mes/HACKING

-*-mode:org-*-

* Fully source-based bootstrapping

** R6RS-like scheme interpreter
This first part is prototyped in C by the mes.c core and Scheme
bootstrap code in module/.  Of course, while mes.c is pretty small it
cannot serve as a fully source-based solution.

The initial idea was to have the minimal core support LISP-1.5 (or
something very close to that as a tribute to John McCarthy) and extend
eval/apply from LISP-1.5 source with define, define-macro etc. and
metamorphose into R6RS.  It seemed to work but performance of the
LISP-intepreted RRS was so bad (~1000x slower than initial LISP-1.5)
that this track was abandoned after the initial ANNOUNCE.

The route changed, trying to strike a balance between core size and
performance: still writing as much as possible in Scheme but having a
mescc compiler that takes not more than some seconds to run.

Now that portable R6RS syntax-case runs and mes.c has grown to
~1200LOC with another ~300LOC of optional C code, some effort must
probably be directed into making that smaller.

** Move mes.c into hex?
One idea is to use OrianJ's amazing self-hosting [[https://github.com/oriansj/stage0][stage0]] hex assembler
and minimal bootstrap binaries and rewrite the mes.c core to directly
bootstrap into Scheme.

** Rewrite mes.c and generate hex?
Another idea (thanks Rutger!) is to rewrite the mes.c core in a
C/Assembly variant and thave mescc produce the simple, annotated
bootstrap binary.

* Bugs
** Core is too fat
mes.c is ~1500 lines (~10,000LOC Assembly) which seems much too big to
start translating it to assembly/hex.

** (mes-use-module ...) is a fake, see module/mes/base.mes.
All top level scripts and test files (scripts/*.mes tests/*.test)
now include appropriate (mes-use-module ...) stanzas.

This hack allows for scripts/includes.mes to generate the list of
files to be prepended.  Previously, this information was put in
GNUmakefile.
** Garbage collection?
Mes is using malloc without freeing anything, memory is patient these
days :-) Sadly, a factor 10^6 less patient than the future that SICP
authors were hoping for (we have 10^3 less memory and 10^3 more
instructions).

SICP's stop and copy Garbage Colletor (Jam Scraper?) algorithm is now
available, but it cannot be hooked up yet as even in boot mode the
core mes eval/apply is still running; it executes the Scheme based
eval/apply.
** Actually do something useful, build: [[https://en.wikipedia.org/wiki/Tiny_C_Compiler][Tiny C Compiler]]
* OLD: Booting from LISP-1.5 into Mes

Mes started out experimenting with booting from a hex-coded minimal
LISP-1.5 (prototype in mes.c), into an almost-RRS Scheme.

When EOF is read, the LISP-1.5 machine calls loop2 from loop2.mes,
which reads the rest of stdin and takes over control.  The functions
readenv, eval and apply-env in mes.mes introduced define, define-macro
quasiquote and macro expansion.

While this works, it's amazingly slow.  We implemented a full reader
in mes.c, which makes running mes:apply-env mes:eval somewhat
bearable, still over 1000x slower than running mes.c.

Bootstrapping has been removed and mes.c implements enough of RRS to
run a macro-based define-syntax and syntax-rules.

loop.mes and mes.mes are unused and lagging behind.  Probably it's not
worth considering this route without a VM.  GNU Epsilon is taking the
more usual VM-route to provide multiple personas.  While that sounds
neat, Lisp/Scheme, bootstrapping and trusted binaries are probably not
in scope as there is no mention of such things; only ML is mentioned
while Guile is used for bootstrapping.

* Assorted ideas and info 
** C parser/compiler
*** [[https://savannah.gnu.org/projects/nyacc][nyacc]]
*** PEG: [[http://piumarta.com/software/peg/][parse C using PEG]]
*** [[https://en.wikipedia.org/wiki/Tiny_C_Compiler][Tiny C Compiler]]
*** [[http://www.t3x.org/subc/index.html][Sub C]]
*** [[https://groups.google.com/forum/#!topic/comp.lang.lisp/VPuX0VsjTTE][C intepreter in LISP/Scheme/Python]]

** C assembler/linker
*** [[http://www.tldp.org/HOWTO/Assembly-HOWTO/linux.html][Assembly HOWTO]]
*** System call clue bat
Basically, you issue an int 0x80, with the __NR_syscallname number
(from asm/unistd.h) in eax, and parameters (up to six) in ebx, ecx,
edx, esi, edi, ebp respectively.
*** ELF
7f 45 4c 46
*** [[http://www.muppetlabs.com/~breadbox/software/tiny/][Small ELF programs]]
*** [[http://www.cirosantilli.com/elf-hello-world/][Elf hello world]]

** RNRS
*** [[http://www.scheme-reports.org/][Scheme Reports]] 
*** [[ftp://publications.ai.mit.edu/ai-publications/pdf/AIM-349.pdf][Scheme - Report on Scheme]]
*** [[ftp://publications.ai.mit.edu/ai-publications/pdf/AIM-452.pdf][RRS - Revised Report on Scheme]]