and because
of optimizations in openAndCloseAnnotations() which lead to the relevant
compareToForSerialization() code path being taken very rarely.
Bug: 51948
Change-Id: If1bcc3eee4fd14d107db1935d89dcc5516643b53
This only affects debug mode, but things look broken when the
background of every single node is set to transparent (e.g. a
<pre> looks weird with a white instead of grey background).
It also leaves the DOM dirty full of inline styles.
Though setting a grey background isn't guaranteed to be visible
either, and all of these redraws and stuff really slow things
down (we should perhaps only start doing these after the initial
document is painted).. the least we can do is undo it and not
leave it there.
Change-Id: I9abfd46765914828ad8618748be5716a8c6b185c
By using annotation indexes only we can avoid a lot of
ve.getHash calls. This reduces the number of getHash calls
on load of [[:en:Argentina]] from ~60,000 to ~2,000.
Bug: 52013
Change-Id: I0bc9aa8feea5f7e4e90a5fcd829de57cab803c15
We would dirty-diff "</span>\n<!-- comment -->\n<span>" to
"</span>\n\n<!-- comment --><span>", i.e. the second newline made
a bunny-hop to the left over the comment.
The actual bug turned out to involve a double bunny-hop, with
"</span> <!-- comment -->\n<span>" turning into
"</span>\n <!--comment --><span>", i.e. the newline bunny-hops
both the comment and the space.
This happened because outputWrappedMetaItems() didn't take
wrappedWhitespace into account when restoring meta items and
associated whitespace. I hacked a check for wrappedWhitespace into it,
but we should really just rewrite this pile of hacks into a unified
system for queuing and processing both whitespace and metadata.
Change-Id: I4375f4c07983ffec6877d0371aeaa9bf6e65fd6e
To avoid confusion between IV store indexes and the index
within the set, rename them to storeIndex and offset.
Change-Id: Ic7d741bd5d39240d63fdc04a2df45658a64441de
As that method makes expensive ve.compare calls, we can quickly
avoid most cases by testing AnnotationSet#contains first.
On page load this reduces the number of ve.compare calls on
[[:en:Argentina]] from ~6,000,000 to about ~2,000.
Also reduces ve.compare calls per backspace keystroke from
~300 to 2.
We can optimise this further, but this is a good simple start.
Bug: 52013
Change-Id: Ie3b4517fd13383c48acb64b3c4e82051c34e7484
Previously, we'd clone the data but convert it in the context of
the existing dm.Document, whose nodes had pointers to elements in the
old data array, not to the cloned ones. Because dm.MWReferenceNode
has logic like if ( something === dataElement ), this caused the sanity
check conversion to behave slightly differently compared to the real
conversion that happens on save, and so a references corruption
bug went unnoticed.
Change-Id: I79a42ae21f91cb8eb410ae26ea638036db19e217
Pretty straightforward, although we should start thinking about
grouping/hiding 'advanced' formatting options in the toolbar.
Making this button experimental for now until we've come up with
a way to deal with this problem.
Bug: 51590
Change-Id: Ieb1935b742aced4b883d8a194e6cb69be68473d0
Server-side, plugins can register themselves by adding to
$wgVisualEditorPluginModules. This is the recommended way for
MW extensions to extend VE. Client-side, plugins can register
themselves through mw.libs.ve.addPlugin(), which takes a string
(RL module name) or a callback.
When VisualEditor loads, we load the registered plugin modules in
parallel with ext.visualEditor.core. Note that they're loaded in
parallel, not after, and so the plugins should explicitly depend
on ext.visualEditor.core if they use or extend classes in VE core.
Once the modules finish loading and user and site scripts have run,
we execute the registered plugin callbacks. These callbacks can
optionally return a promise. We gather these promises and wait for
all of them to be resolved, then initialize the editor.
This allows Gadgets to extend VE by top-loading a small module that
depends on ext.visualEditor.viewPageTarget.init and calls
mw.libs.ve.addPlugin( 'ext.gadget.bottomHalfGadget' ); , the bottom
half being a hidden Gadget that depends on ext.visualEditor.core and
contains the actual code. The addPlugin() call needs to be in a
top-loading module because otherwise there's no guarantee that the
plugin will be registered before the user clicks edit and VE loads.
User and site scripts can extend VE by simply calling addPlugin()
directly, as mw.libs.ve is already present when user scripts run (since
it's top-loaded) and VE waits for 'user' and 'site' to run before
executing plugins.
If user/site scripts need to load additional JS files, they can load
these with $.getScript() and return the corresponding promise:
mw.libs.ve.addPlugin( function() { return $.getScript( 'URL' ); } );
For a diagram of all this, see
https://www.mediawiki.org/wiki/File:VE-plugin-infrastructure.jpg :)
VisualEditor.php:
* Add $wgVisualEditorPluginModules
VisualEditor.hooks.php:
* Expose $wgVisualEditorPluginModules in JS
ve.init.mw.ViewPageTarget.init.js:
* Add mw.libs.ve.addPlugin function that just stores the registered
values in an array and passes them into the mw.Target when it's
being initialized
ve.init.mw.Target.js:
* Add $wgVisualEditorPluginModules to the set of modules to load when
initializing VE
* Add a Deferred (this.modulesReady) to track module loading
* Add addPlugin() and addPlugins() methods that add to either
this.modules or this.pluginCallbacks
* In load(), instead of mw.loader.load()ing this.modules, use using()
to load this.modules plus user and site, and fire onModulesReady()
when they're loaded
* In onModulesReady(), execute the registered callbacks, gather the
returned promises, wait for all of them to be resolved, then resolve
this.modulesReady
* Fire onReady based on this.modulesReady being resolved, rather than
using a second using() call
Bug: 50514
Change-Id: Ib7d87a17eaac6ecdb8b0803b13840d7ee58902df
