- Reorder README to have implementation list after "learning tool" bullet. - This also moves tests/ and libs/ into impls. It would be preferrable to have these directories at the top level. However, this causes difficulties with the wasm implementations which need pre-open directories and have trouble with paths starting with "../../". So in lieu of that, symlink those directories to the top-level. - Move the run_argv_test.sh script into the tests directory for general hygiene. |
||
---|---|---|
.. | ||
core.asm | ||
Dockerfile | ||
env.asm | ||
exceptions.asm | ||
macros.mac | ||
Makefile | ||
printer.asm | ||
reader.asm | ||
README.md | ||
run | ||
step0_repl.asm | ||
step1_read_print.asm | ||
step2_eval.asm | ||
step3_env.asm | ||
step4_if_fn_do.asm | ||
step5_tco.asm | ||
step6_file.asm | ||
step7_quote.asm | ||
step8_macros.asm | ||
step9_try.asm | ||
stepA_mal.asm | ||
system.asm | ||
types.asm |
x86_64 NASM implementation
Notes and known issues:
-
No library dependencies, only Linux system calls
-
Simple readline implemented, just supporting backspace for editing
-
Reference counting used for memory management. No attempt is made to find circular references, so leaks are possible. In particular defining a function with def! creates a circular reference loop.
-
The exception/error handling just resets the stack and jumps to a handler, so does not release memory
-
Memory is allocated by two fixed-size allocators (
Cons
andArray
objects) which have limits specified in types.asm. If more memory is needed then this must currently be done at compile-time, but adding sys_brk calls could be done. -
The hash map implementation is just a list of key-value pairs. Moving symbols around in the core environment makes a significant difference (20-30%) to the performance test. A simple optimisation could be to move items when found to the start of the list so that frequently searched keys are nearer the front.
-
conj
function not yet implemented -
*env*
symbol evaluates to current Environment.