4 ms·
I'm a complete outsider to Firefox development (a former housemate used to work for Mozilla on something other than Firefox, that's about it), but as someone wi
by geofft 11y ago
I'm a complete outsider to Firefox development (a former housemate used to work for Mozilla on something other than Firefox, that's about it), but as someone with strong opinions on browsers and security and not very strong opinions on extensions, it seems pretty reasonable to me. I'm really not comfortable with the extent to which XPCOM lets you load native-code libraries and call functions in them, and I'm not really sure I see a way that you can have something that works like current XPCOM but also uses strong sandboxing. Firefox is neither sandboxed nor multiprocess, unlike Chrome, and while this may not yet matter in the marketplace it's very much a "long-term strategic necessity" -- the architecture's technical debt will start biting them in terms of security profile.
Meanwhile, the Servo folks (according to Google results for 'servo xul') seem to want to actively avoid XUL for Servo and are vaguely hoping to build the entire browser UI in HTML. If they make that work, then yes, the time and energy spent fighting with XUL will be a sunk cost, but I suspect most extension authors are familiar with HTML.
To be fair, these are both decisions you make if your primary priority is building the best web browser possible. There is an argument that Firefox has already lost there to Chrome and that's not a battle worth fighting, and the priority should be building the best extensible web browser possible: it's something Firefox is already much better at, it's something that's harder for Chrome to be good at (because Chrome has already done this rearchitecting, being a much younger codebase), and it'll let Firefox compete on niche instead of competing head-to-head. And, clearly, Firefox is not a smoldering security disaster today, so maybe things will be fine in practice with a more moderate approach.
If you take that worldview, then yes, they should be asking people who primarily care about extensions (developers and users). But if you don't take that worldview, if you ask people who care about browsers in general, they'll be supportive of this rearchitecture.