If I can throw in a couple thoughts:
My experience with open-source projects has usually been that they are, largely, labors of love. Almost none of them manage to run as commercial enterprises unless they somehow managed to capture a new market segment at some point (like Red Hat Linux trying the "software is free, support costs money" model, which was relatively novel at the time). The rest, it seems, that have managed to support their costs seem to have done so by running as foundations: they exist as a nonprofit to promote whatever ethos their software supports (in my case, my flavor of Linux has a foundation that supports the web forum and pays some nominal fees to devs), and take in donations.
It's a hassle getting the initial foundation set up, with no small amount of paperwork. But it also makes it possible to apply for grant funding from larger public-interest charities (I'm thinking here specifically of research foundations that give out grants for specific research, but *only* to other registered nonprofits), and then paying out for dev work becomes more straightforward.
It's also possible to do what my foundation does, which is to register with the Boost project. It's a partnership site that works with major web retailers (like
amazon.com, but also the major chain stores, and [here, at least] the online ticket counter for the national railway). Once users have the app installed in their browser (I know, I know), the partner vendors will donate some single-digit percentage of their sales to the charity of the users' choice.
It'd be a way to send a few € a month to Lilypond development without having to spend anything extra, which -- while it won't support a dev fulltime -- can certainly help.
Cheers,
A