These may not "last a lifetime" in some sense. But early computing books, especially by well-known authors, may come close.
Algorithms, unix networking, assembly books, the art of computer programming, Structure and Interpretation of Computer Programs, anything by Benjamin Pierce, good compiler books, discrete mathematics, the art of electronics, Princeton Companion to Mathematics, Spivak's Calculus, ...
A few months ago, I picked up my significantly dog-eared copy of TUPE and leafed through it to a section I'd remembered being about 'hoc', the higher-order calculator language they derive from first principles. I'm currently writing a 6502 compiler and I was implementing the assembler stage. I wanted my assembler to have macros and be able to calculate expressions etc., because I'm planning on having an asm [...] block in the high-level code. Being able to call the macro "_xfer32 r1,r2" is a lot more appealing than a long series of lda/sta...
So I sat and read about hoc, and worked my way through it, eventually coming up with something I could put into my code and call for any expression calculation. The compiler is a full-blown AST-based model but assemblers are much simpler beasts, and accepting of more limitations. Hoc (or rather an adaption of it) was an ideal basis for what I wanted.
Some old books have timeless wisdom within.
For archival purposes you can go see if it's in the Library of Congress or scanned into Internet Archive, and if not, that's a project you can opt to take on(but it's pretty high effort). Quickly checking for resale value as other comments suggest is also not hard, just expect that most of your stash will go straight to the bin, and getting more out of it means making it your job.