Because sending HTML like <h2> </h2> or <h2></h2> to Parsoid
produces undesirable output like == == or ==<nowiki />==
Bug: T51452
Bug: T52100
Bug: T57769
Bug: T61647
Change-Id: If15a1b4b31d4f08c23ecdf2ecf61a8a14a77259a
New changes:
68e20d4 Create AlignableContextItem to quickly adjust AlignableNodes
0150df2 Update OOjs UI to v0.9.3
b333fd3 FragmentInspector: Execute action on enter, rather than closing with data
2d14f7a Fix webkit column hack
be780eb Don't drop whitespace when removing empty slug paragraphs
cc19787 Split handlesOwnChildren and ignoreChildren
7f1c9a7 Localisation updates from https://translatewiki.net.
Local changes:
Add ignoreChildren to handlesOwnChildren nodes
Change-Id: Id3dc7efae8d30b6551b2fc3104ed00bc86339176
Since all transclusion nodes can be interacted with (including the
'hidden' ones,) there is no need for MWTransclusionMetaNodes.
Change-Id: I23d37e3d82029b7475ec68ebb04883c7e05370cc
By removing from and re-inserting into the InternalList when the right
attributes change, rather than trying to hack it into the model's
updateInternalItem (which won't get run on undo).
Bug: T71119
Change-Id: Iaf7a3127576b94666e05691902d2782394d24afb
Also appending both the default value and the example value to the
field description. The input placeholder will show the default
value, the example value, or nothing.
Related to TemplateData change I9e2d66a805e25 but works without.
Bug: T53049
Change-Id: I4e5e57546f6d3c22e778ba061a5344ddee085c17
Also, fix @returns comments (should be @return) and remove unnecessary
@method comments from the documentation.
Change-Id: Icd303626ac745c7ab5bff164f9b8cac276de1523
We unconditionally retrieved originalHtml from itemNode,
even if there was no data-mw.body.id, so as long as
the reference itself wasn't edited, we would never overwrite
data-mw. This meant that if the <ref> with the contents
was deleted, the contents wouldn't be transfered to the
first <ref> tag with that name like they're supposed to be.
The solution is to only look at originalHtml from itemNode
if data-mw.body.id is set.
Change-Id: Ib87491b6fa6a77d62384158f8e8f7dcc2a36c23a
For the masonry view, the result widget resizes itself according to
the rows and other images in the row. Up until now the resize was
a bit buggy, occasionally missing height/width values. This is
mostly due to rounding errors, but also due to a small double
calculation that is now fixed with this code.
Change-Id: If18adc5280e4bdbb9174b7c7e6e4eadf3c1ab26d
Make sure that the validation method accounts for local api
vs remote api when checking for valid variables.
Bug: T90795
Change-Id: Ia6877ca6ccc4e65d1db053f107e7702721e2da3c
New changes:
0ceafd5 Fix V8 deopt for ve.ce.BranchNode#onSplice
01d2d4b [BREAKING CHANGE] Store originalDomElements instead of htmlAttributes
96005a7 Localisation updates from https://translatewiki.net.
42e672e ve.ui.Toolbar: Don't call #updateToolState from #initialize
373562d Localisation updates from https://translatewiki.net.
Local changes:
Update for htmlAttributes going away
Change-Id: I576a1e5d5883b03d1fff6666926f80a5682a280f
Generalize MWResourceQueue and MWResourceProvider so they are
extending the general classes from ve-core. This also allows for
tests in ve-core on the resource queue system.
* Depends on ve-core change I5346081317e *
Change-Id: If34cd99334f4ccb93015c3646f68ae3040b9c5d7
<link> and <meta> tags are thrown away by the clipboard so make sure the
data-mw element is converted to a span.
Bug: T86401
Change-Id: Ia2869b9496ea5ce4f474e254bb4125f85152c7cf
Iff in clipboard mode:
* Set data-ve-no-generated-contents on the main element
regardless of whether we are using orignal DOM elements.
* Set data-ve-ignore on siblings, again, regardless or
where they came from.
Bug: T89658
Change-Id: I82ad60f0b90e0c6ce88c42737bd36d1a289bcf34
These are currently in page-specific mw.config. As these are
not related any page, they should be in the global export.
This is needed to have our tests (and code base) run error-free
without a MediaWiki page.
Bug: T89434
Change-Id: Ic42bcd14028772ff0c06e28c76ffaf156f8b985f
Empty file pages are returned as results in search but have no
useful information, and, more importantly, no 'imageinfo' property.
Make sure these are skipped by the provider.
Bug: T89303
Change-Id: I3374f7373efc1f5751acb16819973594abf56a56
Add the display for the actual uploader, and change the label
for the artist to "Artist" for differentiating the two roles.
Bug: T88835
Change-Id: Ie939572e7c1240321556f6f81ed9e67d792a339d
If data-mw.body.html is not set, search for the element
whose ID matches data-mw.body.id and use its contents.
.body.html is still used for new content: new references
use .body.html, and if a reference using .body.id is
changed, the changed HTML is output in .body.html.
Bug: T88650
Change-Id: I5d614927256d59a3ab20b0bd6e882b5cb8fcda20
Change the media search widget to work with resource queues and
providers. Create providers based on the user's filerepo settings
and aggregate their responses with the media queue. Stop asking
for more results from providers that are depleted.
Also fixes a rather nasty infinite-loop bug where the API returns
only very few images, and the UI keeps asking for more.
Bug: T78161
Bug: T88764
Change-Id: I65aed3446cd1f056476c56e6e04522c70e49e595
When the user changes an image to another in the media dialog, we
already have API info that includes the original dimensions, media
type and other information that is required for the Scalable object
to compute the new current dimensions. We can use this info if it
exists instead of asking for another API call. Also make sure to
update the filename when we change an image in the dialog.
Bug: T87267
Change-Id: I9d86959b99b3f3dfed92255a9dba9a8fdd352dcb
* Set the src attribute only when the image is actually
visible in the search results.
* Display the thumbnail image we have from the search
results and then update with a larger one from the API.
* Request for more media results on a higher threshhold,
when the user views 2 rows above the last available
result.
* Correct the resizeToBoundingBox and simplify it to work
properly for a non-square bounding box regardless of
constraints.
Change-Id: If024b0335ce6a5d2d0eafdbfdfe1030dcaac3a75
To avoid triggering state thrashing in Sizzle
( https://github.com/jquery/sizzle/issues/311 )
Also account for the fact that missing attributes are returned
as undefined by jQuery but null by .getAttribute().
Bug: T87416
Change-Id: Ib3bc7971920c084568abb6ea0bd822d763b276f2
Show the currently used filename at the top of the media dialog so
it can be copied and used in other pages if needed.
Bug: T54459
Change-Id: Ib1844c997189f7ead0656c5a811d3cab70030f89
Add an info panel to the MediaDialog to show the user information
about the chosen file from the new extmetadata API.
Also, change the way results are viewed:
* Present the search results with their proper ratio.
* Use masonry fit for the search results.
* Change the resizeToBoundingBox method in ve.dm.MWImageNode to accept
a specific side constraint, either 'width' or 'height'. This is also
going to be useful for adding and calculating 'scale' considerations
that affects the image height.
* Add info panel with details from extmetadata including author, license,
date of upload and creation, etc.
* Get descriptions and data according to wiki language.
Bug: T78161
Bug: T74061
Change-Id: I7b4d019825aa89dd178665bd9492a07af15954ac
Make sure all surface widgets have documents that are initialized
with the direction of the parent document, so the alignment is set up
correctly.
Bug: T71969
Change-Id: I6e5f003e18a9c5808d9a4b148eddf0d0f7e29e67
Aside from the confusion and differences between substr() and
substring() and IE8 bugs with substr(), substr() was removed from
the spec as of ECMAScript 5. It's been standardised in the
optional Annex B section of ES5.
Change-Id: I221ef6ae6956ce20dd9bb74510500f747d04c3b1
Deep-extending originalData is asking for trouble, because there's
no way to delete anything that way.
Bug: 73134
Change-Id: I44d3e582e04942efceae3e0bdbbaa147095b5146
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
0165a53 overlooked the fact that 'name' is also passed to
the MWTemplateModel constructor, and there it really does
always need to be a string, not magically either a string
or an mw.Title object. Oops :(
Bug: 72961
Change-Id: I0b20f0768aae4d9cc9f7af268abd0a704b6adc3a
Now that we have .getTitle() as a method in MWTitleInputWidget, replacing the callers
of MWTitleInputWidget.getValue() with .getTitle(). This fixes a bug in the .getTitle()
method in MWTitleInputWidget which was not taking the namespaces into account before.
Also fixes: the error when entering "Talk:" as a template title. The button to add template
remains disabled when "Talk:" is given as an input to the title here.
Bug: 71998
Change-Id: I1e629a61ec8b035d93a4b7acfecab81934019166
Parsoid has added a "key" property to parameter objects,
so now we have to preserve that rather than regenerating
all parameter objects.
Do this by changing the existing use of this.originalData
to a deep extend.
Change-Id: Ib4d31c2e0f7b567da1a148b51c42e688a2590b8e
Bonus: avoid modifying this.originalData
(Followup on If17b50cc4a39993)
Make sure initialHash is set before checking and changing its value
in the MWImageModel method. If a user chooses a brand new image to
insert into the document (an insertion rather than update) then the
initial hash is empty, and changing its values will fail with an
error message.
Bug: 72492
Change-Id: I3f9ea74891cc0ab77a07fa5d4e9cbc591f5b93e6
Methods for new "maps" property in
Template Data JSON root added to
the Template Specification Model
* Added maps property to MWTemplateSpecModel
* Maps filled in extend function
* Added getter function for maps
Depends on I3bf5e002
Change-Id: Ic555093625358d8fc4cb1ca8c68793b5912e8ced
On initialization, the image model has an initial scalable that has
the given currentDimensions from Parsoid; these are usually correct.
However, in cases where the wiki settings do not fit the user settings
and the images appear smaller or bigger in practice than the values
of the wiki-defaults, the initial hash will store the wrong values.
We will only know what the real values for the comparison will be
after we get them from the API and the calculation; only at that
point we can update the image model initial hash.
This is important so that later the dialog can properly understand
whether to enable the "apply" button if a user changed an image
to custom size and then back to default.
Change-Id: If17b50cc4a39993f98a20a3fec3ddf5d8cb400b3
Wikia has done some work on the template user experience, including
automatically showing all available parameters without the use of
TemplateData. In order to make our changes, we had to make some changes
to VE-MW.
ve.dm.MWTransclusionModel.js
* this.specCache is created so subclasses can reference it.
* Promise handlers in the fetch() method have been broken out as class
methods so they can be overridden in subclasses.
ve.ui.MWTemplateDialog.js
* addPromptedParameters() has been moved to the
initializeNewTemplateParameters() class method so subclasses can
overwrite. In Wikia's implementation, we have a method of getting
all parameters and a dialog that shows all of the parameters, so the
request to addPromptedParameters is overwritten.
* Added a done() handler to the transclusionModel promise for Wikia
extensibility.
Change-Id: I073c5850420e7719e82957f879423c2717af674a
Only template nodes for now. Not sure what we can do about generated content nodes in general...
Bug: 65353
Change-Id: I848f36764b446ed30c74c0e641d0973008f6880b
New changes:
6bbcd6a Localisation updates from https://translatewiki.net.
b8d8a5b [BREAKING CHANGE] The Great Selection Rewrite of 2014
Local changes:
Update to use new selection/range API
Change-Id: I5480d5c77d599c93c2d374fac88bb2fdb68b0024
This doesn't happen in real life (as resource is always set) but
it allows us to contruct these nodes from the factory with no data.
Change-Id: Iea146f5dc128ad9d0f36a79b6fe67d622767d46c
Allow for an auto-value for parameters in TemplateData, and make sure
this autovalue is inserted as text when parameter is loaded. This is
mostly useful for referencing date or other dynamic templates inside
parameters.
Bug: 52029
Change-Id: I6203530798ffaa62cb26190c70f7ce391bddf123
New changes:
a047e72 [BREAKING CHANGE] Create ve.dm.Focusable and store focusable flag there
Local changes:
* Add new ve.dm.FocusableNode mixin where required
Change-Id: Ie9ee2de05acfe3c4c62c4864e65bd20397d6a069
Create an image model hash and check for changes to the image every
time the dialog is changed, so we can activate and deactivate the
apply button properly.
Bug: 68058
Change-Id: I94b7e4879c6e752432c6f937a8cf1b9f15d1b56d
Make sure default dimensions are only set if we have the original
dimensions. Also fix a bug where image default dimensions were
calculated according to current width/height rather than those
of the original dimensions.
Bug: 68000
Change-Id: I8c3acc18b317f82c6b88966950e74b956936dda4
Make sure the image model can switch parameters when the image
source is switched, rather than having to instantiate a new
image model.
Also, when the image is changed, retain aspect ratio but make sure
to preserve the width of the previous image (unless the new image
is smaller)
Added 'resizeToBoundingBox' and 'scaleToThumbnailSize' static methods
in MWImageNode that return scaled dimensions based on the image
dimensions and the bounding box detals.
Depends on core change Idbb8b29db8e
Bug: 69450
Bug: 68000
Change-Id: I4327c86a9127f46da5522dc1197411fda757d5a4
So static methods are actally in the section marked 'Static methods',
before non-static methods, per our coding standards.
Change-Id: I17bea17a1409b02f263375c6434be872e4d239ba
This change creates the scalable object that's attached to the image
model through its constructor. Also adds a 'resourceName' variable
to store attributes.resource, which is mostly used for API calls.
Change-Id: I713c59d1cec5f87387a87c98a59e75b3d5c1050e
Previously one could select over a link node and linkify the
selection, with tremendously funny results.
Change-Id: Ic5a2132b2d18280099a6024bc4ec2b1333eca98f
Also cleanup redundant overrides, documentation and
unused messages.
Depends on If22a5197 in core.
Change-Id: I533235f4eb5d703783a8fb45dff5e7be465f4ebb
New changes:
76bbf3b [BREAKING CHANGE] Fix camel-casing of countNonInternalElements
Local changes:
Use renamed countNonInternalElements function
Change-Id: Ib5f86ea5cfeb5e577f61349ed4bbb231ea155638
We normalize underscores to spaces in sortkeys in toDataElement,
but when toDomElements recomputes the normalization to decide
if the sortkey really changed, it doesn't normalize underscores.
This is broken, both code paths need to use the same normalization;
we do this correctly for the category name.
This bug appears to date back to the introduction of category
support in March 2013 :(
Spotted in https://it.wikipedia.org/?diff=67220927
Change-Id: Ie3257fa335f3bd24a36c8654f3f5a24e89d7a9b9
The image model should go by image attributes (whether it comes from
an existing node or a new one) and not a node.
Bug: 68242
Bug: 68000
Change-Id: I4cabd678157619331b3c47d3b44bc16a28fd7bd7
* Each item builds an MWReferenceModel which creates a document
slice clone. This is very expensive and we only use the model
for getting basic attributes, so defer the evaluation of the
document.
* $.show is expensive and, in this case, unnecessary.
Change-Id: I99abc4c1b17f05559a9cae68b15121a8be6d23fb
New changes:
65042de Fix alien test data
83dccbe Add toolbar menu labels to each group for consistency
2e40ef9 Move 'language' tool from the insert to the styling menu, like in VE-MW
6fb21b2 Add toolbar action section to SA demo with keyboard shortcuts tool
f3eb000 ve.ui.Toolbar: Optimise onSurfaceViewKeyUp performance
eb4e9eb Make sure the document is focused when setting a selection
48584d3 Fixes to getRelativeOffset and getRelativeRange
fcc427f [BREAKING CHANGE] Remove generatedContents argument from afterRender
825e4be Documentation fixes in test utils
dbbbe0b Unit tests for showSelection
1ba9234 Don't handle up/down on focusableNodes even if backwards
a667b87 Remove label from 'insert' menu, and instead use a large plus icon
88885b1 Localisation updates from https://translatewiki.net.
1f3b79f Initialize data first
18369b1 Add 'clone' and make 'getDimensionsFromValue' static in ve.dm.Scalable
92e6b7b Switch to 'custom' when MediaSizeWidget updates
d0cd87e Localisation updates from https://translatewiki.net.
1ee21ff Remove lies from documentation
164f07f Safely ignore requests to open unrecognized windows
Local changes:
Work on a cloned scalable when editing images
Make sure the image model handles a cloned copy of the node's
scalable object, which is then shared with the media size widget.
Bug: 67687
Change-Id: Ie3676b715f5bae3d83da03d667cef3a6a847e212
Looks like I messed up in Ia67f9054 and got a setting backwards
Also prefix titles in a couple more places
Bug: 67384
Bug: 66893
Change-Id: Id6d5841ee1645be62a024308da48ddc6769c78b6
Merge the media edit and media insert dialogs. Allow editing
of media properties when inserting a new image, and also
allow the user to replace the image on edit.
Bug: 67170
Change-Id: I35bd95503565dbf95090ee8110284db5a8fbba83
Stop prefixing all requested titles with 'Template:', and don't try to use
colon prefixes when transcluding except when deliberately transcluding from the
main namespace.
(A title prefixed with a colon on MediaWiki is used to create a link to
something that would normally be included as a File/Category)
Bug: 52609
Change-Id: Ia67f9054a56e71153339f961bbbcadb565115748
Our coding style has been upstreamed and first released
in node-jscs v1.5.0 (grunt-jscs-checker v0.6.0).
Change-Id: Ifc0cfc7177946b9d11407f60e38b0177883724c5
* Return the fragment as it may have been cloned in the process
* Do away with DM node tree traversal, we can do everything
with offsets.
Bug: 65706
Bug: 65716
Change-Id: I9a44d562d9365d72adac199725b52555cf0a8009
Some surface fragment methods return a clone, so make sure
that is written back to this.fragment in dialogs, as this.fragment.select()
is called on teardown.
Functionally depends on If26cc0a2d in core.
Bug: 65706
Bug: 65716
Change-Id: Ia552b2a4c4c59ffc308a4acdecac78a7803a1c1f
Depends on I468d4eb4 in core.
Uses various hacks to trick the test runner into thinking
an MW target is in use, when in fact we still use SA targets.
Change-Id: If4611307d5d7aaee4af84f86ef82faf9078043b6
When converting from block to inline image inside a list, insert
the image into the existing paragraph instead of wrapping it in
a new paragraph.
Bug: 65924
Change-Id: Iefbb2a9ae536f51987bdd1aa8b7269bf6245d2b0
Some parts of this are heavily based on how the existing defaultsort stuff works.
Relies on I9bb80b01 and Iba2fff95
Bug: 56868
Change-Id: I5f0ce2ec36e82b04d9e1357b14046867c1dd3b60
Frame images are not resizable; disable the media size widget when
the frame type is selected.
Depends on the change in ve-core:
Id1010798d30e2928358acb5694c429b9ab68ad57
Bug: 66353
Change-Id: I8c8da8933c51ba709a834731425d19445009673f
New changes:
a53f101 Update OOjs UI to v0.1.0
1fa518b Update OOjs UI to v0.1.0-pre (15f4156bac)
257af53 Use ContextWidget instead of toolbar inside context menu
5f10e73 Follow-up Ia2076a42: Mark getDescription() as @inheritable for Annotation, Node
Local changes to adjust for the ContextWidget changes.
Change-Id: I47f68f5cf1b9583dab9bd0109fa6504481bdfc67
MWBlockExtensionNodes would duplicate their children in the output DOM:
they would appear once as restored from originalDomElements, and again
because the contents then got converted to linear model data and
that got converted back to DOM as well.
Bug: 66396
Change-Id: I4513ffc1cf3eb8726ac94252fe1526fac6d8bdf3
Make sure that image model and media edit dialog initialize the
default size properties according to the scalable correctly.
Also make sure that on creation of a new node or edit of existing
node's properties follow on whether the image is set to default
size or custom size.
Depends on the ve-core change:
Ia6bb041bfc050a48d8c40bbff4b73ed2893085e2
Bug: 66149
Change-Id: I6f60fba28e11a111bb40a100148ed1f12d518890
Previously they were always inline.
MWExtensionNode is now a mixin and classes which inherit it should
now inherit MWBlockExtensionNode or MWInlineExtensionNode.
Change-Id: I0ddcabeb63acdf852e252c41add5233c00723f4c
Update DM node constructors for I81c60d4a in VE core.
New changes:
2587549 [BREAKING CHANGE] Fix Node construction arguments
Change-Id: I8d4f37b631fe85a78c72407937477963ac49f87a
Fixing the behavior of alignment select and checkbox in the edit
dialog. Adding a check on all UI events to make sure they update
the model only if the model value is different than the UI value.
Also adding the ability to check the default direction of a
specific node type, to predict default alignment values.
Bug: 65916
Change-Id: I82f624fa788383dec0a12afb473aef01593e670e
Make sure that if the image has default dimensions set up in the
media edit dialog, the scalable (and image CE rendering) receives
the correct dimensions.
Bug: 65886
Change-Id: I6bd2d1681a3d780554767fe77b2640795170184f
Make sure svg drawings on block images have the default size as
the wiki default, no matter their original set size.
Also took the opportunity to clarify the synchronization method
for scalable and image type in ve.dm.MWImageNode and call it when
the image model is loaded to make sure the details are always
synchronized before edit.
Followup to I682d832e456a501836f33ed1dfc7dbd78a4edf53
Bug: 62893
Change-Id: Ic47016d171634c532092067957cc2c3d431e9746
Once more with feeling; default alignment when switching between
inline to block and vise versa. This time, the problem was mainly
that the checkbox for position cannot be updated without triggering
its update event. It was triggered, however, before setting up the
positioning, which caused the model to run positioning twice and
get things wrong.
Another smaller fix in this commit was done to make sure that
the image model remembers the previous alignment, so when image
type changes from inline to block and vise versa, and the alignment
was default, it can recalculate default value for the new node.
Bug: 65564
Change-Id: I6c16a7855ad7e3ce302fa3d042d0a402251b31f0
This time in the installment: Allow the model to predict the type
and default alignment based on parameters rather than just the
current status. This is important for initializing images, especially
inline images whose 'align' attribute is null or undefined, but must
be examined for default alignment.
Also changed:
* Make sure that model operations change each other, rather than
let the editor change 'unrelated' model parameters.
* Use selectItem rather than chooseItem to avoid emiting the 'choose'
event on edit dialog initialization.
Bug: 65564
Change-Id: I8c23fc53e7738d59565294d8b98fe7351bebbc67
In getFilename(), decode URL encoding and replace the search
for 'File:' with a more standard ./ stripping regex.
The ./ regex came from generateContents) in the CE class;
made that code use getFilename() instead.
Also apply decodeURIComponent() to the result; but protect against
exceptions because it's a useless piece of garbage.
Bug: 65611
Change-Id: I7e81a8dc26d1c636c6571d96f43a394a53490587
Make default alignment a computed value, and make sure the GUI listens
to 'choose' instead of 'select' events for SelectWidgets so to prevent
recursive calls and make sure the GUI is updated properly.
Also added a default alignment check on type change so the user can
visually have an indication that the new image will have a different
alignment. The user can then change the value manually.
Bug: 65564
Bug: 65596
Bug: 62852
Bug: 65617
Change-Id: I55e16775d18e0a4bf7f1fa73bbb3aa5e0aa51a79
Relies on If6fb5856 to do anything useful, but degrades gracefully to current
behaviour.
Bug: 50964
Change-Id: Ia35c2f433a1a604d7794f367fd10dbccac102b89
When inserting a new node, get the proper fragment range and ask
for the node from within the search loop. Also added a fallback
for the case that the node wasn't found, and a caption cannot
be added.
Change-Id: Ic19d9b1cd2739435137f633de09a13b25a11a0cb
* Fixing a typo in the getImageNodeType() method ('frame' was supposed
to be 'frameless').
* Adding a check for type 'mwBlockImage' before editing/adding captions
* Using the node's getType() method instead of the model's where possible.
* Adjusting default alignment, so that when we check if it is set, it gives
us the actual answer, which is always useful.
Bug: 65568
Change-Id: I1687d8c7430c723b52448a9e24641e2c30d85d6a
Creating an MWImageModel that handles all image edit operations.
If the new attributes mean a change of image types from inline
to block or vice versa, the model will handle the creation and
insertion of the new node.
Change-Id: Ibe71bc8bd74e4ba5a024ac722432ccf0b8f65e71
