Any code asking for a declaration change preview will trigger unecessary
updates if the value is the same as the current value.
Because of this we were awaiting for more than one rule-view-changed
event in a couple of tests.
It looks like we can also get rid of `flushCount` workaround.
Shapes highlighter are now doing full updates instead of light preview+commit.
Differential Revision: https://phabricator.services.mozilla.com/D294905
The swatches are currently broken when a rule view updates while they
are already opened. This is hard to reproduce yet as the rule view
rarely updates and it often requires interacting with the page or the
browser window to trigger one.
But this is much easier to reproduce with bug 2032221 patches.
Differential Revision: https://phabricator.services.mozilla.com/D294904
This reverts commit d1e8e2dad9.
Revert "Bug 2032220 - [devtools] Re-attach swatches when the declaration value is updated. r=devtools-reviewers,nchevobbe"
This reverts commit 70bb651cb9.
Revert "Bug 2032220 - [devtools] Fix unregistering for filter swatches. r=devtools-reviewers,nchevobbe"
This reverts commit a054f4fc68.
Any code asking for a declaration change preview will trigger unecessary
updates if the value is the same as the current value.
Because of this we were awaiting for more than one rule-view-changed
event in a couple of tests.
It looks like we can also get rid of `flushCount` workaround.
Shapes highlighter are now doing full updates instead of light preview+commit.
Differential Revision: https://phabricator.services.mozilla.com/D294905
The swatches are currently broken when a rule view updates while they
are already opened. This is hard to reproduce yet as the rule view
rarely updates and it often requires interacting with the page or the
browser window to trigger one.
But this is much easier to reproduce with bug 2032221 patches.
Differential Revision: https://phabricator.services.mozilla.com/D294904
Bug 2033191 made sure we only call ownerGlobal on nodes.
As per the discussion in dev-platform, the front-end ~always wants
.documentGlobal. This removes ownerGlobal from the bindings. Following
patch will rename internal GetOwnerGlobal in gecko to GetRelevantGlobal
as per the HTML spec terminology.
Differential Revision: https://phabricator.services.mozilla.com/D297183
Bug 2033191 made sure we only call ownerGlobal on nodes.
As per the discussion in dev-platform, the front-end ~always wants
.documentGlobal. This removes ownerGlobal from the bindings. Following
patch will rename internal GetOwnerGlobal in gecko to GetRelevantGlobal
as per the HTML spec terminology.
Differential Revision: https://phabricator.services.mozilla.com/D297183
Bug 2033191 made sure we only call ownerGlobal on nodes.
As per the discussion in dev-platform, the front-end ~always wants
.documentGlobal. This removes ownerGlobal from the bindings. Following
patch will rename internal GetOwnerGlobal in gecko to GetRelevantGlobal
as per the HTML spec terminology.
Differential Revision: https://phabricator.services.mozilla.com/D297183
Bug 2033191 made sure we only call ownerGlobal on nodes.
As per the discussion in dev-platform, the front-end ~always wants
.documentGlobal. This removes ownerGlobal from the bindings. Following
patch will rename internal GetOwnerGlobal in gecko to GetRelevantGlobal
as per the HTML spec terminology.
Differential Revision: https://phabricator.services.mozilla.com/D297183
Also, only show the element highlighter when the "objectBox-node" is hovered,
like we do in the console.
Tests are updated to assert the new behavior.
Differential Revision: https://phabricator.services.mozilla.com/D295750
This patch make it so the tooltip we show when hovering a `@container` query condition
in the rule view will have the properties that are directly referenced in the condition.
So for example, if the query is `@container (width > 10px)` it'll have `width`,
if it's `@container style(--x)` it'll have `--x`, if it's `@container style(attr(data-x px))`,
it'll have `data-x` and so on.
We also indicate when referenced variables and attributes are not set.
Tests case are added in !!browser_rules_container-queries.js!! to cover the different
features we can have. We also now assert the condition text in `assertQueryContainerTooltip`
so it's easier to understand which condition the function is asserting.
Differential Revision: https://phabricator.services.mozilla.com/D294420
We now show the `container-name` property (when it's not `none`), in the tooltip
that is displayed on `@container` condition tooltip.
Along this change, we also show the tooltip when the condition does not match
any container element (e.g. there's no parent element with the specified `container-name`
value), with a label indicating that there's no container with such name.
From the server, that means that instead of returning a vague `unmatched` property,
we now pass a `hasContainer` property that we can use to include the "open in inspector"
icon in the condition.
The existing test for container queries is updated to assert those updates.
Differential Revision: https://phabricator.services.mozilla.com/D293026
container queries can now have multiple conditions, so we're computing those
on the server and sending it to the client, where we render them separately.
As each condition might refer to a different container, we need to update how
we were handling hovering the container query to show the container tooltip.
Before, we were underlining the `@container` part plus the eventual container
name; now, we underline the whole condition (eventual container name + eventual
container query).
This also impacts how we were triggering the box model highlighter for the container,
it is now displayed when the tooltip is visible, and the tooltip is actually
responsible to toggle it.
The tooltip code is also tweaked so we would always show the tooltip based on
the condition element, not any of its children (e.g. the jump button).
And since we were updating things here, we now add a tooltip-anchor class to
the element that is used as the anchor for the tooltip, so it can be styled.
In this case, we use it to have the same style as we put on hover on the condition
so things feels a bit more stable.
Finally, in order for us to get the proper container, we had to update
`CSSContainerRule.queryContainerFor` in the previous patch so it would know which
condition we should get the container for.
This patch now properly pass the condition index to the RDP method, and the server
functions pass it to `CSSContainerRule.queryContainerFor`.
Differential Revision: https://phabricator.services.mozilla.com/D291482
Three callers:
* fixed attribute is only used as fixed=true for treecolumns, so no
behavior change.
* noautofocus is already only used as =true or removed, so no behavior
change.
* incontentshell is only used with ="false", rename to
escapecontentshell and flip the meaning.
Remove the API and similar GetXUL* stuff that did the same.
Differential Revision: https://phabricator.services.mozilla.com/D286520
We don't add a test for it as the issue only impact the baseline-shift property,
for which the compat data will have a MDN url the next time we fetch the data in
the repository.
Differential Revision: https://phabricator.services.mozilla.com/D285995
Add specific class from the OutputParser for the attribute name and fallback
spans so they're easier to target in tests.
Refactor and augment browser_rules_attr.js so it covers showing the preview
tooltip.
Differential Revision: https://phabricator.services.mozilla.com/D282021
This reverts commit 6c5157b828.
Revert "Bug 2000644 - Introduce a superclass of color picker to prepare for multiple subclasses r=devtools-reviewers,nchevobbe,desktop-theme-reviewers,frontend-codestyle-reviewers,hjones"
This reverts commit 0604161977.
Revert "Bug 2004409 - Split window opener logic from DateTimePicker classes r=emilio"
This reverts commit dbe80fc5ce.
This reverts commit a0d863ae05.
Revert "Bug 2000644 - Introduce a superclass of color picker to prepare for multiple subclasses r=devtools-reviewers,nchevobbe,desktop-theme-reviewers,frontend-codestyle-reviewers,hjones"
This reverts commit 11f938e477.
Revert "Bug 2004409 - Split window opener logic from DateTimePicker classes r=emilio"
This reverts commit 16894a9e46.
Nicolas Chevobbe