We need to strip the protocol on both sides. This might have been broken
by Ic00b38b04ce78178c64c13bab7f1b2e4b6c5b803 in MediaWiki core.
Bug: T321437
Change-Id: I11903b767aebfdb189a8d54fbf6fb7f8ce9ffb6a
This patch follows the audit made on the extensions to check the usage
of the "rel" attribute and check that it's compatible with multi-values.
Bug: T315209
Change-Id: Ib323736d93ea96c86f9d56599e515c9e6d72a76e
This code was for when the dialog had a trash can icon for every
parameter, and parameters could actually be deleted. It's unreachable
now.
We missed this when removing the old workflow.
Change-Id: Ic94df506ea84009a1e1863a4e9847a70498df448
The need for something like this was anticipated in
I2bf43c7e83283f43e047229eb53c244918fcbb0c.
As of version 2.5.0 of Parsoid's output, if alternate text is missing
for an image but a caption is present and image isn't displaying the
caption (ie. it isn't a thumb or frame), then the text content of the
caption will be set as the alt attribute. Parsoid will then drop the
alt attribute when serializing if it matches the caption text, since
it's unnecessary.
However, if the caption is modified and the alt text isn't, the alt will
be serialized. This is likely to be unexpected to editor. They may
have missed that the both the caption and alt are populated in VE and
only edited one place.
Since all of the above is happening only for images where the caption
isn't visible, it doesn't appear to be a much used feature since, at
least for inline images, the experience of caption editing was already
less than optimal.
However, because of a quirk in how galleries are rendered in Parsoid,
this affects gallery caption editing, which is visible and presumably
used more often. See T268250 for a discussion on an improved gallery
structure. But for now, gallery images are effectively inline and set
the alternate text, thus subject to the above.
Here we add a checkbox so that the default is to ignore the alt if it's
the same as the caption. And only make use of it if it differed
originally or was explicitly unchecked to modify.
Bug: T311677
Change-Id: Idf297d8a98995971c5835b0cea56c3317a3626e2
It is no longer necessary to prepend a colon in Parsoid HTML
to ensure they are interpreted as links rather than an image
inclusion or categorization.
Instead, the colon causes Parsoid to generate piped links
when they could be unpiped, so remove it.
This code was added in 1e62e9f64c (2012),
the Parsoid bug was fixed in b62b93c678 (2013).
Bug: T312700
Change-Id: I3d71fd658b5dd627445e60b850f647081ef842e7
Only 'altText' should be described, everything else
is computed or, in the case of `resource`, results in
nodes being incomparable.
Change-Id: I586b67a0cfa30fae10a86fe3791f7e532c0ed754
This was done in 3 rather "random" places:
1. Whenever a template is manually added. But rather late, after the
template was added, in an event handler that is about focus
behavior. It should not continue to manipulate the template that
was just added.
2. When the dialog opens with a template preloaded by name, as it is
done from the citation menu.
3. When the dialog is about to finish loading.
This patch fixes 2 issues:
* Get rid of a duplicate call (number 2 and 3) when using the
citation menu.
* Move number 1 to a place where it's executed much earlier, and
only when the user clicks "add template" in a template placeholder.
There is no other way to add a template to an existing transclusion,
but it's still a more appropriate place I feel.
Bug: T311069
Change-Id: I8a65ad703b95ba2092e9ef73493e9903e96b0dd6
The "mediaClass" property now only serves to capture the original class
found on the media so that it can be roundtripped without causing dirty
diffs. In the 2.4.0 version of Parsoid's output, that will still be
the usual Image/Audio/Video. As of 2.5.0, it will always be File and
the mediaClass property can be dropped.
Parsoid is currently forward compatible with serializing mw:File, so
edited or new media can use that type already.
The contextmenu item for media has been updated to make use of the
"mediaTag" instead of mediaClass to continue distinguishing media types.
That was the only place a grep of mediaClass turned up any use.
Bug: T273505
Change-Id: If5dc6b794dacd6973d3b2093e6b385591b91d539
Before the (intentional) design decision was to not do anything special
when the same parameter is used multiple times (via aliases). Garbage
in, garbage out. Only the first usage of the parameter would work as
intended. The rest was ignored and subsequently removed from the
wikitext.
New design decision: Track and display duplicates as they appear in the
wikitext.
Notes:
* It's not possible to create such a situation in VE. Do this via
wikitext.
* Labels will be made distinguishable via T309198.
* Possible warning messages will be added later.
* The behavior when unchecking a duplicate will be specified later.
Bug: T309198
Bug: T310248
Change-Id: I6011344638cdad8529d8f57513ef51b5237eb878
Parsoid stopped emitting figure-inline in content version 2.2.0 and VE
requests 2.4.0, which isn't satisfied by the earlier versions.
Change-Id: I5f47fbe85fdca7fe429952709b62f5d1cc548daf
The previous comment was true, but we forgot another edge-case. The
same parameter can appear multiple times. The old code tried to
delete it 2 times, which deleted something else.
Note that the behavior for duplicates is larely unspecified. We
will work on this soon. This is only a first quick step to fix this
specific bug.
Bug: T309203
Change-Id: If0afb2c19626c3d9db0d109d6559ae74698ed378
* Implement ve.dm.MWGalleryImageNode.static.isDiffComparable to
match ve.dm.MWImageNode, in that images with different resources
are not compared.
* Diff galleries as documents so remove/inserts are rendered.
Bug: T308747
Change-Id: Ide6f4110e65cad7f6bb6d13766815413602fd991
New changes:
72bd71dbc Localisation updates from https://translatewiki.net.
b7f38d993 Re-apply "Fix for showing "invisible icon" when root node is hidden""
Local changes:
* Ensure templates are attached to DOM before rendering invisible
icon in converter hack.
Bug: T305110
Change-Id: I0e4c02619b2a06b9932f214ae482093fdb0cb22c
Such comments don't add any new information. The method signature
alone already tells the full story.
We did this already to a lot of the template related code we touched.
This is just a bit of cleanup to make it consistent.
Change-Id: I932b620910924a16dc0d31d6c8a3ab11818316fe
The numeric part of these ids is never used on it's own. There is no
need to expose it.
Note we renamed the method not long ago in I6eeab8b to reflect better
what it does. This is the next step. We just forgot it back then.
Change-Id: I5da82855e99ea3a42a5d91379c6974ae9c154518
New changes:
91a762803 ve.dm.ModelRegistry: Disallow unknown types on tag and tag+func matches
8575e8ad8 PreviewElement: Document event
d24f96874 Allow unknown types for ve.dm.AlienMetaItem
106354b38 Follow-up I534ce003d: Fix height restriction calculation of desktop context
Local changes:
* Allow unknown types for ve.dm.MWExternalLinkAnnotation, it is meant
to match all <a> tags and handles unknown types in toDataElement()
Depends-On: I785be75f389136769084a8e14832a09e3a519fc1
Bug: T303298
Bug: T304847
Change-Id: I75d1e82a43f5f96a9093cd334c524527918efaa5
In detail:
* Just compare strings where possible. Sometimes the first character
alone is distinct enough.
* Add missing ^ anchor to avoid possibly running unbound pattern
searches. This is not strictly necessary when we know the input is
fixed-length. Still I find it more robust and more obvious what it's
meant to do. Not to "search for some match somewhere" but to check if
the input _is_ a match.
Change-Id: I9b7651d8ec1266521c22b38ccabb95a947903f72
.test() is the dedicated syntax for a boolean "does match? yes/no?"
check. .match() returns an array of matches, or null. This is just not
needed in these situations.
Change-Id: Ibb996ab843d1a6c7d7af98d6a112990665d543b2
A class, mw-broken-media, was added in
I92359048b42d32fe8a0f2cb79cd348cf5f2c56cc
Bug: T304010
Bug: T270150
Change-Id: I542d7d879868c9030ce00b2686cdfd0077e49491
Parsoid added a class and, without it, we get selser complaining
about wrappers being modified, similar to T214649.
The "image" class is removed since Parsoid never added it (although it
now has "mw-file-description" for a similar purpose) and the legacy
parser doesn't apply it indiscriminately.
It doesn't seem like VE supports editing the |link= media option; it
just tries to roundtrip what's there and drops it on edit. The patch
here works with that limitation.
Galleries are found to drop href's, breaking selser, and should be fixed
in a follow up.
Bug: T292657
Bug: T303469
Change-Id: I92359048b42d32fe8a0f2cb79cd348cf5f2c56cc
This was added back in 2020 so mw:DisplaySpace would work while the
RESTBase cache still contained old data. It's been two years, it should
be good now.
Follow-up to ed3579245e
Change-Id: I65665231b78198633682d52c48068c89807df55e
Same random finds while working on something else. I carefully
checked and made sure these methods are actually called without the
optional parameter.
Change-Id: Iab36fd130258322985b5d6e7f8e1f7b4ee235ba2
These are only needed when we need to access a specific `this` from
within another `function () {}` context. This is not the case in the
situations here.
This is split from Ibf25d7e to make it smaller and easier to argue
about.
Change-Id: Ide1476de91fc343aa992ad92a1321d3a38b06dd0
It's not a getter, but a generator. I found the name confusing.
Getters typically don't return something different every time you
call them.
Change-Id: I6eeab8b6a8644e430003f6e1ad77ab4b28e0d8c9
* Make ve.ce.MWBlockImageNode autofocus=false, remove
unused transition property
* Remove ignoreChildren from ve.dm.MWBlockImageNode
based on new definition
* Remove tests which assert that deleting in a list next
to a block image always de-indents. If this is desired
behaviour it should be fixed without reference to
ignoreChildren.
Bug: T295905
Depends-On: Idc0cccbe73d1b49d07b60c14a192a40f47d64608
Change-Id: Ib79a070f5d36dbe7742fa0760f8cdf55fe3046ed
This is a more radical change, compared to the previous patch.
I will post more detailled explanations as comments on Gerrit.
Change-Id: I6909b3f0b2c153b7ee9995441e995ffa793eab40
Preserve the place of annotation meta tags; adds information for the
users about annotation and, if necessary, annotation range extension.
The messages and individual handling of annotations for the annotation
range can be defined by the extensions: see I0b58a418 for an example
of how that can look like.
The structure of this patch closely follows the one from I104e7abbd
(handling of <noinclude> et al.).
Bug: T261181
Change-Id: I39029e4a63d22b37107edec066006557bcff34bf
The signature of createInvisibleIcon was changed but this
was never updated.
This fixes invisible template renderings inside previews, e.g.
inside the reference context item for:
<ref>{{InvisibleTemplate}} Content</ref>
Change-Id: I3d1b7a177408032957ac3fa8ead813438aa6bda7
When I type e.g. "subst:example" as the template name, we made this
work as the user would expect: the template named "Example" is found
and it's TemplateData documentation used. But the dialog title shows
"Subst:example". Note the uppercase "S". It means this string is
parsed as a title, including the "subst:". This is confusing. Just
show the template name.
Change-Id: I9817786991a8379cf48b0a664aef1413abddee2d
Some cleanup to improve readability and reduce the amount of code.
Relevant bits:
* One method name was wrong. It can actually return parameter ids,
not only top-level part ids.
* I got rid of some fail-safe checks that are never needed or moved
them to a more central place.
Change-Id: I08f2ad7bc7d3f985d6189dff170dda554f3d37c2
I moved some code around and found that quite a lot of code wants to
know:
* Is the length of this transclusion exactly 1?
* I need that 1st part.
There is more that can potentially moved from the dialog to the model.
But I don't want to make this patch to big.
Bug: T292371
Change-Id: Ia94ed0450d04dd97c4c41f5bf7c266f9a534e821
The TemplateData extension is not limited to the Template: namespace.
And even if pages outside of the Template: namespace typically don't
have TemplateData information, the API is still able to automatically
extract parameter information from every page.
Or:
It's rare that a page outside of the Template: namespace is used as
a template. But if it is, this is not a mistake. The code here in
VisualEditor should not try to be "more clever" than the user is. If
this is what the user want's to do, let's not block them.
Bug: T291883
Change-Id: Iaf3fd5530b74fab7cedfc85ce04c8c40632df11f
The code that uses it is commented out
Bug: T291729
Follow-up: I7af2bc91524e832555b66f090a671672cd14f294
Change-Id: I4cceb9ca83a2274fa93783af3608b9486b773522
The previous patch Id314ee8 was incomplete. The event changed.
The id in the event is not guaranteed to be a top-level partId any
more, but can be a template parameter's id.
Note: "Parameter id" and "pageName" is the same. The fact that
these ids match is how the left and the right side of the dialog
communicate.
Bug: T289043
Bug: T291151
Change-Id: I391f0f8edb96398fd33a2e0b01003013c52776da
This patch improves the error handling for when a user tries to add
a parameter which is either an alias of a existing parameter, the
primary name of a existing aliased parameter, or a name/alias of an
existing parameter which is shown with an override label.
The error message was modified to always refer to the conflicting
parameter using the same name that is has in the sidebar.
Example: A parameter named "Parameter B" is already present in the
sidebar under its alias "B". When a user tries to add "Parameter B",
the new error message will inform the user that the parameter they
are trying to add already exists as "B".
Bug: T285869
Change-Id: I762b72b6cf14eb8ff5fcef63b4dcb70e297050de
I tried to review all of them. Some of the changes I did:
* Make sure the `config` parameter is not marked as optional
when it is not.
* Make sure default values are mentioned.
* List individual `@cfg` options when it makes sense.
Note I don't list all options a class could accept (e.g. via all
its parent classes and mixins). That's too much. Instead I checked
how a class is actually used and list only these options.
Even then I don't list everything, e.g. unspecific options
like "classes" that can be used pretty much everywhere.
Change-Id: Idf4fbe1dc3608ace277df9e385f2f140df3a2f50
We want to assert that value is true-ish, and that it doesn't equal a
default or auto string.
Bug: T290554
Change-Id: I454dda8d0085a8d3898a0d5b1a3ecc6dd7c2c9e4
The code in .cacheTemplateDataApiResponse() where the `specCache`
is filled skips missing pages. .setTemplateData() is never called.
While we could – in theory – check the `missing` flag (as done in
patchset 1), this flag never makes it to the spec.
Rather simple solution: Mark everything as undocumented, as long
as .setTemplateData() is not called.
This affects only missing pages. .setTemplateData() is called in
all other situations.
Bug: T272487
Bug: T276574
Bug: T290136
Change-Id: I7045e84f2f2ba5aa4591c94ea495b0249e6c40d6
Method names have been changed in I8fa47ed, assuming these are
private. It looks like some hacks exist out there. Let's make
these peoples life easier.
Change-Id: I63c80761fe06e2f3a4bb104fe3e8c17d1c7faa02
There are 2 situations:
1. Either the template name is used in a [[…]] link. In this case
we must provide the namespace. MWTemplateModel.getTitle() does
this. However, it's not a mw.Title object and therefor not really
guaranteed to be a valid title. This is fine. The worst thing
that can happen is that the link points to an error message.
But this should be entirely unreachable anyway.
2. Some messages want to display the name of the template.
Ideally without the namespace. That's what
MWTemplateSpecModel.getLabel() is for. Again this is not
guaranteed to be a valid mw.Title. But it doesn't need to. It's
only used as a label.
Change-Id: I03d0481201620a2f5c444ee32b656bcaade98aac
It appears like it's currently not possible to reach this code
with an invalid template name like `{{foo}}`. But this is not
guaranteed.
The purpose of this code is to call the TemplateData API. This is
pointless when a title is invalid. We know a page with this name
can't exist. So we skip it.
But that's all this code cares about. It should not crash. Nor
does it need to report this situation.
This is related to the discussion in Ic364e75.
Change-Id: If9bacc91b1c7bb110b33bfd395e1cbdf538e6c22
This is just the smallest possible boilerplate to get the first
trivial test running. More test cases will be added in the
following patches.
Bug: T289560
Change-Id: I3a4e49a7b9761db00b211e933386bad71d4f0942
Names like "fetch" or "resolve" are heavily ambiguous and
continue to confuse me. I hope these new names reflect better
what's going on.
Bug: T288827
Depends-On: I587a203a9370e4742f87586b4f1867b37459c375
Change-Id: I8fa47ed313e7d7b2c114a5638a67c4f3c8b830f1
This is what actually happens:
* We call `addParameter()`.
* This triggers an `add` event.
* This calls an `MWTemplateDialog.onAddParameter` event handler.
* This code doesn't check if a parameter already exists (because
it shouldn't). It detroys the page in the content pane on the
right and recreates it from scratch.
The only reason we do this is to focus the input field on the
right. This patch introduces a dedicated event to do this.
Bug: T288827
Change-Id: I47effe05427cfabfcf534920edee79521eaa033f
This check makes sure the user doesn't loose work when clicking
the back button. I would like to argue that neither of these
values is valuable enough to block the user with a confirmation
dialog:
* Literally nothing is lost when the input is empty.
* The auto-value is only temporarily lost, but will
automatically be restored when the user decides to add the
template back. The input field is pre-filled with this value.
* The default value doesn't need to be manually entered. It will
show up anyway when the parameter is missing.
There is a rare edge-case, but it is not really relevant in this
situation. Some templates allow to override a default value with
the empty string. This will be considered irrelevant by this
code. However, this was already happening before and doesn't
change with this patch.
The only edge-case where this patch makes a difference is if a
parameter is marked as required or suggested, is documented to
have a default value, _and_ the template allows to override it
with an empty string. But this combination is rather crazy, if
not bogus, and not worth considering here, I believe.
Bug: T274551
Change-Id: Ib176a82844335c3d4dd5b720d335ec28245e1637
This is really only about the methods name, but doesn't change
any behavior.
I realized we work with several different definitions of what
"empty" means. There are at least two significant definitions:
1. When a parameter's value is the empty string or identical
to the default value, the behavior of the template is the same.
It will use the default value just as if the user entered it.
The auto-value is a meaningful value in this scenario and can't
be considered equal to the empty string.
2. The context here is when the user presses the back button.
This will destroy all user input. But an auto-value is not user
input. It will appear again when the user realizes they made a
mistake. Nothing is lost.
Personally, I would not use the word "empty" to describe this
concept. Things like "containsUserProvidedValue",
"isCustomValue", "isMeaningfulValue", … come to mind. These are
all still a big vague. A "user provided" value can be identical
to the default or auto-value. "Custom" how? I went for
"containsValuableData" for now.
Bug: T274551
Change-Id: I2912a35556795c867a6b2396cbad291e947f0ed6
This method already exists in the ve.dm.MWTransclusionPartModel
base class where it does the exact same.
Bug: T274551
Change-Id: I19d5914ed9b4b435c83ea4d64019bc46ce1ce8fd
This reverts commit 0d4dee341b.
Reason for revert: This made it entirely impossible to add a
deprecated parameter, even if done intentionally. Needs more tests.
Bug: T274551
Change-Id: I7389bad0845cd1ce78f9d7ef71592cb1ce2a063e
