Re:Improving Firefox Responsiveness on macOS
by saagarjha on Monday October 10, 2022 @08:33PM
(#33158062)
Yes but also do note that WebKit is probably among the most NIH codebase on the platform.
[ Parent
| on HN ]
Re:Improving Firefox Responsiveness on macOS
by olliej on Monday October 10, 2022 @10:37PM
(#33158868)
Those reasons are largely that the things that are perf critical for a browser engine don't match anything that normal apps do. That's the same reason gecko also isn't using the system allocator, as we see here.
You really don't want the rest of the OS making the performance trade offs people expect from browser engines these days.
Re:Improving Firefox Responsiveness on macOS
by ben-schaaf on Tuesday October 11, 2022 @01:40AM
(#33159863)
Blink uses os_unfair_lock on macOS: https://source.chromium.org/chromium/chromium/src/+/main:bas.... As does WebKit: https://github.com/WebKit/WebKit/blob/4c09a48ea418ff15f65d6c....
Re:Improving Firefox Responsiveness on macOS
by olliej on Tuesday October 11, 2022 @04:18AM
(#33160697)
Yes. Blink and WebKit, and I suspect a bunch of other - correct - software uses os_unfair_lock. Because os_unfair_lock is public, documented, and stable API.
The function this article is talking about is not API, it is part of the OS internals, presumably used to implement the standard API functions.