The language and standard library
As others, e.g graydon2 and Steve Klabnik has pointed out, rust doesn't currently need much when it comes to new language features, and 2019 may be a good year to mature that is already there (i.e. catching up on documentation, examples, etc).
However, there are some more or less existing stuff that would be very nice to have stabilized to a release. The `!` (never) type and TryFrom trait may be the ones I look forward to the most. Then of course, there is async/await, that just about everyone seems to be eagerly awaiting …
Rust has a great eco system, with library crates for just about everything. But there is one thing I hope for: A more mature web framework.
A whole bunch of good web frameworks already exists; rocket, gotham, warp and tide are all showing promise, but warp is explicitly experimental, tide is not on stable yet and rocket probably won't be on stable in a long time.
I don't really like the way gotham stores extracted parameters in a state where the handler needs to extract them by name. I prefer the parameter handling used in warp (which is similar to what i simulate with macros around nickel in rphotos). The way it's done in warp tends to lead to rather long compile times, but I think that can be solved.
There is also actix-web, but that seems like a rather different beast that I haven't got my head around yet. I should probably try it out.
Lots of progress in this area is just awaiting async, and to some extent TryFrom, to land in stable.
I'm happily hacking at ructe and rsass and expect to continue with both those and my end-user projects rphotos and fanrs through 2019. If more people are interested in joining my efforts, especially for ructe, that would be awesome. But this is spare-time projects for me, and my main skills are as a programmer, not as a project lead, so I don't really know how to make that happen other than trying to respond timely and nicely when there is an issue report or a pull request on them. There is some community involvement, which is awesome. I just hope I don't drive those contributors off.
It would be nice if 2019 can be the year where I start coding Rust at work and not only for hobby projects. When any suitable project comes up, I'll try to take the chance.