non-webshit templating engines/preprocessors

so, let's say I have a source code with a lot of large text strings embeded. for licensing-related reasons, I'd like to separate those strings from the source file, but still have them embedded in the compiled output. now, the language I'm using is a bit retarded and assumes that the program code is in one file, and the only way around it that I know of is by making it a library - but it comes with its own pitfalls so I'd rather avoid it.

the natural choice here would be some kind of a template or macro language. the problem with those is that their're are either an unintuitive mess that will bite you in the ass (like m4 or scribble/text) or webshit-oriented crapware which builds DOM, escapes XSS and in general assumes that you want to generate HTML. unfortunately for me, my retarded language is not HTML. because I don't want to work around HTML-specific shit, I'd rather use something that is outout-agnostic.

my desired workflow would be to have a first step of a build process that takes in a bunch of templated source files, fills in the values from corresponding text files and outputs one big piece of shit file that I can compile further.

what would /prague/ do? should I just freeze my anus and use a C preprocessor?

I use ERB to generate C code every now and then. It is completely orthogonal to HTML, other than using quasi-xml syntax <% %> as delimiters and including a few functions for escaping HTML in an Util module.