Common Lisp

Lisp dates from 1958, which makes it the second-oldest language family still in use, after Fortran. For decades it was the language of artificial intelligence research. Common Lisp, standardized in the 1980s and by ANSI in 1994, merged the dialects of that era into one. Here it runs on SBCL, a compiler that turns each function into machine code.

Code and data share one shape: everything is a list, and (+ 1 2) calls + with 1 and 2. The standard gives you lists, vectors, hash tables, sort, reduce, a powerful loop, and format, but no statistics, no string split, and no way to read command-line arguments. So each program reads its data path from SBCL’s own sb-ext:*posix-argv*, and nothing from Quicklisp, the usual library installer, is installed.

Quirks

• The reader turns 28.6 into a single float, accurate to about 7 digits: as a double it’s 28.600000381469727. Every program here sets *read-default-float-format* to double-float, inside eval-when so that it also covers the decimals in the program itself, which the compiler reads before the program runs.
• (/ 7 2) is the exact fraction 7/2, not 3 or 3.5. Integers never overflow: (expt 2 100) is 1267650600228229401496703205376.
• sort may take its list apart. After (sort x #'<) on (5 4 3 2 1), x holds only (4 5); the sorted list is what sort returns.
• Names are case-insensitive: 'Rosa and 'rosa are the same symbol. Strings aren’t, and string= compares them.
• The run script compiles first and stops on any warning, so an unused variable or a discarded sort result fails the build.

git clone https://github.com/qnixon/dswithx.git
cd dswithx
python3 harness/run_tests.py --language common-lisp