What I learned today:
* Window doesn't have a scrollTop property, body does (that's why animate
doesn't work on window)
* jQuery.scrollTop() doesn't work on body (in firefox) but works on
window everywhere
* jQuery.scrollTop() uses scroll offset, not the scrollTop property
* Body doesn't have an onscroll event, window does
What I really learned today:
* Browsers are very poorly designed
Objective:
* Make clippable elements properly resize in Firefox when scrolled
Diagnosis:
* Scroll events were not being emitted from the scrollable container
after the merge of Ifec0dae598f7fd99270588bd8ca77777a07e9669 because
such events are not emitted from body tags, only scrollable divs and
windows
* jQuery.scrollTop was giving incorrect values when called on the body
instead of the window, so also due to the aforementioned change, the
clipping was being calculated incorrectly
Treatment:
* Add $clippableScroller property, which is either a scrollable div or
the window (could this have side-effects if someone did something
ridiculous like made the body absolutely positioned and overflow:auto?
Yes, but I have no other option and that's a strange edge case don't
you think?)
* Use $clippableScroller for listening to scroll events and getting the
scrollTop value from jQuery
Bug: 55343
Change-Id: I819aba60b200059886b347115fda437b3dc9cb7a
Objective:
* Remove surface dependencies in dialogs, inspectors, windows and window sets
* Introduce surface-specific versions of dialogs, inspectors and window sets
Change-Id: I2db59127d2085b02e173a3605e174317e419e213
After 6ec34a3de the edit notice button was no longer hidden by
default if there were no notices. The alert icon was always
visible (when clicked it would show "0 notices").
In addition, on any page (except pages with edit notices) it
would throw a fatal exception at load time because method .hide()
doesn't exist.
As a result, current master shows an incomplete toolbar (e.g. not even
a Save button!)
Change-Id: Ib6e91c4756664c25fbb7403ef54b4fffcc0f9938
The header should comprise 'key=value; ' pairs rather than 'key: value', and
the key name 'cached-response' is less tied to the particulars of the current
Parsoid storage setup (which is being revised). Follows I8085fdba3.
Change-Id: I8bcf947763512ba700793974ab0c48d65c171bbb
That commit removed message 'visualeditor-help-tool' for some reason.
Could be a race condition (relatively new message), but unlikely as other
translations of that message were not removed and are still in the file.
Anyway, reverting so they can try again. And to avoid this messages's
translations from being deleted next time translatewiki reads this from git.
This reverts commit 73f8299046.
Change-Id: If457b61e3fd3703a59ee67a9a0c26b0c9a36d0d4
Class ve-init-mw-viewPageTarget-pageTitle added various
transition settings that were never used. Meanwhile, we're doing
fadeTo which sets inline opacity css every X ms until the
animation is finished.
* Changed the inline jQuery animation to use css transitions
instead.
* Removed the inexistent and obsolete ms-transition declaration.
* Removed ":visible" from selector query. This makes the selector
more performant (since :visible is a proprietary Sizzle keyword)
and it was obsolete anyway. The classes don't affect display none/hide,
this is handled naturally by the browser now.
Change-Id: Ibdfb442ff6c743ef16b514a7696796ee27821887
When deactivating before the surface became active (e.g.
this.active is still false, as case is the case when Parsoid
isn't running), the teardown sequence was incomplete.
Most notable, the page title (h1.firstHeading) was still dimmed
after cancelling the alert for Parsoid error, eventhough
everything else was shown and restored.
* Moved call to #showTableOfContents in #deactivate up for
consistency with #activate.
* Added call to #restorePageTitle in #deactivate so that the
title is restored even if the surface didn't activate yet.
* Removed calls to various methods in #tearDownSurface that
were already called by #deactivate.
Now activate/deactivate and setUpSurface/tearDownSurface are
in balance.
Change-Id: Ibb2fbf0e5ab9b6a028d4e139c13aa7ff8c82be82
Previously we returned ElementLinearData from the converter, then
stripped out the MetaLinearData. This meant that before processing
the ElementLinearData from the converter actually contained metadata
which is confusing.
The new document constructor stores the converter results in a
FlatLinearData object and simultaneously populates element and meta
data stores.
Also in this commit I have moved various methods from ElementLinearData
to FlatLinearData, from which ElementLinearData inherits.
Change-Id: I64561bde2c31d8f703c13ac7b0a0c5f7ade9f3d4
Objectives:
* Use widgets to render toolbar actions
* Remove labels next to help notices and edit notices buttons
* Add a close button to the help notices and edit notices
Overview:
* ve.ui.ButtonWidget is now abstract, use ve.ui.PushButtonWidget instead
* ve.ui.IconButtonWidget now inherits from ve.ui.ButtonWidget
* ve.ui.PopupWidget's display method no longer takes x and y arguments
* Fixup naming issues in MWCategoryPopupWidget
* Fixup naming issues with some ve-init-mw CSS classes
* Rename ve-mw/ui/styles/ve.ui.Widget.css to ve.ui.MWWidget.css
* Change uses of "callout" to "tail"
* Add hyperlink functionality to buttons
* Make buttons accessible through focusing, but make unfocusable by
clicking
* Add head option to popup for rendering a title and close button
Bug: 52386
Change-Id: Iea2c8df1be64d40f9c039873d89ee540cc56e687
Objective:
* Use <body> instead of the closest surface overlay as the default
overlay (wasn't actually being used anyway)
* Use this.$.offsetParent instead of closest surface when measuring
parent frame offset to position text input menu
Change-Id: I04fc5ff1da4bafd342093922a5bd3b3a784b327c
This commit extracts the page and outline handling logic from
ve.ui.PagedDialog (RIP) and moves it into two layouts: ve.ui.PagedLayout
and ve.ui.PagedOutlineLayout, respectively. These layouts are now implemented
inside the dialogs that used to mixin the ve.ui.PagedDialog class. This
brings a much cleaner separation of concerns between Dialogs and Layouts
and allows the use of page handling logic without the accompanying
outline logic.
Change-Id: I5efa2f893f4b7e962438b3aff34b737573bbd5ca
Change I1d7fd94f2 to Parsoid adds a custom header on Parsoid HTTP responses,
'X-Parsoid-Performance', which contains data about how long it took Parsoid to
handle the request. This patch makes the API pass through the header from
Parsoid to the client, where it will be logged as an event (along with
additional datapoints that are only available on the client). When the request
is fulfilled entirely by Varnish, the XPP value is replaced with
'backend-varnish-hit: true'.
Change-Id: I8085fdba35c860eaaa4df020d06d1d45d4ea5813
Ignoring all bounced change events led to the logic for unhighlighting
the previously selected FocusableNode being skipped. This caused a bug
where if you clicked a FocusableNode, it would stay highlighted even
if you then selected some text, until you highlighted another node
(which would then be highlighted forever, etc.)
Change-Id: Ia8d74ef85eaa47326d49ef6c0f395b44b90da4dc
Objective:
* Make ve.Factory behave like ve.NamedClassFactory
* Remove the only remaining use of ve.Factory (actions)
* Remove ve.NamedClassFactory
Change-Id: Ie302ef5ea31081de7ab0db6091058a59946aef4c
Objectives:
* Pass a specific tool factory into a toolbar, allowing it to be used
with different collections of tools and not depend on the
ve.ui.toolFactory global
* Move syntax highlight editor tools to their own factory
Change-Id: I307bf180bd6817bc044bc474a77861e13f431ddb
InternalList.clone() assumed that all properties are automatically rebuilt
when a new document is built, but that's not true for .nextUniqueNumber
(or for .itemHtmlQueue for that matter). This meant that, in practice,
.nextUniqueNumber was being reset to 0 after auto/N numbers for existing
references had been assigned, but before assigning numbers to newly
created references. This caused all sorts of naming collision fun.
Bug: 54712
Change-Id: I1d087a5f3c23979d7d488e3ab32eb064ebc23e94
Document slice only ever contained linear data, with extra functionality
to preserve the range. It pre-dated LinearData, but now we should
refactor it to reflect its purpose.
Change-Id: Ifc908f7526c83a43a51372c8d2494d7260e7facd
We already getSlice which returns a ve.dm.DocumentSlice, so using
the word slice in this method is very confusing. What we are actually
doing is creating a ve.dm.Document from a range. Also remove argument
overloading as it's not particularly helpful and would make the new
name a lie.
Change-Id: I93da3419510410b170396e6765fbe2a87f9795be
The way we implemented undoing transactions was horrible. We'd process
the original transaction, but with a reversed=true flag. That meant we
had to keep track of the 'reversed' flag everywhere, and use ternaries
like insert = reversed ? op.remove : op.insert; all over the place to
access transaction operations. Redo then worked by reapplying the
transaction. We would verify that this was OK by tracking whether the
transaction was in an applied state or an undone state.
This commit makes it so every transaction can only be applied once. To
undo, you obtain a mirror image of the transaction with tx.reverse(),
then apply that. To redo, you clone the original transaction with
tx.clone() and apply that. All the code that had to use ternaries to
check whether the transaction was being applied in reverse or not is
gone now, because you can only apply a given transaction forwards,
never in reverse.
Bonus:
* Make ve.dm.Document's .completeHistory a simple array of
transactions, rather than transaction/boolean pairs
* In the protection of double application test, clone the example
document properly; it modified ve.dm.example.data, which was "fine"
because it ran .commit() and .rollback() the same number of times
Change-Id: I3050c5430be4a12510f22e20853560b92acebb67
