Upgrade Electron to v1.6.x #12696

This ensures smooth scheduling interactions with autocomplete-plus overlays so they measure their dimensions at the right time.

This avoids forcing a reflows in some circumnstances.

The shouldUpdate method is just returning true for now. We probably need to find a new approach to representing line number decorations that's easier to diff, perhaps a sparse array?

See https://jsperf.com/let-compound-assignment

We pass the bound scheduleUpdate method as an event handler to a variety of subscription methods, some of which supply arguments.

Albeit (potentially) slower, this will allow to add a dynamic number of block decoration nodes before and after a given line.

Also, remove some barely used public APIs around cursor visibility that don't make much sense and are not ideal for performance. We don't want to subscribe to the visibility of each cursor.

When two or more decorations in the same tile were invalidated but the sum of their height didn't change, we were previously failing to recognize that the line numbers gutter needed to be re-rendered. With this commit, whenever a block decoration is visible and gets invalidated, we will force the line number gutter to always update.

Signed-off-by: Nathan Sobo <nathan@github.com>

This ensures the component's shouldUpdate method works correctly.

Due to the way element-resize-detector schedules the delivering of resize events, this will ensure that creating an editor while the clock is mocked won't prevent subsequent tests using the real clock from getting such events.

...and fix spec/workspace-element-spec.js

Signed-off-by: Nathan Sobo <nathan@github.com>

…ances Signed-off-by: Nathan Sobo <nathan@github.com>

Signed-off-by: Nathan Sobo <nathan@github.com>

That's the language we use throughout the implementation now and a more accurate name for the concept.

These methods require us to render off-screen lines in some circumstances in order to measure them, so this commit extends the rendering of the longest line to include arbitrary lines.

This preserves the expected behavior for positioning overlays, etc so that package tests keep passing.

I wish we didn't need this, but it's currently relied on by several packages including bookmarks.

This avoids a non-failure error message when resetting the environment in some specs.

Signed-off-by: Nathan Sobo <nathan@github.com>

Previously, when the gutter container was removed due to the editor becoming mini, the lack of keys caused the gutter to be updated with the recycled cursors vnode. But then we tried to remove the cursors vnode not realizing it had been moved and tore down all the references. We probably need to revisit whether it makes sense to recycle vnodes.

Rather than storing these values on the editor model.

Ideally, packages would resize and then wait for an update. But we set up an example of calling measureDimensions directly in find-and-replace so the easiest thing for now is just to keep this method around.

These methods are sometimes called by the model before the editor has been attached to the DOM.

These are now tested in text-editor-component-spec

...because packages like `.git-diff` are relying on this behavior to position their decorations. This didn't seem to degrade layout times, so it makes sense to just add it to keep package breakage to a minimum.

This commit introduces the concept of a scrollTopRow and scrollLeftColumn which is used to query and update the logical scroll position.

Unfortunately Chromium does not trigger a `compositionstart` before firing the text input event for the non accented character. Using `undo` to remove such character from the history is risky because it could be grouped with a previous change, thus making Atom undo too much. With this commit we simply keep the behavior master exhibits as of today. In the process of rewriting this code path, however, we fixed a bug that occurred when opening the accented character menu while holding another key, and improved test coverage as well by simulating the events the browser triggers.

This ensures that direct calls to updateSync from places like scroll handlers never take this optimization path.

@t9md

These decorations allow the class and style of a cursor associated with any marker to be customized. /cc @t9md

This method can be used to translate a pixel position to a screen position even if the line is not currently rendered on screen.

The only event-based APIs we kept are for listening to changes in block decoration markers.

Signed-off-by: Nathan Sobo <nathan@github.com>

This was a private method in the previous implementation that was used by some packages. Signed-off-by: Nathan Sobo <nathan@github.com>

Signed-off-by: Nathan Sobo <nathan@github.com>

There are still a few rootComponent references remaining in the LineNumberGutterComponent. These should be removed and instead we should consult this data when constructing the line numbers to render.

Signed-off-by: Nathan Sobo <nathan@github.com>

@as-cii

I was hoping to strictly contain the layouts of highlights an lines separately, since they are updated during different render phases. Unfortunately, strict containment requires both divs to be positioned absolutely. This in turn creates separate stacking contexts for lines and highlights, which makes it impossible to render highlights in front lines which themes sometimes need to do. For example, atom-material-syntax pushes bracket matcher highlights to the front so they are not obscured by the theme's solid black cursor line background. /cc @as-cii. You should examine my work here and make sure I'm not screwing something up with your line/block decoration update code.

Taking the initial measurement was setting the soft wrap column, which was triggering a display layer reset, which was scheduling an update. This update occurred at an unexpected time causing an exception.

@t9md

/cc @t9md

This was happening when a text decoration overlapped a row, but the next boundary was located exactly at the beginning of it.

Some packages are interacting with this method assuming this behavior, so this commit eliminates `screenPositionForPixelPositionSync` and instead just performs the DOM update in `screenPositionForPixelPosition` if it is needed.

This avoids content being shifted over due to rendering and measuring the gutter on element creation and then subsequently hiding it.

This was leaving a measurement request in the map that was getting picked up on the next frame. In some cases, the requested measurement row was not present, causing an exception.

This was causing problems in measurements because in that code path we assume that text nodes are never empty. This commit also adds a test verifying this invariant when a text decoration ending right after a text tag is added.

This fixes #13880 (comment) once again.

Signed-off-by: Nathan Sobo <nathan@github.com>

Rewrite editor rendering layout to use new browser features and virtual DOM

Previously, we were accidentally depending on the state of the display layer when forcing it to update its index. This caused us to not index enough content to cover the visibile area, which meant we weren't querying enough lines to fill the screen in some situations.

Signed-off-by: Nathan Sobo <nathan@github.com>

See 1c8847c

This avoids updating tiles unnecessarily when changing the number of rendered tiles. Signed-off-by: Nathan Sobo <nathan@github.com>

Signed-off-by: Nathan Sobo <nathan@github.com>

This means we may render more tiles than necessary when we have block decorations, but it prevents changing the number of rendered tiles during scrolling with certain combinations of line height and editor height. If it ever becomes a problem we can get smarter about subtracting the height of the visible block decorations from the editor height, but for now this gives us more reliable performance for the common case.

See 4eecf8d.

Previously the DisplayLayer would have issues when closing tags that had not been previously opened, requiring us to guard against mismatched scopes at the frontier of asynchronous tokenization (see 5cfe971 for more information). Now the DisplayLayer gracefully handles closing tags that had not been opened, meaning we can eliminate this costly logic. Signed-off-by: Nathan Sobo <nathan@github.com>

Signed-off-by: Nathan Sobo <nathan@github.com>

…enRow

# Conflicts: # script/package.json # src/task.coffee

This avoid an extra screen-to-buffer coordinate translation. Signed-off-by: Nathan Sobo <nathan@github.com>

Signed-off-by: Nathan Sobo <nathan@github.com>

Signed-off-by: Max Brunsfeld <maxbrunsfeld@gmail.com>

It's causing test failures locally and enough time has passed that most user data should be transitioned to indexed DB by now.

We now pass an isolated document into the AtomEnvironment instance under test to avoid accidentally handling clicks on the test runner itself.

Adding a source map for the entire snapshot was expensive in terms of memory and seemed to be triggering some sort of bug in Chromium when reloading with the DevTools open. The custom row translation relies on a much more compact representation of the data and avoids the crash. Signed-off-by: Nathan Sobo <nathan@github.com>

Signed-off-by: Nathan Sobo <nathan@github.com>

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Upgrade Electron to v1.6.x #12696

Upgrade Electron to v1.6.x #12696

Commits on May 5, 2017

Commits on May 8, 2017

Commits on May 9, 2017

Commits on May 10, 2017

Commits on May 11, 2017

Commits on May 12, 2017

Commits on May 16, 2017

Commits on May 19, 2017