The current roadmap for that is "never". :) There just isn't enough ROI for rewriting all old code.

Surely at one point the return for getting rid of the rest of the C++ code would be getting rid of the C++ build process?

I'd go further and say that the ROI is not only "insufficient", it is negligible. Unless your legacy code is written in a language/framework where available experts are disappearing (which is not the case here), there is practically zero value on rewriting code that works as intended.

Plenty of legacy code was reimplemented from C++ to Java (a memory-safe language, not unlike Rust) when the latter was released. The ROI is definitely there. But given the scale of that task, it makes sense to pick the lowest-hanging fruit first. AIUI, most of the effort wrt. "oxidation" is now around WebRender, specifically making it the default render backend.

I think you should change your question to "which -parts- of firefox will be redone in rust and is there a road map for that?"

It's been a while, but I don't believe that is a goal, so I don't believe there's a roadmap for it.

Servo [1] was the project to do that. Unfortunately they don't care (enough) about it anymore [2].

[1] https://servo.org/ [2] https://mjtsai.com/blog/2020/08/11/mozilla-layoffs/

Servo was a research project, AIUI it was never meant to completely replace Gecko.