Charles Kerr
Charles Kerr
Very much a work in progress. Testers, don't waste your time yet. WIll eventually fix #3333 but there are several steps to go first.
Not sure if this is something to merge or just an experiment, so pushing up to see how much CI screams about it :smile_cat:
No functional changes; just a minor const-correctness change. Use case: client code calls `utp_write()` or `utp_writev()` from a function that receives the data in `const` form, but currently has to...
#### Description of Change Replace our calls to the deprecated API `DeprecatedLayoutImmediately()` with `InvalidateLayout()` instead. It looks like all of the occurrences were just to trigger a new draw &...
the small `test` object in preload should expose itself to the renderer process. The preload code will also need to check electron's version & state to see how to expose...
Backport of https://github.com/electron/electron/pull/42006 See that PR for details. Notes: None.
#### Description of Change Fix a performance wart in `ProtocolRegistry`. The current workflow is to call `ProtocolRegistry::handlers().at(scheme)` after a safeguard check to `ProtocolRegistry::IsRegistered(scheme)`. This workflow requires a redundant lookup for...
#### Description of Change We have a lot of classes that inherit from upstream observer classes to manage internal state, e.g. our `AutofillDriverFactory` inherits from `content::WebContentsObserver` so that it can...
#### Description of Change Make the ownership / nullability contract for `RootView` fields a little stricter: - [Use `raw_ref` instead of `raw_ptr`](https://chromium.googlesource.com/chromium/src/+/main/styleguide/c++/c++.md#non_owning-pointers-in-class-fields) for class fields that can never be `nullptr`...