Note, since the page issues code is not feature flagged,
the new treatment will only be accessible via query
string until T206179 is taken care of.
Bug: T206178
Change-Id: I5ab2f3396e642f7b973263e2bb3963e0e82721b3
The extractMessage function has a lot to do with parsing - so this
and its tests are moved into the pageIssuesParser.
Change-Id: I62d79fbba166eff2c3ca573ef94ff86a269a7f9a
We have two type defs - IssueSummary and PageIssue.
I'd like to consolidate these two types by making
IssueSummary a combination of the two
Change-Id: Ic831b463fa66b0cacdd0b9b79aff741e55c0ec24
In I06ef42cb1461fde7ca0aa903f174c3b1f39ab154 a button became an
a tag and somehow that got through the Jenkins censors.
With this change Jenkins on Minerva will be happy again
Bug: T207480
Change-Id: I3faedb10e46c3e3e237f5b90d71ffe20606fce6d
Also make the parameter $tooltip mandatory to ensure that a mandatory
parameter does follow an optional parameter. ParserOutput always fill
all parameters of doEditSectionLink.
Bug: T207413
Depends-On: Ic49d76c7d9357f99b462471696f7e1eb86e70e3d
Change-Id: I49c1e3fb7bf72002c7f498d4afe14a06caf411f8
There's no reason to have this as separate style file in content.styles
as common text styles are taken care of in 'text.less' and
'print/articles.less' for print. Moving contents rule there.
Change-Id: Ie613d95488e9b5a814b6be8f0c856e9e92ab5aed
Separate the page issue grouping concern so that changes to parsing
don't concern everything else and vice-versa.
Bug: T203449, T202349
Change-Id: I7bddb0c53310805ece71b8f7821b1d6ce05cfae9
The share button should go though proper testing before it hits
production. First, it should be enabled only for beta cluster,
and then when everything works properly we can enable it as a
beta feature on production.
Bug: T181195
Change-Id: I1c5330edd8b5dfe00179bf7d9c6660887208a2e6
Add a new beta feature - share icon. When user opts into beta
the share icon will be visible on supported devices (Android OS
and Chrome browser).
Bug: T181195
Change-Id: Ie4b9dd05eea9b63422bd174048d8b1251cb02bf4
It's presumed that skin options will eventually become the default or be removed from the skin.
While they are not the default, it would be helpful to package them in one single module - as ResourceLoader
modules are costly bloating the dependency graph in the MediaWiki startup module.
In T167713 we talked about grouping our entry points by page rather than feature, which this seems consistent
with. A page with special options enabled is different from a page without.
Change-Id: Id948f913d4743532ba3442d2059a03c122419ff2
All of this would only be used with the configuration setting
`$wgMFEditorOptions['anonymousEditing'] = false;`.
Removed features:
* Call-to-action popup in skins.minerva.editor (note that anonymous
editors still get a CTA from MobileFrontend's EditorOverlay code)
* Pointer towards the edit button shown after registering via the CTA
(entire skins.minerva.newusers module)
Bug: T205382
Change-Id: I66c7035f7a23581811dda87c911dea41d4a8e5da
Dynamic Type is a feature in iOS that lets users pick their reading size
so that apps can adjust their text size accordingly. This commit makes
the mobile site hook into that feature. If you go to
settings > Accessibility > Larger Text in iOS and increase or decrease
the reading size, the mobile site will now increase / decrease
its text size relative to that setting.
Notable changes:
* Moved `font-size: 100% * @fontScalingFactor` from html to body. This
rule would override Dynamic Type sizing which would effectively disable
it. Moving it to the body makes the rule be relative to Dynamic Type
instead of overriding it in iOS browsers while browsers that don't
support Dynamic Type should observe no visible changes.
* Removed intermediate variable @bodyLineHeight because it was only
being used in one place.
* Added @font-size-body-mobile and made it based on % instead of em
(which is what @font-size-body uses). @font-size-mobile-bod is used by a
media query in browsers that support Dynamic Type and by another media
query in browsers that don't support Dynamic Type.
* Added a feature query for Dynamic Type which will make Dynamic Type CSS
rules only affect browsers that support Dynamic Type. This limits the
feature to iOS 9+ feature but ensures that the css rules don't affect
browsers that don't support Dynamic Type.
* Scaled down Dynamic Type so that its default size would appear
identical to the common browser default text size of 16px. By not doing
this, the default font size in iOS browsers would be 17px (Dynamic
Type's default) and would appear 1px larger than Android browsers.
Bug: T204807
Change-Id: I8a4d621dba8dc56190bd8c974543d08dd374ba5e
The project ESLint configuration unintentionally inherited unwanted
configuration from MediaWiki, which allows global jQuery usage. Cap the
project's ESLint hierarchy at the project root.
Change-Id: Iceda39a4eeb492446b952f597af0f153cf345206
Goal: Make skins.minerva.editor not rely on Minerva-specific markup.
SkinMinerva.php:
* Add `class="mw-editsection"` to section edit links in SkinMinerva.
This is the default behavior in SkinTemplate.
* Tweak the page "Edit" link generated in PHP to be the same as the
link we were generating in JS: add class="edit-page" and change the
message for the text.
* (Fix an unrelated code comment that was incorrect.)
skins.minerva.content.styles/hacks.less:
* Remove a hack that was hiding .mw-editsection, since we now use it.
skins.minerva.editor/init.js:
* Stop using the `data-section` attribute on links to decide which
page section to open in the editor. Instead, use the `href`
attribute and extract the `section` URL parameter from it.
* Stop using the `edit-page` class to find section edit links.
Instead, use the `mw-editsection` class.
* Remove super weird code that removed the original "Edit" link from
the page and generated an identical one to replace it, instead of
just adding event handlers to the existing one.
* Centralize event handling for all types of edit links.
Bug: T198765
Change-Id: I79639c738ff1c3ec4b48ee2e462d23060151a21b
These CSS classes are specific to the Minerva skin and we want to move
this file to MobileFrontend.
Now that this code is all in one place, I noticed that it seems to not
do much at all… Everything in this function looks like it could just
be done in the PHP code. If PHP does it all, then we could remove all
of it for a free performance improvement.
Bug: T198765
Change-Id: I6487c2fc520e14e0856b2e9a6f9dfa5066205817
