3 ms·
The way add-ons are currently done significantly hampers the engine architecture, and is a good chunk of the reason why multi-process took so long. All this syn
by metajack 10y ago
The way add-ons are currently done significantly hampers the engine architecture, and is a good chunk of the reason why multi-process took so long. All this synchronous access to engine internals is also why the first advice diagnosing performance problems is to disable all add-ons. NPAPI plugins also have similar issues, which is why everyone is getting rid of those as well.
Servo's architecture will also not support the old Firefox style add-ons, and we've known that the architecture would be incompatible for them (and NPAPI plugins) since the early days.
We want to enable people to build new experiences, but we have to find new ways to achieve this. Perhaps the technology the Browser.html team and us are working on will solve this problem. If you have ideas and want to help, I encourage you to get involved[1][2].
1. https://github.com/servo/servo https://github.com/servo/servo
2. https://github.com/browserhtml/browserhtml https://github.com/browserhtml/browserhtml
- Endy 10y agoIf your choices will remove support for functions which have been essential to Firefox since v2, then you need to make a better choice.
- dman 10y agoWhy not find out the new ways to achieve this before pulling the plug on the existing stack? You are asking your users to make a leap of faith with you by moving to a browser that is less powerful in the short to mid term.