New changes:
ff237d4 Fix z-indexes in core
e88d43e Localisation updates from https://translatewiki.net.
cf61803 Consistently use ve.ui.WindowManager everywhere
f9dfdb8 Update OOjs UI to v0.1.0-pre (23565e7519)
f79f7e3 Update OOjs UI to v0.1.0-pre (8f8896196f)
c8201dd Update OOjs UI to v0.1.0-pre (9ed4cf2557)
Local changes for the breaking change to OptionWidget and sub-classes.
Change-Id: Ife6abd312d4dc97be67cb84eea4cb9c6a0a31b1d
In I19da270a we bumped the z-index for the overlay from 4 down to 2, to avoid
clashing with the toolbar. However, the site logo, search box and personal
tools all have a z-index of 3, so 4 is required. Instead, bump the toolbar's
z-index up by 1 to 5.
Change-Id: I7e1edcf05cde054c7bcb8c13b5633930fb5ed3b5
No combination of WebDriver for any version of Chrome on Linux
works 100% correctly.
Remove the failing tests for now.
Change-Id: I46be2c2c857e93234c839d32f1c0d4224733d0f7
Move ve.track() subscriber to its own file, and have it
route mwtiming.* events (for TimingData) and mwedit.* events
(for Edit schema) differently. Most of the data population
lives in the subscriber, so actual ve.track() calls are
pretty lightweight.
Existing ve.track() calls with timing data were kept with
their names intact for backwards compatibility, but
we may eventually want to throw them out and start from scratch.
ve.init.mw.ViewPageTarget.init.js:
* Remove old track subscriber
* Track init and ready events
* Remove old ve.track( 'Edit', ... ) crap that didn't work
ve.init.mw.ViewPageTarget.js:
* Fire the saveWorkflowBegin event before the save dialog
loads rather than after
* Remove unnecessary this.events.trackSaveError() calls:
TargetEvents already listens to these events itself
* Remove badtoken handler because all it was was an
unnecessary trackSaveError() call
* Add abort tracking
** Pass trackMechanism through deactivate() and cancel()
ve.init.mw.Target.js:
* Add static.integrationType to populate the 'integration'
field in the schema
ve.init.mw.TargetEvents.js:
* Simplify onSaveError* methods away into connect bindings
* Map track topics to mwtiming.* so they can be routed separately
* Track save-related mwedit.* events
Depends on I978eda96c in WikimediaEvents
Change-Id: Iae677d9b15c71d2b18e795bd5179d11876c06abd
New changes:
7d8ed7f Copy in some IE CSS hacks from ve-mw
a2e962e Update OOjs UI to v0.1.0-pre (20c61ec865)
d99c62f Update OOjs UI to v0.1.0-pre (d4cfcce969)
7fe02b3 Fix lots of spelling mistakes and typos
Local changes:
Remove redundant IE CSS rules since core's I013688c5
Bug: 73565
Change-Id: Ic60cd5290932ec38fab26492fffa17c3a8e91398
Make sure it always gets torn down on deactivate, not
just on save. Otherwise we end up with multiple copies of it.
Change-Id: I7b95c316641fc48ce7087a0042ec6174fe03180b
If you clicked "Read" while the editor was loading
(but only while the "Edit" tab was already active, not before)
then you could get in a situation where surfaceReady fired
on an already-aborted target, which caused JS errors.
It seems like we should clean more things up in this
case, but I don't know what they are. In any case,
we should not try to set things up on a non-activating
target when surfaceReady fires.
Change-Id: Id57bd63ff288156725e472e7d89009022090253a
Copy in the font-size:127%; rule from MonoBook's main.css which is applied to
div#globalWrapper
Caused by Id425c56d
Bug: 73660
Change-Id: I05502295b81c62fd1180dff860dea68d76c2dfa9
Follows-up Ica33de675, If505a46f54.
* The setting of 'uri' looked like a redundant local alias before
passing on to pushState, but is actually important on itself.
* Remove binding for hideLoading. It's a detachable method on
a singleton, not an instance method.
Change-Id: Ic3536caf3f42ee893124312fd5981b67336bd480
When you clicked the Back button in the browser, the URL
in the address bar would change (removing veaction=edit),
but we would not go back to read mode. This was broken
by 5c0c11753 almost a month ago but apparently no one noticed.
This is because 5c0c11753 moved the pushState() calls to
be earlier (in init init), making the replaceState() call
in the ViewPageTarget constructor (which is there
specifically for this bug) run too late in those cases.
The simplest way to fix this is to duplicate these replaceState()
calls before the pushState() calls in init init.
I feel a bit bad about copying code, but not very bad
because the code I'm copying already has a FIXME comment
about how there should be a better way :P
Change-Id: I6627a5d1d9377ae815bc58bceeb059ce9f4f19ab
Just like DOM interfaces such as document and location. And JS
constructors like Object.prototype, and libs like jQuery.
No need for the window host object.
Change-Id: Ie39aaea2ca17bdec5fb3afe32a4e6c9f751bd0c6
The location object is a global, just like document.
Using it via 'window' needlessly adds complexity and, for example,
makes it harder to catch typos in static analysis.
Standardise on location.href in place of the many different
variants for accessing and assinging the full url:
location =
location.href =
location.assign() =
And each with 'window', 'document' and without host object.
Change-Id: If88515bd88fc5b93ebacd4ec41fce02fec540f18
mw.msg can return things with <br />, which we want to be interpreted as HTML
rather than just escaped as text.
Bug: 73571
Change-Id: I7861432d9361e01f9def0578c342e60715c3d764
ApiEditPage can give us 'nochange' instead of the
'oldrevid'/'newrevid'/'newtimestamp' keys if we gave it a null edit.
Bug: 73463
Change-Id: Ic22597dfed11de3823471673404090a9bce12928
* put the shared step in shared_steps.rb
* replace "onfocus" with simple "click"
* update RSpec while we're at it
Change-Id: I77754abdbeba1314758f5fe10601f6601e8783bd
preliminary to updating Then statements to RSpec3.0 syntax
zero change to function or syntax, just arranging things so
we can review them later.
Change-Id: Id0f006d0fcf74010c06cb7ae6386ea6bcaff0e55
New changes:
6ec286e Support for generic file drop handlers
4f2ab66 Generic progress dialog for drag and drop uploads
742e0fe Fix pasteSpecial command
87ea9e9 Fix up/down arrowing off a focusable node
643ea34 spinner.gif: Update with new version from Trevor
ffd8b1e Position the table context relative to the first section node
393776a Fix logic for ignoring (col|row)span=1
f94b10e Don't unmerge spanned cells when spanned cell is removed
3416764 TableContext: Simplify row/column overlays styles
e6c0038 Localisation updates from https://translatewiki.net.
fb1a922 Register select all as a command and perform programmatically
3c1ce93 Make caption tool available when caption selected
dd0ac7d Move disabling logic from tools to commands
8bf33a5 Bold & italic icons for Armenian
058b32c Update OOjs UI to v0.1.0-pre (05f0fefc3f)
c01d115 Create filedrophandlers folder for implementations
Local changes:
* Register new files
* Move disabling logic from tools to commands
onUpdateState should only ever call setActive. setDisabled should
already be determined by the command so that it can prevent itself
being called by triggers.
Change-Id: Icdeaa4c152a0dcd839c8e7cc4e64b2215eb2de1e
After 5c0c117, clicking the VE edit tab on /wiki/Foo?oldid=123
sent you to /wiki/Foo?veaction=edit , dropping the oldid parameter
and editing the current version rather than the old version.
Fix this by only recomputing veEditUri from scratch when necessary,
and reusing the current URI when possible.
Change-Id: Iee9a745990f1a2bad40f46b8cdaaa727bb7fead8
Deep-extending originalData is asking for trouble, because there's
no way to delete anything that way.
Bug: 73134
Change-Id: I44d3e582e04942efceae3e0bdbbaa147095b5146
setMode() is called multiple times, but it didn't
check whether the mode being set was already set.
Because it's a setter, make it idempotent.
This fixes the problem where the first input will
be focused first, then blurred later, but it
introduces a new problem because the label for
the 'mode' ActionWidget is set from setMode().
To solve that, factor the setting of this label
out into its own function and call it on setup.
Bug: 73138
Change-Id: I9bb127f22f6c0b745b393c523ec42f320fc85cf3
* A subclass of MenuItemWidget from OOUI, MWLinkMenuItemWidget, is introduced that, when used in MWLinkTargetInputWidget for both
external and internal links, returns actual page links.
* MWInternalLinkMenuItemWidget is extended from MWLinkMenuItemWidget to support the internal links.
* LinkCache is being used in MWInternalLinkMenuItemWidet to stylize the links. Thus, getting rid of some old css classes.
Bug: 51205
Change-Id: I3ac18dabd514ca539fff1db3c67ae97931c3d4fc
We need to figure out where the language screenshot elements reside
merged as an interim step to share the code more easily, no change
to function
Change-Id: Icae480aaa526a9de6e6926e027c0cac0c7de997e
We possibly want to discourage this somehow but our current
functionality achieves this by doing an isolateAndUnwrap
which for a table is catastrophic.
Bug: 72990
Change-Id: I79f1348da20385dfec014afcac0fb5464580cf2f
We were attaching its removal to the wrong promise in 2 of the 3
places where it was being called.
Also this file is horrible and should be refactored. A lot.
Change-Id: If74b1629266f22e2d110305b139bef4a8e69b13b
We were setting the height to 2.25em which just happened
to be right, but it depended on the browser's default
line-height being a certain value. Now that that's changed
with unisolation, the alignment breaks.
Instead of using a magic value, use height: 100%; so it
takes the parent's height. We don't even have to make the
parent position: relative; because it already is.
Also change the width to 2em, which is a much more defensible
value than 2.25em. Previously 2.25em kind of made sense because
that was also the height (resulting in a square box), but
now we don't have that excuse any more.
Bug: 72962
Change-Id: Id617bfaafefd1fc1530fbee29be5d935ec486b92
