feat: expose shouldUseDarkColorsForSystemIntegratedUI
Closes https://github.com/electron/electron/issues/46429.
Refs https://github.com/electron/electron/pull/19735.
This PR adds a new API `shouldUseDarkColorsForSystemIntegratedUI` to the
`nativeTheme` module. This API returns a boolean indicating whether the
system is using dark colors for system integrated UI elements. This is
useful for applications that want to adapt their UI to match the system
theme, especially for those that use system integrated UI elements like
the shell theme or taskbar appearance.
* refactor: make api::View::GetBounds() const
* refactor: make api::View::OnViewBoundsChanged() private
refactor: make api::View::OnViewIsDeleting() private
refactor: make api::View::OnChildViewRemoved() private
* refactor: make ChildPair private
* refactor: use AddChildView(std::unique_ptr<View>) in OpaqueFrameView::CreateButton()
Xref: https://issues.chromium.org/issues/40485510
* refactor: use AddChildView(std::unique_ptr<View>) in MenuBar::RebuildChildren()
* refactor: use AddChildView(std::unique_ptr<View>) for ClientFrameViewLinux labels
* refactor: use AddChildView(std::unique_ptr<View>) for ClientFrameViewLinux buttons
* refactor: use AddChildView(std::unique_ptr<View>) in AutofillPopupView
* refactor: use AddChildViewRaw() to flag the edge cases that we still need to fix
* chore: use west coast const for consistency
* perf: do not clone the map each time we call Archive::HeaderIntegrity()
* perf: use absl::flat_hash_map for the integrity cache
* perf: do not clone the JSON payload string
* perf: preallocate capacity for the integrity cache
* perf: use move variant of insert_or_assign()
* refactor: simplify integrity cache building
remove unnecessary std::optional<>
* refactor: use base::FindOrNull()
* refactor: remove unused #includes
* refactor: make variable types explicit
* fix: make res_size unsigned
* refactor: put GetIntegrityConfigCache() in an unnamed namespace
refator: put LoadIntegrityConfig() in an unnamed namespace
* fix: oops, missing rel_path_utf8 key
* fix: oops, fix Wunreachable-code-return
* refactor: instantiate navigation_entries on the stack instead of the heap
* refactor: reserve the full size of navigation_entries
* refactor: use emplace_back to simplify the code a little
* fix: destroy parent port backend when JS env exits
* fix: close parent port before destroying
---------
Co-authored-by: deepak1556 <hop2deep@gmail.com>
* chore: bump chromium in DEPS to 136.0.7076.0
* chore: bump chromium in DEPS to 136.0.7077.0
* 6368856: Migrate absl variant.h and utility.h in content (part 2/2) | 6368856
* 6356528: Clean up LegacyRenderWidgetHostHWND code | 6356528
* chore: export patches
* 6339113: [Viewport Segments] Add CDP commands to override Viewport Segments without overriding other device properties. | 6339113
* 6352169: [DevTools][MultiInstance] Support new tab in another window on Android | 6352169
* 6368856: Migrate absl variant.h and utility.h in content (part 2/2) | 6368856
* 6360858:Clickiness: Wire response from URLLoader to DB, add e2e tests| 6360858
* chore: bump chromium in DEPS to 136.0.7079.0
* chore: export patches
* chore: bump chromium in DEPS to 136.0.7081.0
* chore: export patches
* chore: bump chromium in DEPS to 136.0.7083.0
* 6361987: Remove double-declaration with gfx::NativeView and gfx::NativeWindow | 6361987
* chore: export patches
* chore: bump chromium in DEPS to 136.0.7087.0
* chore: export patches
* fix: include node patch for missing AtomicsWaitEvent
6385540
* build: add depot_tools python to path
* fix: cppgc init and unregistering v8 isolate
6333562
CppGc is now initialized earlier so Node can skip reinitializing it.
Additionally, gin::IsolateHandle was attempting to destruct an already destructed
v8::Isolate upon electron::JavaScriptEnvironment destruction. By removing the call
to NodePlatform::UnregisterIsolate, this fixes the crash on app shutdown.
* fix: unregister isolate after destruction
See code comment.
* chore: bump chromium in DEPS to 136.0.7095.0
* chore: sync patches
* fix: add script_parsing::ContentScriptType parameter
6298395
* fix: migrate content::BrowserAccessibilityState methods
64014376383275
* feat: enableHappyEyeballs option for host resolver
6332599
* fix: add new cookie exclusion reason
6343479
* fix: add new url loader method
6337340
* fix: add new cppgc header file for electron_node headers
6348644
* fix: disable CREL on Linux ARM64
https://chromium-review.googlesource.com/q/I3a62f02f564f07be63173b0773b4ecaffbe939b9
* fixup! fix: add new cppgc header file for electron_node headers 6348644
* chore: update corner smoothing patch
* fixup! chore: update corner smoothing patch
* chore: disable NAN weak tests
These two tests are incompatible with a V8 change that disallows running JS code from a weak finalizer callback.
Ref: 4733273
* test: fix task starvation in node test
A V8 change makes these contexts get collected in a task that is posted
and run asynchronously. The tests were synchronously GC'ing in an
infinite loop, preventing the task loop from running the task that would
GC these contexts.
This change should be upstreamed in some way.
Ref: 4733273
---------
Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com>
Co-authored-by: alice <alice@makenotion.com>
Co-authored-by: Samuel Maddock <smaddock@slack-corp.com>
Co-authored-by: John Kleinschmidt <jkleinsc@electronjs.org>
Co-authored-by: clavin <clavin@electronjs.org>
New Node.js module resolution system reads `package.json` from imported
modules by reading from the file natively in C++ without calling into
`fs.readFileSync`. The ASAR FS wrapper code had copied files out into a
temporary folder as a workaround, but it is inefficient and does not
cover all module resolution mechanisms in Node.js.
In this change we expose `overrideReadFileSync` method on the `modules`
binding in Node.js, and use this override to call into ASAR-supporting
`fs.readFileSync`.
* fix: gin_helper::Promise in GPUInfoManager must be destroyed before destroying Node/V8
* fix: use CleanedUpAtExit to control the lifetime of GPUInfoManager
* fix: ensure maximize is emitted when reduce motion is enabled on macOS
* Revert "build: reenable graphics effects for macOS tests (#46454)"
This reverts commit 682ea345ac.
* perf: use an absl::flat_hash_set for UsbChooserContext::ephemeral_devices_
* perf: use an absl::flat_hash_set for GlobalMenuBarRegistrarX11::live_windows_
* perf: use an absl::flat_hash_set for NativeWindowViews::forwarding_windows_
* perf: use an absl::flat_hash_set for OffScreenRenderWidgetHostView::guest_host_views_
perf: use an absl::flat_hash_set for OffScreenRenderWidgetHostView::proxy_views_
* perf: use an absl::flat_hash_set for NativeWindow::injected_frames_
* perf: use an absl::flat_hash_set for NativeWindow::background_throttling_sources_
perf: have ErrorThrower lazy-lookup the current isolate
ErrorThrower's default constructor is marked as "should rarely if ever
be used" because it's expensive to call.
Unfortunately, nearly every instance of ErrorThrower comes as an argument
in gin_helper's JS-->C++ function marshalling where a thrower is
default-constructed and then populated in gin_helper::GetNextArgument()
with an assignment operator to a temporary ErrorThrower constructed
with the gin::Arguments' isolate.
tldr: most of the time we use the slow constructor first, then throw
that work away unused by overwriting with a fast-constructed one.
This refactor avoids that cost by deferring the expensive work to
`ErrorThrower::isolate()`, where it happens only as a fallback iff
isolate_ hasn't been set.
* perf: avoid redundant call to popup_bounds_in_view()
* refactor: use a std::optional<> for paint_canvas local
* fix: fix leaked gfx::Canvas in AutofillPopupView::OnPaint()
* refactor: remove redundant get() call when testing smart pointer for nonempty
* refactor: remove unnecessary draw_canvas variable
* refactor: rename bitmap to offscreen_bitmap for symmetry
* refactor: avoid another redundant call to popup_bounds_in_view()
