sforman/Thun
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
349 Commits 2 Branches 0 Tags
cfab1cf3c8b60447ef56c377622bc8d10a21994d
Commit Graph
100 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
Simon Forman cfab1cf3c8 Minor edits. 2020-01-31 12:01:13 -08:00
Simon Forman a14a55defb Remove unused predicates appears_only_once/2 and reg_used_once/2. 2020-01-31 08:30:10 -08:00
Simon Forman 3d1c9bab95 Freeing registers before using them is simpler. 
https://todo.sr.ht/~sforman/thun-der/3
 2020-01-31 08:26:10 -08:00
Simon Forman 400cc7dcde Partial reduction for combinator rule works after all. 
It just looked weird to me and I didn't think it through.
Once I checked it I realized it was okay.
 2020-01-30 10:25:12 -08:00
Simon Forman 8d1797c1b6 Uncovered and fixed a subtle bug in free_reg//3. 
non_alloc/1 for words that don't change the ref count of registers and can be delegated to their semantic relation.
 2020-01-29 11:04:22 -08:00
Simon Forman 95df1fda8c Cons. I should research Lisp compilers... 
I think I should emit high-level code and reduce it to actual
machine code later under whatever model (cons cell heap, etc.)
 2020-01-29 09:48:08 -08:00
Simon Forman 26c92e9a48 Add assoc to kinda sorta track the values in the registers. 
But it doesn't update e.g. if you add two numbers, the value int(N) stays the same.

It could be modified to track the value as it computes?  But then why keep them in registers at all? Sometimes a value must arrive at runtime, eh?
 2020-01-28 18:04:51 -08:00
Simon Forman ca7c23b1ac swap, pop, and + 
compiling is tricky
 2020-01-28 17:16:17 -08:00
Simon Forman bf8cc63c70 dup, add_ref/3. 2020-01-28 12:44:19 -08:00
Simon Forman 712dd24e88 THread through a context to track registers. 
It seems to work to allocate and free registers with a kind of reference counting
by membership in an auxilliary list.
 2020-01-28 12:24:04 -08:00
Simon Forman 768c7f6599 Sort of compile '+'. 2020-01-28 11:01:28 -08:00
Simon Forman bc7b471daf Move immediate to register for int literal. 2020-01-28 10:46:38 -08:00
Simon Forman 6888c71b69 blep. 2020-01-28 10:37:13 -08:00
Simon Forman 5b6bd42ebe A start on machine code generation. 
Just the initial messing around...
 2020-01-28 10:21:37 -08:00
Simon Forman b323402c9b Fix a bug in step. 
You think it would be easy to find all the places where the type tags are needed.
 2020-01-27 12:56:26 -08:00
Simon Forman 6e8151a0b2 docs, formatter 2020-01-27 11:54:24 -08:00
Simon Forman 714aa59c43 Fix a bug in rest. 
It didn't tag the tail list as a list.

Also, spell out bool for true cases.
 2020-01-26 17:06:21 -08:00
Simon Forman ab0a378b56 Minor cleanup. 
I feel like I should keep the un-partially-reduced thun/4 but
you can still read it, and the reduced forms are more efficient.
(And not too much more wordy.)
 2020-01-26 16:50:49 -08:00
Simon Forman 76de590d27 Experiments with partial reduction are very promising. 
Functions become clauses like these:

    thun(symbol(rolldown), [], [C, A, B|D], [A, B, C|D]).
    thun(symbol(rolldown), [A|B], [E, C, D|F], G) :-
        thun(A, B, [C, D, E|F], G).
    thun(symbol(dupd), [], [A, B|C], [A, B, B|C]).
    thun(symbol(dupd), [A|B], [C, D|E], F) :-
        thun(A, B, [C, D, D|E], F).
    thun(symbol(over), [], [B, A|C], [A, B, A|C]).7
    thun(symbol(over), [A|B], [D, C|E], F) :-
        thun(A, B, [C, D, C|E], F).


Definitions become

    thun(symbol(of), A, D, E) :-
        append([symbol(swap), symbol(at)], A, [B|C]),
        thun(B, C, D, E).
    thun(symbol(pam), A, D, E) :-
        append([list([symbol(i)]), symbol(map)], A, [B|C]),
        thun(B, C, D, E).
    thun(symbol(popd), A, D, E) :-
        append([list([symbol(pop)]), symbol(dip)], A, [B|C]),
        thun(B, C, D, E).


These are tail-recursive and allow for better indexing so I would expect
them to be more efficient than the originals.

Notice the difference between the original thun/4 rule for definitions
and this other one that actually works.

    thun(symbol(Def),   E, Si, So) :- def(Def, Body), append(Body, E,    E o),    thun(Eo, Si, So).
    thun(symbol(Def),   E, Si, So) :- def(Def, Body), append(Body, E, [T|Eo]), thun(T, Eo, Si, So)


The latter reduces to:

    thun(symbol(A), C, F, G) :-
        def(A, B),
        append(B, C, [D|E]),
        thun(D, E, F, G).


We want something like...

    thun(symbol(B), [], A, D) :- def(B, [H|C]), thun(H, C, A, D).
    thun(symbol(A), [H|E0], Si, So) :-
        def(A, [DH|DE]),
        append(DE, [H|E0], E),
        thun(DH, E, Si, So).

But it's good enough.  The earlier version doesn't transform into
correct code:

    thun(symbol(B), D, A, A) :- def(B, C), append(C, D, []).
    thun(symbol(A), C, F, G) :- def(A, B), append(B, C, [D|E]), thun(D, E, F, G).

It would probably be good to investigate what goes wrong there.)

It doesn't seem to work right for thun/4 combinator rules either.  Dunno what's
up there.
 2020-01-26 13:21:47 -08:00
Simon Forman ef6d271c85 Partial reduction of thun/3 in the thun/4 relation. 
It mostly works.
 2020-01-26 12:48:38 -08:00
Simon Forman 0af0fb7e8e Map combinator works with types. 2020-01-26 10:15:32 -08:00
Simon Forman e0a36eab8b Remove '==' from definitions. (Bools) 
I decided that the conceptual simplicity of omitting '==' is more useful
than the cosmetic value of keeping it.  The defs.txt file is now just one
definition per line with the first symbol giving the name.

Also, bools are literals.
 2020-01-26 09:48:30 -08:00
Simon Forman ed41395560 Definition for 'not' in terms of 'branch'. 2020-01-26 08:44:57 -08:00
Simon Forman 2aa1765b89 Proper types, checking, inference. 
When I first translated Joy into Prolog I was so blown away by the Prolog
unification over list structures and Triska's CLP(FD) semantics for math
(both in the sense that he wrote CLP(FD) for SWI Prolog and he suggested
it for Joy-in-Prolog specifically) that I didn't realize that it wasn't
quite type inference.

It's kind of like type inference, in that lists are handled correctly and
the CLP(FD) library creates and maintains a kind of context for
constraints, which are sort-of like "dependent types" if you squint a
little.  But you can still do things like 'cons' a number to a number and
get (a Prolog term) like [2|3] which is almost certainly a bug.

So I went through and added type "tags" as Prolog terms: int/1, list/1,
and symbol/1.  The parser assigns them and all the primitive functions
and combinators use them (and so all the definitions do too.)  With this
information Prolog can e.g. prevent attempts to add numbers and lists, and
so on.

This also allows for the thun/3 relation to be implemented a little more
efficiently (without much loss of beauty) by looking ahead one term in
the pending expression and dispatching on structural "type" in a thun/4
relation.  I miss the elegance of the previous version, but this lets
Prolog index on the structural type tags that the parser produces.

(This might mess up tail recursion because now we have a loop between
thun/3 and thun/4 but we can eliminate that problem by partial reduction
on thun/3.  TODO.)

Now that literals are tagged by the parser there's no need for literal/1.
 2020-01-26 08:43:52 -08:00
Simon Forman 6e6e52d206 Don't assert defs twice. 
Each definition is getting parsed with the name of the next one as part
of its body, then the next one fails to parse and the thing backtracks.
So each definition (but the last) gets asserted twice.

    def(--,[1,-,?])
    def(--,[1,-])
    def(?,[dup,bool,++])
    def(?,[dup,bool])
    def(++,[1,+,anamorphism])
    def(++,[1,+])
    def(anamorphism,[[pop,[]],swap,[dip,swons],genrec,app1])
    def(anamorphism,[[pop,[]],swap,[dip,swons],genrec])
    def(app1,[grba,infrst,app2])
    def(app1,[grba,infrst])
    def(app2,[[grba,swap,grba,swap],dip,[infrst],cons,ii,app3])
    def(app2,[[grba,swap,grba,swap],dip,[infrst],cons,ii])

...and so on.
 2020-01-25 16:13:06 -08:00
Simon Forman 1ecb5be278 Change back to CLP(FD) semantics. 
Minor changes to parser to make it less logical but a little firmer.
(E.g. ints can't be backtracked into becoming symbols!)
 2020-01-25 15:50:50 -08:00
Simon Forman 36ec93e46b Docs and minor cleanup to the grammar. 2020-01-25 14:35:44 -08:00
Simon Forman 0588496ca5 Minor cleanup. 2019-12-03 08:41:42 -08:00
Simon Forman fdf0339e16 Make parser REs into module-level "constants". 2019-12-02 14:26:07 -08:00
Simon Forman 39a0e73fac Some helper scripts for windows. 2019-12-02 14:14:38 -08:00
Simon Forman 57446a1179 Minor docs update. 2019-12-02 14:13:50 -08:00
Simon Forman eb591d27e0 Debugging this sucks. 
Even with the RISC emu GUI.

Redesign vm?  Add more tooling?  Use MetaII?

Happy Thanksgiving!
 2019-11-28 07:58:42 -08:00
Simon Forman ff69046a4c Emit a sort of symbol table. 2019-11-13 21:15:33 -08:00
Simon Forman 4a1770b1cc 4 is already an offset 2019-11-13 15:07:27 -08:00
Simon Forman 93caa2ed66 Definitions. 2019-11-13 11:41:41 -08:00
Simon Forman 851f212fb2 "swap" word. 2019-11-12 16:09:13 -08:00
Simon Forman 4faebb4551 "new" word. 2019-11-12 15:47:33 -08:00
Simon Forman 4cba943355 Refactor sub_base_merge_and_store. 2019-11-12 11:42:24 -08:00
Simon Forman 89c58d4f5a dup cons i 2019-11-12 09:23:31 -08:00
Simon Forman e72fa2585f Forgot to "return" from i combinator. 2019-11-12 09:18:58 -08:00
Simon Forman 9cfb06c736 Even "nicer". 2019-11-12 09:06:22 -08:00
Simon Forman 54fe7cbab7 Charming. 2019-11-12 09:03:54 -08:00
Simon Forman f201cd6bb3 dexpr//1 2019-11-12 09:02:19 -08:00
Simon Forman afea54bf08 The i combinator. 2019-11-12 08:37:20 -08:00
Simon Forman 80d127788e head_addr 2019-11-11 07:51:58 -08:00
Simon Forman c3e44e2bf8 Simple push of empty list. 2019-11-11 07:45:22 -08:00
Simon Forman 3126dce307 Dup. 
ANd portray_clause to stablize output logical variable names.
 2019-11-10 14:35:57 -08:00
Simon Forman 0b210d7754 halt. 2019-11-10 10:58:47 -08:00
Simon Forman 2e301fa6c3 word works with negative numbers now. 
symbols moved to head of machine code.
 2019-11-10 10:53:55 -08:00
Simon Forman 4a39d82c75 merge_and_store, chain_link 2019-11-09 21:55:01 -08:00
Simon Forman cde066eb86 Minor refactor. 
It doesn't save space (but if I reuse it once it will.)
 2019-11-09 21:20:31 -08:00
Simon Forman 2b9034f6d5 Pass through label. 2019-11-09 20:19:15 -08:00
Simon Forman 8405d3f863 if_literal and lookup 2019-11-09 18:39:39 -08:00
Simon Forman 7a821a4eaf incr stack 
Really decr, but I'm abstracting.
 2019-11-09 18:28:01 -08:00
Simon Forman 67583716f2 Refactoring and cleanup. 2019-11-09 18:15:56 -08:00
Simon Forman 04b021846a load 2019-11-09 18:02:07 -08:00
Simon Forman 8ee9a05780 Bleah. 2019-11-09 15:20:06 -08:00
Simon Forman e4c17f8c24 Base address in unpack_pair. 2019-11-09 15:12:44 -08:00
Simon Forman bcadc95aa3 unpack_pair 2019-11-09 15:03:17 -08:00
Simon Forman a1ecea341a Minor refactor. 2019-11-09 14:02:18 -08:00
Simon Forman c211848d5c Refactoring, with oddball quoting "symbols". 2019-11-09 13:59:06 -08:00
Simon Forman a38d8153cd That works nicely, again. 2019-11-09 13:42:14 -08:00
Simon Forman b7659bdeaa Convert to ? DCG and it's macro-time! 2019-11-09 13:34:51 -08:00
Simon Forman 632161b6d1 I think that does it for cons. 
Offsets in pair records can be negative.
 2019-11-09 12:14:51 -08:00
Simon Forman ab0ff48c54 That's the mainloop converted to permit negative offsets. 2019-11-09 12:03:13 -08:00
Simon Forman 9ce9d967cf Minor cleanup, bug fixes. 2019-11-09 11:27:29 -08:00
Simon Forman 3c796310a5 Minor bugfix. 
asr not ror.
 2019-11-08 15:24:09 -08:00
Simon Forman 4c13868e7c Cons 2019-11-08 14:06:28 -08:00
Simon Forman d67420ae68 Just do it in asm. 2019-11-08 08:08:53 -08:00
Simon Forman 5172be7a0a hmm... 2019-11-07 20:21:26 -08:00
Simon Forman 3f843ed4cc WIth push2 finished that's the mainloop converted. 2019-11-07 17:25:04 -08:00
Simon Forman 65d55cf7b4 lookup 2019-11-07 16:28:28 -08:00
Simon Forman 7bab15c64a if_literal 2019-11-07 16:15:41 -08:00
Simon Forman 0ce64f2ec4 A start on converting the mainloop. 2019-11-07 15:53:07 -08:00
Simon Forman b924350c6d Take two on the compiler. 2019-11-07 14:57:41 -08:00
Simon Forman 53ef16bee4 Minor cleanup. 2019-11-07 07:55:01 -08:00
Simon Forman 3751107a09 Modify error reporting a lil; words word. 2019-11-07 07:54:04 -08:00
Simon Forman f58f3b9959 Definition of ii combinator. 
ii == [dip] dupdip i

         a [F] ii
      --------------
          F a F
 2019-11-07 07:53:21 -08:00
Simon Forman 99f9534a10 Call for_serial/2 2019-11-07 07:43:21 -08:00
Simon Forman 22b34aba05 Using partial deduction to inline literals, functions, and combinators. 2019-08-22 15:45:24 -07:00
Simon Forman 558f45bf47 minor cleanup 2019-08-20 22:00:27 -07:00
Simon Forman 6e646c012f small func recognizes [] and [X]. 2019-08-19 22:02:06 -07:00
Simon Forman b59e70f96d Rename bar to korf and some docs. 2019-08-18 16:59:31 -07:00
Simon Forman 26c2730273 minor cleanup 2019-08-18 12:17:37 -07:00
Simon Forman 7d44742317 reintroduce definition of fork; it doesn't shadow func(fork, ...). 2019-08-18 11:47:48 -07:00
Simon Forman a5fb17cc48 read child output after local thun/3 2019-08-18 11:09:36 -07:00
Simon Forman 2057d9ee74 truly fork, sort of 2019-08-18 10:53:25 -07:00
Simon Forman 11fdc81409 minor cleanup 2019-08-13 11:58:31 -07:00
Simon Forman 8116b83086 Minor cleanup. 2019-08-13 11:51:41 -07:00
Simon Forman bbc0f750bc Move DCG stuff to own file. 2019-08-13 11:48:08 -07:00
Simon Forman defedcbd12 Parse ints; move line/{1,2} to main.pl. 2019-08-13 11:29:49 -07:00
Simon Forman 3f19b2b210 Parse floating point numbers. 2019-08-13 11:24:28 -07:00
Simon Forman 4c78a31bd9 minor cleanup 2019-08-12 22:29:09 -07:00
Simon Forman d22e6e86dc minor cleanup 2019-08-12 21:59:19 -07:00
Simon Forman 86bf875841 Numbers can be followed by space or [. 2019-08-12 21:29:26 -07:00
Simon Forman 8947dff010 Parse negative numbers. 2019-08-12 21:13:11 -07:00
Simon Forman 143b711028 swapd function 2019-08-12 21:08:35 -07:00
Simon Forman aae0e91c44 assert_defs/1 got upset about not finding combo/5 2019-08-12 21:08:17 -07:00
Simon Forman e67ef3951c Minor cleanup of the parser. 2019-08-12 20:59:02 -07:00
Simon Forman 120610cd6b rework parser DCGs 2019-08-12 20:12:35 -07:00