forth: JIT cooperation hooks (vm-eligible flag + call-count + forth-hot-words)

lib/forth/ans-tests/README.md

@@ -0,0 +1,14 @@
+ANS Forth conformance tests — vendored from
+https://github.com/gerryjackson/forth2012-test-suite (master, commit-locked
+on first fetch: 2026-04-24).
+
+Files in this directory are pristine copies of upstream — do not edit them.
+They are consumed by the conformance runner in `lib/forth/conformance.sh`.
+
+- `tester.fr` — John Hayes' test harness (`T{ ... -> ... }T`). (C) 1995
+  Johns Hopkins APL, distributable under its notice.
+- `core.fr` — Core word set tests (Hayes, ~1000 lines).
+- `coreexttest.fth` — Core Extension tests (Gerry Jackson).
+
+Only `core.fr` is expected to run green end-to-end for Phase 3; the others
+stay parked until later phases.

lib/forth/ans-tests/coreexttest.fth

@@ -0,0 +1,775 @@
+\ To test the ANS Forth Core Extension word set
+
+\ This program was written by Gerry Jackson in 2006, with contributions from
+\ others where indicated, and is in the public domain - it can be distributed
+\ and/or modified in any way but please retain this notice.
+
+\ This program is distributed in the hope that it will be useful,
+\ but WITHOUT ANY WARRANTY; without even the implied warranty of
+\ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
+
+\ The tests are not claimed to be comprehensive or correct 
+
+\ ------------------------------------------------------------------------------
+\ Version 0.15 1 August 2025 Added two tests to VALUE
+\         0.14 21 July 2022 Updated first line of BUFFER: test as recommended
+\              in issue 32
+\         0.13 28 October 2015
+\              Replace <FALSE> and <TRUE> with FALSE and TRUE to avoid
+\              dependence on Core tests
+\              Moved SAVE-INPUT and RESTORE-INPUT tests in a file to filetest.fth
+\              Use of 2VARIABLE (from optional wordset) replaced with CREATE.
+\              Minor lower to upper case conversions.
+\              Calls to COMPARE replaced by S= (in utilities.fth) to avoid use
+\              of a word from an optional word set.
+\              UNUSED tests revised as UNUSED UNUSED = may return FALSE when an
+\              implementation has the data stack sharing unused dataspace.
+\              Double number input dependency removed from the HOLDS tests.
+\              Minor case sensitivities removed in definition names.
+\         0.11 25 April 2015
+\              Added tests for PARSE-NAME HOLDS BUFFER:
+\              S\" tests added
+\              DEFER IS ACTION-OF DEFER! DEFER@ tests added
+\              Empty CASE statement test added
+\              [COMPILE] tests removed because it is obsolescent in Forth 2012
+\         0.10 1 August 2014
+\             Added tests contributed by James Bowman for:
+\                <> U> 0<> 0> NIP TUCK ROLL PICK 2>R 2R@ 2R>
+\                HEX WITHIN UNUSED AGAIN MARKER
+\             Added tests for:
+\                .R U.R ERASE PAD REFILL SOURCE-ID 
+\             Removed ABORT from NeverExecuted to enable Win32
+\             to continue after failure of RESTORE-INPUT.
+\             Removed max-intx which is no longer used.
+\         0.7 6 June 2012 Extra CASE test added
+\         0.6 1 April 2012 Tests placed in the public domain.
+\             SAVE-INPUT & RESTORE-INPUT tests, position
+\             of T{ moved so that tests work with ttester.fs
+\             CONVERT test deleted - obsolete word removed from Forth 200X
+\             IMMEDIATE VALUEs tested
+\             RECURSE with :NONAME tested
+\             PARSE and .( tested
+\             Parsing behaviour of C" added
+\         0.5 14 September 2011 Removed the double [ELSE] from the
+\             initial SAVE-INPUT & RESTORE-INPUT test
+\         0.4 30 November 2009  max-int replaced with max-intx to
+\             avoid redefinition warnings.
+\         0.3  6 March 2009 { and } replaced with T{ and }T
+\                           CONVERT test now independent of cell size
+\         0.2  20 April 2007 ANS Forth words changed to upper case
+\                            Tests qd3 to qd6 by Reinhold Straub
+\         0.1  Oct 2006 First version released
+\ -----------------------------------------------------------------------------
+\ The tests are based on John Hayes test program for the core word set
+
+\ Words tested in this file are:
+\     .( .R 0<> 0> 2>R 2R> 2R@ :NONAME <> ?DO AGAIN C" CASE COMPILE, ENDCASE
+\     ENDOF ERASE FALSE HEX MARKER NIP OF PAD PARSE PICK REFILL
+\     RESTORE-INPUT ROLL SAVE-INPUT SOURCE-ID TO TRUE TUCK U.R U> UNUSED
+\     VALUE WITHIN [COMPILE]
+
+\ Words not tested or partially tested:
+\     \ because it has been extensively used already and is, hence, unnecessary
+\     REFILL and SOURCE-ID from the user input device which are not possible
+\     when testing from a file such as this one
+\     UNUSED (partially tested) as the value returned is system dependent
+\     Obsolescent words #TIB CONVERT EXPECT QUERY SPAN TIB as they have been
+\     removed from the Forth 2012 standard
+
+\ Results from words that output to the user output device have to visually
+\ checked for correctness. These are .R U.R .(
+
+\ -----------------------------------------------------------------------------
+\ Assumptions & dependencies:
+\     - tester.fr (or ttester.fs), errorreport.fth and utilities.fth have been
+\       included prior to this file
+\     - the Core word set available
+\ -----------------------------------------------------------------------------
+TESTING Core Extension words
+
+DECIMAL
+
+TESTING TRUE FALSE
+
+T{ TRUE  -> 0 INVERT }T
+T{ FALSE -> 0 }T
+
+\ -----------------------------------------------------------------------------
+TESTING <> U>   (contributed by James Bowman)
+
+T{ 0 0 <> -> FALSE }T
+T{ 1 1 <> -> FALSE }T
+T{ -1 -1 <> -> FALSE }T
+T{ 1 0 <> -> TRUE }T
+T{ -1 0 <> -> TRUE }T
+T{ 0 1 <> -> TRUE }T
+T{ 0 -1 <> -> TRUE }T
+
+T{ 0 1 U> -> FALSE }T
+T{ 1 2 U> -> FALSE }T
+T{ 0 MID-UINT U> -> FALSE }T
+T{ 0 MAX-UINT U> -> FALSE }T
+T{ MID-UINT MAX-UINT U> -> FALSE }T
+T{ 0 0 U> -> FALSE }T
+T{ 1 1 U> -> FALSE }T
+T{ 1 0 U> -> TRUE }T
+T{ 2 1 U> -> TRUE }T
+T{ MID-UINT 0 U> -> TRUE }T
+T{ MAX-UINT 0 U> -> TRUE }T
+T{ MAX-UINT MID-UINT U> -> TRUE }T
+
+\ -----------------------------------------------------------------------------
+TESTING 0<> 0>   (contributed by James Bowman)
+
+T{ 0 0<> -> FALSE }T
+T{ 1 0<> -> TRUE }T
+T{ 2 0<> -> TRUE }T
+T{ -1 0<> -> TRUE }T
+T{ MAX-UINT 0<> -> TRUE }T
+T{ MIN-INT 0<> -> TRUE }T
+T{ MAX-INT 0<> -> TRUE }T
+
+T{ 0 0> -> FALSE }T
+T{ -1 0> -> FALSE }T
+T{ MIN-INT 0> -> FALSE }T
+T{ 1 0> -> TRUE }T
+T{ MAX-INT 0> -> TRUE }T
+
+\ -----------------------------------------------------------------------------
+TESTING NIP TUCK ROLL PICK   (contributed by James Bowman)
+
+T{ 1 2 NIP -> 2 }T
+T{ 1 2 3 NIP -> 1 3 }T
+
+T{ 1 2 TUCK -> 2 1 2 }T
+T{ 1 2 3 TUCK -> 1 3 2 3 }T
+
+T{ : RO5 100 200 300 400 500 ; -> }T
+T{ RO5 3 ROLL -> 100 300 400 500 200 }T
+T{ RO5 2 ROLL -> RO5 ROT }T
+T{ RO5 1 ROLL -> RO5 SWAP }T
+T{ RO5 0 ROLL -> RO5 }T
+
+T{ RO5 2 PICK -> 100 200 300 400 500 300 }T
+T{ RO5 1 PICK -> RO5 OVER }T
+T{ RO5 0 PICK -> RO5 DUP }T
+
+\ -----------------------------------------------------------------------------
+TESTING 2>R 2R@ 2R>   (contributed by James Bowman)
+
+T{ : RR0 2>R 100 R> R> ; -> }T
+T{ 300 400 RR0 -> 100 400 300 }T
+T{ 200 300 400 RR0 -> 200 100 400 300 }T
+
+T{ : RR1 2>R 100 2R@ R> R> ; -> }T
+T{ 300 400 RR1 -> 100 300 400 400 300 }T
+T{ 200 300 400 RR1 -> 200 100 300 400 400 300 }T
+
+T{ : RR2 2>R 100 2R> ; -> }T
+T{ 300 400 RR2 -> 100 300 400 }T
+T{ 200 300 400 RR2 -> 200 100 300 400 }T
+
+\ -----------------------------------------------------------------------------
+TESTING HEX   (contributed by James Bowman)
+
+T{ BASE @ HEX BASE @ DECIMAL BASE @ - SWAP BASE ! -> 6 }T
+
+\ -----------------------------------------------------------------------------
+TESTING WITHIN   (contributed by James Bowman)
+
+T{ 0 0 0 WITHIN -> FALSE }T
+T{ 0 0 MID-UINT WITHIN -> TRUE }T
+T{ 0 0 MID-UINT+1 WITHIN -> TRUE }T
+T{ 0 0 MAX-UINT WITHIN -> TRUE }T
+T{ 0 MID-UINT 0 WITHIN -> FALSE }T
+T{ 0 MID-UINT MID-UINT WITHIN -> FALSE }T
+T{ 0 MID-UINT MID-UINT+1 WITHIN -> FALSE }T
+T{ 0 MID-UINT MAX-UINT WITHIN -> FALSE }T
+T{ 0 MID-UINT+1 0 WITHIN -> FALSE }T
+T{ 0 MID-UINT+1 MID-UINT WITHIN -> TRUE }T
+T{ 0 MID-UINT+1 MID-UINT+1 WITHIN -> FALSE }T
+T{ 0 MID-UINT+1 MAX-UINT WITHIN -> FALSE }T
+T{ 0 MAX-UINT 0 WITHIN -> FALSE }T
+T{ 0 MAX-UINT MID-UINT WITHIN -> TRUE }T
+T{ 0 MAX-UINT MID-UINT+1 WITHIN -> TRUE }T
+T{ 0 MAX-UINT MAX-UINT WITHIN -> FALSE }T
+T{ MID-UINT 0 0 WITHIN -> FALSE }T
+T{ MID-UINT 0 MID-UINT WITHIN -> FALSE }T
+T{ MID-UINT 0 MID-UINT+1 WITHIN -> TRUE }T
+T{ MID-UINT 0 MAX-UINT WITHIN -> TRUE }T
+T{ MID-UINT MID-UINT 0 WITHIN -> TRUE }T
+T{ MID-UINT MID-UINT MID-UINT WITHIN -> FALSE }T
+T{ MID-UINT MID-UINT MID-UINT+1 WITHIN -> TRUE }T
+T{ MID-UINT MID-UINT MAX-UINT WITHIN -> TRUE }T
+T{ MID-UINT MID-UINT+1 0 WITHIN -> FALSE }T
+T{ MID-UINT MID-UINT+1 MID-UINT WITHIN -> FALSE }T
+T{ MID-UINT MID-UINT+1 MID-UINT+1 WITHIN -> FALSE }T
+T{ MID-UINT MID-UINT+1 MAX-UINT WITHIN -> FALSE }T
+T{ MID-UINT MAX-UINT 0 WITHIN -> FALSE }T
+T{ MID-UINT MAX-UINT MID-UINT WITHIN -> FALSE }T
+T{ MID-UINT MAX-UINT MID-UINT+1 WITHIN -> TRUE }T
+T{ MID-UINT MAX-UINT MAX-UINT WITHIN -> FALSE }T
+T{ MID-UINT+1 0 0 WITHIN -> FALSE }T
+T{ MID-UINT+1 0 MID-UINT WITHIN -> FALSE }T
+T{ MID-UINT+1 0 MID-UINT+1 WITHIN -> FALSE }T
+T{ MID-UINT+1 0 MAX-UINT WITHIN -> TRUE }T
+T{ MID-UINT+1 MID-UINT 0 WITHIN -> TRUE }T
+T{ MID-UINT+1 MID-UINT MID-UINT WITHIN -> FALSE }T
+T{ MID-UINT+1 MID-UINT MID-UINT+1 WITHIN -> FALSE }T
+T{ MID-UINT+1 MID-UINT MAX-UINT WITHIN -> TRUE }T
+T{ MID-UINT+1 MID-UINT+1 0 WITHIN -> TRUE }T
+T{ MID-UINT+1 MID-UINT+1 MID-UINT WITHIN -> TRUE }T
+T{ MID-UINT+1 MID-UINT+1 MID-UINT+1 WITHIN -> FALSE }T
+T{ MID-UINT+1 MID-UINT+1 MAX-UINT WITHIN -> TRUE }T
+T{ MID-UINT+1 MAX-UINT 0 WITHIN -> FALSE }T
+T{ MID-UINT+1 MAX-UINT MID-UINT WITHIN -> FALSE }T
+T{ MID-UINT+1 MAX-UINT MID-UINT+1 WITHIN -> FALSE }T
+T{ MID-UINT+1 MAX-UINT MAX-UINT WITHIN -> FALSE }T
+T{ MAX-UINT 0 0 WITHIN -> FALSE }T
+T{ MAX-UINT 0 MID-UINT WITHIN -> FALSE }T
+T{ MAX-UINT 0 MID-UINT+1 WITHIN -> FALSE }T
+T{ MAX-UINT 0 MAX-UINT WITHIN -> FALSE }T
+T{ MAX-UINT MID-UINT 0 WITHIN -> TRUE }T
+T{ MAX-UINT MID-UINT MID-UINT WITHIN -> FALSE }T
+T{ MAX-UINT MID-UINT MID-UINT+1 WITHIN -> FALSE }T
+T{ MAX-UINT MID-UINT MAX-UINT WITHIN -> FALSE }T
+T{ MAX-UINT MID-UINT+1 0 WITHIN -> TRUE }T
+T{ MAX-UINT MID-UINT+1 MID-UINT WITHIN -> TRUE }T
+T{ MAX-UINT MID-UINT+1 MID-UINT+1 WITHIN -> FALSE }T
+T{ MAX-UINT MID-UINT+1 MAX-UINT WITHIN -> FALSE }T
+T{ MAX-UINT MAX-UINT 0 WITHIN -> TRUE }T
+T{ MAX-UINT MAX-UINT MID-UINT WITHIN -> TRUE }T
+T{ MAX-UINT MAX-UINT MID-UINT+1 WITHIN -> TRUE }T
+T{ MAX-UINT MAX-UINT MAX-UINT WITHIN -> FALSE }T
+
+T{ MIN-INT MIN-INT MIN-INT WITHIN -> FALSE }T
+T{ MIN-INT MIN-INT 0 WITHIN -> TRUE }T
+T{ MIN-INT MIN-INT 1 WITHIN -> TRUE }T
+T{ MIN-INT MIN-INT MAX-INT WITHIN -> TRUE }T
+T{ MIN-INT 0 MIN-INT WITHIN -> FALSE }T
+T{ MIN-INT 0 0 WITHIN -> FALSE }T
+T{ MIN-INT 0 1 WITHIN -> FALSE }T
+T{ MIN-INT 0 MAX-INT WITHIN -> FALSE }T
+T{ MIN-INT 1 MIN-INT WITHIN -> FALSE }T
+T{ MIN-INT 1 0 WITHIN -> TRUE }T
+T{ MIN-INT 1 1 WITHIN -> FALSE }T
+T{ MIN-INT 1 MAX-INT WITHIN -> FALSE }T
+T{ MIN-INT MAX-INT MIN-INT WITHIN -> FALSE }T
+T{ MIN-INT MAX-INT 0 WITHIN -> TRUE }T
+T{ MIN-INT MAX-INT 1 WITHIN -> TRUE }T
+T{ MIN-INT MAX-INT MAX-INT WITHIN -> FALSE }T
+T{ 0 MIN-INT MIN-INT WITHIN -> FALSE }T
+T{ 0 MIN-INT 0 WITHIN -> FALSE }T
+T{ 0 MIN-INT 1 WITHIN -> TRUE }T
+T{ 0 MIN-INT MAX-INT WITHIN -> TRUE }T
+T{ 0 0 MIN-INT WITHIN -> TRUE }T
+T{ 0 0 0 WITHIN -> FALSE }T
+T{ 0 0 1 WITHIN -> TRUE }T
+T{ 0 0 MAX-INT WITHIN -> TRUE }T
+T{ 0 1 MIN-INT WITHIN -> FALSE }T
+T{ 0 1 0 WITHIN -> FALSE }T
+T{ 0 1 1 WITHIN -> FALSE }T
+T{ 0 1 MAX-INT WITHIN -> FALSE }T
+T{ 0 MAX-INT MIN-INT WITHIN -> FALSE }T
+T{ 0 MAX-INT 0 WITHIN -> FALSE }T
+T{ 0 MAX-INT 1 WITHIN -> TRUE }T
+T{ 0 MAX-INT MAX-INT WITHIN -> FALSE }T
+T{ 1 MIN-INT MIN-INT WITHIN -> FALSE }T
+T{ 1 MIN-INT 0 WITHIN -> FALSE }T
+T{ 1 MIN-INT 1 WITHIN -> FALSE }T
+T{ 1 MIN-INT MAX-INT WITHIN -> TRUE }T
+T{ 1 0 MIN-INT WITHIN -> TRUE }T
+T{ 1 0 0 WITHIN -> FALSE }T
+T{ 1 0 1 WITHIN -> FALSE }T
+T{ 1 0 MAX-INT WITHIN -> TRUE }T
+T{ 1 1 MIN-INT WITHIN -> TRUE }T
+T{ 1 1 0 WITHIN -> TRUE }T
+T{ 1 1 1 WITHIN -> FALSE }T
+T{ 1 1 MAX-INT WITHIN -> TRUE }T
+T{ 1 MAX-INT MIN-INT WITHIN -> FALSE }T
+T{ 1 MAX-INT 0 WITHIN -> FALSE }T
+T{ 1 MAX-INT 1 WITHIN -> FALSE }T
+T{ 1 MAX-INT MAX-INT WITHIN -> FALSE }T
+T{ MAX-INT MIN-INT MIN-INT WITHIN -> FALSE }T
+T{ MAX-INT MIN-INT 0 WITHIN -> FALSE }T
+T{ MAX-INT MIN-INT 1 WITHIN -> FALSE }T
+T{ MAX-INT MIN-INT MAX-INT WITHIN -> FALSE }T
+T{ MAX-INT 0 MIN-INT WITHIN -> TRUE }T
+T{ MAX-INT 0 0 WITHIN -> FALSE }T
+T{ MAX-INT 0 1 WITHIN -> FALSE }T
+T{ MAX-INT 0 MAX-INT WITHIN -> FALSE }T
+T{ MAX-INT 1 MIN-INT WITHIN -> TRUE }T
+T{ MAX-INT 1 0 WITHIN -> TRUE }T
+T{ MAX-INT 1 1 WITHIN -> FALSE }T
+T{ MAX-INT 1 MAX-INT WITHIN -> FALSE }T
+T{ MAX-INT MAX-INT MIN-INT WITHIN -> TRUE }T
+T{ MAX-INT MAX-INT 0 WITHIN -> TRUE }T
+T{ MAX-INT MAX-INT 1 WITHIN -> TRUE }T
+T{ MAX-INT MAX-INT MAX-INT WITHIN -> FALSE }T
+
+\ -----------------------------------------------------------------------------
+TESTING UNUSED  (contributed by James Bowman & Peter Knaggs)
+
+VARIABLE UNUSED0
+T{ UNUSED DROP -> }T                  
+T{ ALIGN UNUSED UNUSED0 ! 0 , UNUSED CELL+ UNUSED0 @ = -> TRUE }T
+T{ UNUSED UNUSED0 ! 0 C, UNUSED CHAR+ UNUSED0 @ =
+         -> TRUE }T  \ aligned -> unaligned
+T{ UNUSED UNUSED0 ! 0 C, UNUSED CHAR+ UNUSED0 @ = -> TRUE }T  \ unaligned -> ?
+
+\ -----------------------------------------------------------------------------
+TESTING AGAIN   (contributed by James Bowman)
+
+T{ : AG0 701 BEGIN DUP 7 MOD 0= IF EXIT THEN 1+ AGAIN ; -> }T
+T{ AG0 -> 707 }T
+
+\ -----------------------------------------------------------------------------
+TESTING MARKER   (contributed by James Bowman)
+
+T{ : MA? BL WORD FIND NIP 0<> ; -> }T
+T{ MARKER MA0 -> }T
+T{ : MA1 111 ; -> }T
+T{ MARKER MA2 -> }T
+T{ : MA1 222 ; -> }T
+T{ MA? MA0 MA? MA1 MA? MA2 -> TRUE TRUE TRUE }T
+T{ MA1 MA2 MA1 -> 222 111 }T
+T{ MA? MA0 MA? MA1 MA? MA2 -> TRUE TRUE FALSE }T
+T{ MA0 -> }T
+T{ MA? MA0 MA? MA1 MA? MA2 -> FALSE FALSE FALSE }T
+
+\ -----------------------------------------------------------------------------
+TESTING ?DO
+
+: QD ?DO I LOOP ;
+T{ 789 789 QD -> }T
+T{ -9876 -9876 QD -> }T
+T{ 5 0 QD -> 0 1 2 3 4 }T
+
+: QD1 ?DO I 10 +LOOP ;
+T{ 50 1 QD1 -> 1 11 21 31 41 }T
+T{ 50 0 QD1 -> 0 10 20 30 40 }T
+
+: QD2 ?DO I 3 > IF LEAVE ELSE I THEN LOOP ;
+T{ 5 -1 QD2 -> -1 0 1 2 3 }T
+
+: QD3 ?DO I 1 +LOOP ;
+T{ 4  4 QD3 -> }T
+T{ 4  1 QD3 -> 1 2 3 }T
+T{ 2 -1 QD3 -> -1 0 1 }T
+
+: QD4 ?DO I -1 +LOOP ;
+T{  4 4 QD4 -> }T
+T{  1 4 QD4 -> 4 3 2 1 }T
+T{ -1 2 QD4 -> 2 1 0 -1 }T
+
+: QD5 ?DO I -10 +LOOP ;
+T{   1 50 QD5 -> 50 40 30 20 10 }T
+T{   0 50 QD5 -> 50 40 30 20 10 0 }T
+T{ -25 10 QD5 -> 10 0 -10 -20 }T
+
+VARIABLE ITERS
+VARIABLE INCRMNT
+
+: QD6 ( limit start increment -- )
+   INCRMNT !
+   0 ITERS !
+   ?DO
+      1 ITERS +!
+      I
+      ITERS @  6 = IF LEAVE THEN
+      INCRMNT @
+   +LOOP ITERS @
+;
+
+T{  4  4 -1 QD6 -> 0 }T
+T{  1  4 -1 QD6 -> 4 3 2 1 4 }T
+T{  4  1 -1 QD6 -> 1 0 -1 -2 -3 -4 6 }T
+T{  4  1  0 QD6 -> 1 1 1 1 1 1 6 }T
+T{  0  0  0 QD6 -> 0 }T
+T{  1  4  0 QD6 -> 4 4 4 4 4 4 6 }T
+T{  1  4  1 QD6 -> 4 5 6 7 8 9 6 }T
+T{  4  1  1 QD6 -> 1 2 3 3 }T
+T{  4  4  1 QD6 -> 0 }T
+T{  2 -1 -1 QD6 -> -1 -2 -3 -4 -5 -6 6 }T
+T{ -1  2 -1 QD6 -> 2 1 0 -1 4 }T
+T{  2 -1  0 QD6 -> -1 -1 -1 -1 -1 -1 6 }T
+T{ -1  2  0 QD6 -> 2 2 2 2 2 2 6 }T
+T{ -1  2  1 QD6 -> 2 3 4 5 6 7 6 }T
+T{  2 -1  1 QD6 -> -1 0 1 3 }T
+
+\ -----------------------------------------------------------------------------
+TESTING BUFFER:
+
+T{ 2 CELLS BUFFER: BUF:TEST -> }T
+T{ BUF:TEST DUP ALIGNED = -> TRUE }T
+T{ 111 BUF:TEST ! 222 BUF:TEST CELL+ ! -> }T
+T{ BUF:TEST @ BUF:TEST CELL+ @ -> 111 222 }T
+
+\ -----------------------------------------------------------------------------
+TESTING VALUE TO
+
+T{ 111 VALUE VAL1 -999 VALUE VAL2 -> }T
+T{ VAL1 -> 111 }T
+T{ VAL2 -> -999 }T
+T{ 222 TO VAL1 -> }T
+T{ VAL1 -> 222 }T
+T{ : VD1 VAL1 ; -> }T
+T{ VD1 -> 222 }T
+T{ : VD2 TO VAL2 ; -> }T
+T{ VAL2 -> -999 }T
+T{ -333 VD2 -> }T
+T{ VAL2 -> -333 }T
+T{ VAL1 -> 222 }T
+T{ 444 TO VAL1 -> }T
+T{ VD1 -> 444 }T
+T{ 123 VALUE VAL3 IMMEDIATE VAL3 -> 123 }T
+T{ : VD3 VAL3 LITERAL ; VD3 -> 123 }T
+
+\ -----------------------------------------------------------------------------
+TESTING CASE OF ENDOF ENDCASE
+
+: CS1 CASE 1 OF 111 ENDOF
+           2 OF 222 ENDOF
+           3 OF 333 ENDOF
+           >R 999 R>
+      ENDCASE
+;
+
+T{ 1 CS1 -> 111 }T
+T{ 2 CS1 -> 222 }T
+T{ 3 CS1 -> 333 }T
+T{ 4 CS1 -> 999 }T
+
+\ Nested CASE's
+
+: CS2 >R CASE -1 OF CASE R@ 1 OF 100 ENDOF
+                            2 OF 200 ENDOF
+                           >R -300 R>
+                    ENDCASE
+                 ENDOF
+              -2 OF CASE R@ 1 OF -99  ENDOF
+                            >R -199 R>
+                    ENDCASE
+                 ENDOF
+                 >R 299 R>
+         ENDCASE R> DROP
+;
+
+T{ -1 1 CS2 ->  100 }T
+T{ -1 2 CS2 ->  200 }T
+T{ -1 3 CS2 -> -300 }T
+T{ -2 1 CS2 -> -99  }T
+T{ -2 2 CS2 -> -199 }T
+T{  0 2 CS2 ->  299 }T
+
+\ Boolean short circuiting using CASE
+
+: CS3  ( N1 -- N2 )
+   CASE 1- FALSE OF 11 ENDOF
+        1- FALSE OF 22 ENDOF
+        1- FALSE OF 33 ENDOF
+        44 SWAP
+   ENDCASE
+;
+
+T{ 1 CS3 -> 11 }T
+T{ 2 CS3 -> 22 }T
+T{ 3 CS3 -> 33 }T
+T{ 9 CS3 -> 44 }T
+
+\ Empty CASE statements with/without default
+
+T{ : CS4 CASE ENDCASE ; 1 CS4 -> }T
+T{ : CS5 CASE 2 SWAP ENDCASE ; 1 CS5 -> 2 }T
+T{ : CS6 CASE 1 OF ENDOF 2 ENDCASE ; 1 CS6 -> }T
+T{ : CS7 CASE 3 OF ENDOF 2 ENDCASE ; 1 CS7 -> 1 }T
+
+\ -----------------------------------------------------------------------------
+TESTING :NONAME RECURSE
+
+VARIABLE NN1
+VARIABLE NN2
+:NONAME 1234 ; NN1 !
+:NONAME 9876 ; NN2 !
+T{ NN1 @ EXECUTE -> 1234 }T
+T{ NN2 @ EXECUTE -> 9876 }T
+
+T{ :NONAME ( n -- 0,1,..n ) DUP IF DUP >R 1- RECURSE R> THEN ;
+   CONSTANT RN1 -> }T
+T{ 0 RN1 EXECUTE -> 0 }T
+T{ 4 RN1 EXECUTE -> 0 1 2 3 4 }T
+
+:NONAME  ( n -- n1 )    \ Multiple RECURSEs in one definition
+   1- DUP
+   CASE 0 OF EXIT ENDOF
+        1 OF 11 SWAP RECURSE ENDOF
+        2 OF 22 SWAP RECURSE ENDOF
+        3 OF 33 SWAP RECURSE ENDOF
+        DROP ABS RECURSE EXIT
+   ENDCASE
+; CONSTANT RN2
+
+T{  1 RN2 EXECUTE -> 0 }T
+T{  2 RN2 EXECUTE -> 11 0 }T
+T{  4 RN2 EXECUTE -> 33 22 11 0 }T
+T{ 25 RN2 EXECUTE -> 33 22 11 0 }T
+
+\ -----------------------------------------------------------------------------
+TESTING C"
+
+T{ : CQ1 C" 123" ; -> }T
+T{ CQ1 COUNT EVALUATE -> 123 }T
+T{ : CQ2 C" " ; -> }T
+T{ CQ2 COUNT EVALUATE -> }T
+T{ : CQ3 C" 2345"COUNT EVALUATE ; CQ3 -> 2345 }T
+
+\ -----------------------------------------------------------------------------
+TESTING COMPILE,
+
+:NONAME DUP + ; CONSTANT DUP+
+T{ : Q DUP+ COMPILE, ; -> }T
+T{ : AS1 [ Q ] ; -> }T
+T{ 123 AS1 -> 246 }T
+
+\ -----------------------------------------------------------------------------
+\ Cannot automatically test SAVE-INPUT and RESTORE-INPUT from a console source
+
+TESTING SAVE-INPUT and RESTORE-INPUT with a string source
+
+VARIABLE SI_INC 0 SI_INC !
+
+: SI1
+   SI_INC @ >IN +!
+   15 SI_INC !
+;
+
+: S$ S" SAVE-INPUT SI1 RESTORE-INPUT 12345" ;
+
+T{ S$ EVALUATE SI_INC @ -> 0 2345 15 }T
+
+\ -----------------------------------------------------------------------------
+TESTING .(
+
+CR CR .( Output from .() 
+T{ CR .( You should see -9876: ) -9876 . -> }T
+T{ CR .( and again: ).( -9876)CR -> }T
+
+CR CR .( On the next 2 lines you should see First then Second messages:)
+T{ : DOTP  CR ." Second message via ." [CHAR] " EMIT    \ Check .( is immediate
+     [ CR ] .( First message via .( ) ; DOTP -> }T
+CR CR
+T{ : IMM? BL WORD FIND NIP ; IMM? .( -> 1 }T
+
+\ -----------------------------------------------------------------------------
+TESTING .R and U.R - has to handle different cell sizes
+
+\ Create some large integers just below/above MAX and Min INTs
+MAX-INT 73 79 */ CONSTANT LI1
+MIN-INT 71 73 */ CONSTANT LI2
+
+LI1 0 <# #S #> NIP CONSTANT LENLI1
+
+: (.R&U.R)  ( u1 u2 -- )  \ u1 <= string length, u2 is required indentation
+   TUCK + >R
+   LI1 OVER SPACES  . CR R@    LI1 SWAP  .R CR
+   LI2 OVER SPACES  . CR R@ 1+ LI2 SWAP  .R CR
+   LI1 OVER SPACES U. CR R@    LI1 SWAP U.R CR
+   LI2 SWAP SPACES U. CR R>    LI2 SWAP U.R CR
+;
+
+: .R&U.R  ( -- )
+   CR ." You should see lines duplicated:" CR
+   ." indented by 0 spaces" CR 0      0 (.R&U.R) CR
+   ." indented by 0 spaces" CR LENLI1 0 (.R&U.R) CR \ Just fits required width
+   ." indented by 5 spaces" CR LENLI1 5 (.R&U.R) CR
+;
+
+CR CR .( Output from .R and U.R)
+T{ .R&U.R -> }T
+
+\ -----------------------------------------------------------------------------
+TESTING PAD ERASE
+\ Must handle different size characters i.e. 1 CHARS >= 1 
+
+84 CONSTANT CHARS/PAD      \ Minimum size of PAD in chars
+CHARS/PAD CHARS CONSTANT AUS/PAD
+: CHECKPAD  ( caddr u ch -- f )  \ f = TRUE if u chars = ch
+   SWAP 0
+   ?DO
+      OVER I CHARS + C@ OVER <>
+      IF 2DROP UNLOOP FALSE EXIT THEN
+   LOOP  
+   2DROP TRUE
+;
+
+T{ PAD DROP -> }T
+T{ 0 INVERT PAD C! -> }T
+T{ PAD C@ CONSTANT MAXCHAR -> }T
+T{ PAD CHARS/PAD 2DUP MAXCHAR FILL MAXCHAR CHECKPAD -> TRUE }T
+T{ PAD CHARS/PAD 2DUP CHARS ERASE 0 CHECKPAD -> TRUE }T
+T{ PAD CHARS/PAD 2DUP MAXCHAR FILL PAD 0 ERASE MAXCHAR CHECKPAD -> TRUE }T
+T{ PAD 43 CHARS + 9 CHARS ERASE -> }T
+T{ PAD 43 MAXCHAR CHECKPAD -> TRUE }T
+T{ PAD 43 CHARS + 9 0 CHECKPAD -> TRUE }T
+T{ PAD 52 CHARS + CHARS/PAD 52 - MAXCHAR CHECKPAD -> TRUE }T
+
+\ Check that use of WORD and pictured numeric output do not corrupt PAD
+\ Minimum size of buffers for these are 33 chars and (2*n)+2 chars respectively
+\ where n is number of bits per cell
+
+PAD CHARS/PAD ERASE
+2 BASE !
+MAX-UINT MAX-UINT <# #S CHAR 1 DUP HOLD HOLD #> 2DROP
+DECIMAL
+BL WORD 12345678123456781234567812345678 DROP
+T{ PAD CHARS/PAD 0 CHECKPAD -> TRUE }T
+
+\ -----------------------------------------------------------------------------
+TESTING PARSE
+
+T{ CHAR | PARSE 1234| DUP ROT ROT EVALUATE -> 4 1234 }T
+T{ CHAR ^ PARSE  23 45 ^ DUP ROT ROT EVALUATE -> 7 23 45 }T
+: PA1 [CHAR] $ PARSE DUP >R PAD SWAP CHARS MOVE PAD R> ;
+T{ PA1 3456
+   DUP ROT ROT EVALUATE -> 4 3456 }T
+T{ CHAR A PARSE A SWAP DROP -> 0 }T
+T{ CHAR Z PARSE
+   SWAP DROP -> 0 }T
+T{ CHAR " PARSE 4567 "DUP ROT ROT EVALUATE -> 5 4567 }T
+ 
+\ -----------------------------------------------------------------------------
+TESTING PARSE-NAME  (Forth 2012)
+\ Adapted from the PARSE-NAME RfD tests
+
+T{ PARSE-NAME abcd  STR1  S= -> TRUE }T        \ No leading spaces
+T{ PARSE-NAME      abcde STR2 S= -> TRUE }T    \ Leading spaces
+
+\ Test empty parse area, new lines are necessary
+T{ PARSE-NAME
+  NIP -> 0 }T
+\ Empty parse area with spaces after PARSE-NAME
+T{ PARSE-NAME         
+  NIP -> 0 }T
+
+T{ : PARSE-NAME-TEST ( "name1" "name2" -- n )
+    PARSE-NAME PARSE-NAME S= ; -> }T
+T{ PARSE-NAME-TEST abcd abcd  -> TRUE }T
+T{ PARSE-NAME-TEST abcd   abcd  -> TRUE }T  \ Leading spaces
+T{ PARSE-NAME-TEST abcde abcdf -> FALSE }T
+T{ PARSE-NAME-TEST abcdf abcde -> FALSE }T
+T{ PARSE-NAME-TEST abcde abcde
+   -> TRUE }T         \ Parse to end of line
+T{ PARSE-NAME-TEST abcde           abcde         
+   -> TRUE }T         \ Leading and trailing spaces
+
+\ -----------------------------------------------------------------------------
+TESTING DEFER DEFER@ DEFER! IS ACTION-OF (Forth 2012)
+\ Adapted from the Forth 200X RfD tests
+
+T{ DEFER DEFER1 -> }T
+T{ : MY-DEFER DEFER ; -> }T
+T{ : IS-DEFER1 IS DEFER1 ; -> }T
+T{ : ACTION-DEFER1 ACTION-OF DEFER1 ; -> }T
+T{ : DEF! DEFER! ; -> }T
+T{ : DEF@ DEFER@ ; -> }T
+
+T{ ' * ' DEFER1 DEFER! -> }T
+T{ 2 3 DEFER1 -> 6 }T
+T{ ' DEFER1 DEFER@ -> ' * }T
+T{ ' DEFER1 DEF@ -> ' * }T
+T{ ACTION-OF DEFER1 -> ' * }T
+T{ ACTION-DEFER1 -> ' * }T
+T{ ' + IS DEFER1 -> }T
+T{ 1 2 DEFER1 -> 3 }T
+T{ ' DEFER1 DEFER@ -> ' + }T
+T{ ' DEFER1 DEF@ -> ' + }T
+T{ ACTION-OF DEFER1 -> ' + }T
+T{ ACTION-DEFER1 -> ' + }T
+T{ ' - IS-DEFER1 -> }T
+T{ 1 2 DEFER1 -> -1 }T
+T{ ' DEFER1 DEFER@ -> ' - }T
+T{ ' DEFER1 DEF@ -> ' - }T
+T{ ACTION-OF DEFER1 -> ' - }T
+T{ ACTION-DEFER1 -> ' - }T
+
+T{ MY-DEFER DEFER2 -> }T
+T{ ' DUP IS DEFER2 -> }T
+T{ 1 DEFER2 -> 1 1 }T
+
+\ -----------------------------------------------------------------------------
+TESTING HOLDS  (Forth 2012)
+
+: HTEST S" Testing HOLDS" ;
+: HTEST2 S" works" ;
+: HTEST3 S" Testing HOLDS works 123" ;
+T{ 0 0 <#  HTEST HOLDS #> HTEST S= -> TRUE }T
+T{ 123 0 <# #S BL HOLD HTEST2 HOLDS BL HOLD HTEST HOLDS #>
+   HTEST3 S= -> TRUE }T
+T{ : HLD HOLDS ; -> }T
+T{ 0 0 <#  HTEST HLD #> HTEST S= -> TRUE }T
+
+\ -----------------------------------------------------------------------------
+TESTING REFILL SOURCE-ID
+\ REFILL and SOURCE-ID from the user input device can't be tested from a file,
+\ can only be tested from a string via EVALUATE
+
+T{ : RF1  S" REFILL" EVALUATE ; RF1 -> FALSE }T
+T{ : SID1  S" SOURCE-ID" EVALUATE ; SID1 -> -1 }T
+
+\ ------------------------------------------------------------------------------
+TESTING S\"  (Forth 2012 compilation mode)
+\ Extended the Forth 200X RfD tests
+\ Note this tests the Core Ext definition of S\" which has unedfined
+\ interpretation semantics. S\" in interpretation mode is tested in the tests on
+\ the File-Access word set
+
+T{ : SSQ1 S\" abc" S" abc" S= ; -> }T  \ No escapes
+T{ SSQ1 -> TRUE }T
+T{ : SSQ2 S\" " ; SSQ2 SWAP DROP -> 0 }T    \ Empty string
+
+T{ : SSQ3 S\" \a\b\e\f\l\m\q\r\t\v\x0F0\x1Fa\xaBx\z\"\\" ; -> }T
+T{ SSQ3 SWAP DROP          ->  20 }T    \ String length
+T{ SSQ3 DROP            C@ ->   7 }T    \ \a   BEL  Bell
+T{ SSQ3 DROP  1 CHARS + C@ ->   8 }T    \ \b   BS   Backspace
+T{ SSQ3 DROP  2 CHARS + C@ ->  27 }T    \ \e   ESC  Escape
+T{ SSQ3 DROP  3 CHARS + C@ ->  12 }T    \ \f   FF   Form feed
+T{ SSQ3 DROP  4 CHARS + C@ ->  10 }T    \ \l   LF   Line feed
+T{ SSQ3 DROP  5 CHARS + C@ ->  13 }T    \ \m        CR of CR/LF pair
+T{ SSQ3 DROP  6 CHARS + C@ ->  10 }T    \           LF of CR/LF pair
+T{ SSQ3 DROP  7 CHARS + C@ ->  34 }T    \ \q   "    Double Quote
+T{ SSQ3 DROP  8 CHARS + C@ ->  13 }T    \ \r   CR   Carriage Return
+T{ SSQ3 DROP  9 CHARS + C@ ->   9 }T    \ \t   TAB  Horizontal Tab
+T{ SSQ3 DROP 10 CHARS + C@ ->  11 }T    \ \v   VT   Vertical Tab
+T{ SSQ3 DROP 11 CHARS + C@ ->  15 }T    \ \x0F      Given Char
+T{ SSQ3 DROP 12 CHARS + C@ ->  48 }T    \ 0    0    Digit follow on
+T{ SSQ3 DROP 13 CHARS + C@ ->  31 }T    \ \x1F      Given Char
+T{ SSQ3 DROP 14 CHARS + C@ ->  97 }T    \ a    a    Hex follow on
+T{ SSQ3 DROP 15 CHARS + C@ -> 171 }T    \ \xaB      Insensitive Given Char
+T{ SSQ3 DROP 16 CHARS + C@ -> 120 }T    \ x    x    Non hex follow on
+T{ SSQ3 DROP 17 CHARS + C@ ->   0 }T    \ \z   NUL  No Character
+T{ SSQ3 DROP 18 CHARS + C@ ->  34 }T    \ \"   "    Double Quote
+T{ SSQ3 DROP 19 CHARS + C@ ->  92 }T    \ \\   \    Back Slash
+
+\ The above does not test \n as this is a system dependent value.
+\ Check it displays a new line
+CR .( The next test should display:)
+CR .( One line...)
+CR .( another line)
+T{ : SSQ4 S\" \nOne line...\nanotherLine\n" TYPE ; SSQ4 -> }T
+
+\ Test bare escapable characters appear as themselves
+T{ : SSQ5 S\" abeflmnqrtvxz" S" abeflmnqrtvxz" S= ; SSQ5 -> TRUE }T
+
+T{ : SSQ6 S\" a\""2DROP 1111 ; SSQ6 -> 1111 }T \ Parsing behaviour
+
+T{ : SSQ7  S\" 111 : SSQ8 S\\\" 222\" EVALUATE ; SSQ8 333" EVALUATE ; -> }T
+T{ SSQ7 -> 111 222 333 }T
+T{ : SSQ9  S\" 11 : SSQ10 S\\\" \\x32\\x32\" EVALUATE ; SSQ10 33" EVALUATE ; -> }T
+T{ SSQ9 -> 11 22 33 }T
+
+\ -----------------------------------------------------------------------------
+CORE-EXT-ERRORS SET-ERROR-COUNT
+
+CR .( End of Core Extension word tests) CR
+
+

lib/forth/ans-tests/tester.fr

@@ -0,0 +1,66 @@
+\ From: John Hayes S1I
+\ Subject: tester.fr
+\ Date: Mon, 27 Nov 95 13:10:09 PST  
+
+\ (C) 1995 JOHNS HOPKINS UNIVERSITY / APPLIED PHYSICS LABORATORY
+\ MAY BE DISTRIBUTED FREELY AS LONG AS THIS COPYRIGHT NOTICE REMAINS.
+\ VERSION 1.2
+
+\ 24/11/2015 Replaced Core Ext word <> with = 0=
+\ 31/3/2015 Variable #ERRORS added and incremented for each error reported.
+\ 22/1/09 The words { and } have been changed to T{ and }T respectively to
+\ agree with the Forth 200X file ttester.fs. This avoids clashes with
+\ locals using { ... } and the FSL use of } 
+
+HEX
+
+\ SET THE FOLLOWING FLAG TO TRUE FOR MORE VERBOSE OUTPUT; THIS MAY
+\ ALLOW YOU TO TELL WHICH TEST CAUSED YOUR SYSTEM TO HANG.
+VARIABLE VERBOSE
+   FALSE VERBOSE !
+\   TRUE VERBOSE !
+
+: EMPTY-STACK   \ ( ... -- ) EMPTY STACK: HANDLES UNDERFLOWED STACK TOO.
+   DEPTH ?DUP IF DUP 0< IF NEGATE 0 DO 0 LOOP ELSE 0 DO DROP LOOP THEN THEN ;
+
+VARIABLE #ERRORS 0 #ERRORS !
+
+: ERROR      \ ( C-ADDR U -- ) DISPLAY AN ERROR MESSAGE FOLLOWED BY
+      \ THE LINE THAT HAD THE ERROR.
+   CR TYPE SOURCE TYPE       \ DISPLAY LINE CORRESPONDING TO ERROR
+   EMPTY-STACK               \ THROW AWAY EVERY THING ELSE
+   #ERRORS @ 1 + #ERRORS !
+\   QUIT  \ *** Uncomment this line to QUIT on an error
+;
+
+VARIABLE ACTUAL-DEPTH         \ STACK RECORD
+CREATE ACTUAL-RESULTS 20 CELLS ALLOT
+
+: T{      \ ( -- ) SYNTACTIC SUGAR.
+   ;
+
+: ->      \ ( ... -- ) RECORD DEPTH AND CONTENT OF STACK.
+   DEPTH DUP ACTUAL-DEPTH !      \ RECORD DEPTH
+   ?DUP IF            \ IF THERE IS SOMETHING ON STACK
+      0 DO ACTUAL-RESULTS I CELLS + ! LOOP \ SAVE THEM
+   THEN ;
+
+: }T      \ ( ... -- ) COMPARE STACK (EXPECTED) CONTENTS WITH SAVED
+      \ (ACTUAL) CONTENTS.
+   DEPTH ACTUAL-DEPTH @ = IF      \ IF DEPTHS MATCH
+      DEPTH ?DUP IF         \ IF THERE IS SOMETHING ON THE STACK
+         0  DO            \ FOR EACH STACK ITEM
+           ACTUAL-RESULTS I CELLS + @   \ COMPARE ACTUAL WITH EXPECTED
+           = 0= IF S" INCORRECT RESULT: " ERROR LEAVE THEN
+         LOOP
+      THEN
+   ELSE               \ DEPTH MISMATCH
+      S" WRONG NUMBER OF RESULTS: " ERROR
+   THEN ;
+
+: TESTING   \ ( -- ) TALKING COMMENT.
+  SOURCE VERBOSE @
+   IF DUP >R TYPE CR R> >IN !
+   ELSE >IN ! DROP [CHAR] * EMIT
+   THEN ;
+

lib/forth/conformance.sh

@@ -0,0 +1,170 @@
+#!/usr/bin/env bash
+# Run the Hayes/Gerry-Jackson Core conformance suite against our Forth
+# interpreter and emit scoreboard.json + scoreboard.md.
+#
+# Method:
+#   1. Preprocess lib/forth/ans-tests/core.fr — strip \ comments, ( ... )
+#      comments, and TESTING … metadata lines.
+#   2. Split into chunks ending at each `}T` so an error in one test
+#      chunk doesn't abort the run.
+#   3. Emit an SX file that exposes those chunks as a list.
+#   4. Run our Forth + hayes-runner under sx_server; record pass/fail/error.
+
+set -e
+FORTH_DIR="$(cd "$(dirname "$0")" && pwd)"
+ROOT="$(cd "$FORTH_DIR/../.." && pwd)"
+SX_SERVER="${SX_SERVER:-/root/rose-ash/hosts/ocaml/_build/default/bin/sx_server.exe}"
+SOURCE="$FORTH_DIR/ans-tests/core.fr"
+OUT_JSON="$FORTH_DIR/scoreboard.json"
+OUT_MD="$FORTH_DIR/scoreboard.md"
+TMP="$(mktemp -d)"
+PREPROC="$TMP/preproc.forth"
+CHUNKS_SX="$TMP/chunks.sx"
+
+cd "$ROOT"
+
+# 1. preprocess
+awk '
+{
+  line = $0
+  # protect POSTPONE \ so the comment-strip below leaves the literal \ alone
+  gsub(/POSTPONE[ \t]+\\/, "POSTPONE @@BS@@", line)
+  # strip leading/embedded \ line comments (must be \ followed by space or EOL)
+  gsub(/(^|[ \t])\\([ \t].*|$)/, " ", line)
+  # strip ( ... ) block comments that sit on one line
+  gsub(/\([^)]*\)/, " ", line)
+  # strip TESTING … metadata lines (rest of line, incl. bare TESTING)
+  sub(/TESTING([ \t].*)?$/, " ", line)
+  # restore the protected backslash
+  gsub(/@@BS@@/, "\\", line)
+  print line
+}' "$SOURCE" > "$PREPROC"
+
+# 2 + 3: split into chunks at each `}T` and emit as a SX file
+#
+# Cap chunks via MAX_CHUNKS env (default 638 = full Hayes Core). Lower
+# it temporarily if later tests regress into an infinite loop while you
+# are iterating on primitives.
+MAX_CHUNKS="${MAX_CHUNKS:-638}"
+
+MAX_CHUNKS="$MAX_CHUNKS" python3 - "$PREPROC" "$CHUNKS_SX" <<'PY'
+import os, re, sys
+preproc_path, out_path = sys.argv[1], sys.argv[2]
+max_chunks = int(os.environ.get("MAX_CHUNKS", "590"))
+text = open(preproc_path).read()
+# keep the `}T` attached to the preceding chunk
+parts = re.split(r'(\}T)', text)
+chunks = []
+buf = ""
+for p in parts:
+    buf += p
+    if p == "}T":
+        s = buf.strip()
+        if s:
+            chunks.append(s)
+        buf = ""
+if buf.strip():
+    chunks.append(buf.strip())
+chunks = chunks[:max_chunks]
+
+def esc(s):
+    s = s.replace('\\', '\\\\').replace('"', '\\"')
+    s = s.replace('\r', ' ').replace('\n', ' ')
+    s = re.sub(r'\s+', ' ', s).strip()
+    return s
+
+with open(out_path, "w") as f:
+    f.write("(define hayes-chunks (list\n")
+    for c in chunks:
+        f.write('  "' + esc(c) + '"\n')
+    f.write("))\n\n")
+    f.write("(define\n")
+    f.write("  hayes-run-all\n")
+    f.write("  (fn\n")
+    f.write("    ()\n")
+    f.write("    (hayes-reset!)\n")
+    f.write("    (let ((s (hayes-boot)))\n")
+    f.write("      (for-each (fn (c) (hayes-run-chunk s c)) hayes-chunks))\n")
+    f.write("    (hayes-summary)))\n")
+PY
+
+# 4. run it
+OUT=$(printf '(epoch 1)\n(load "lib/forth/runtime.sx")\n(epoch 2)\n(load "lib/forth/reader.sx")\n(epoch 3)\n(load "lib/forth/interpreter.sx")\n(epoch 4)\n(load "lib/forth/compiler.sx")\n(epoch 5)\n(load "lib/forth/hayes-runner.sx")\n(epoch 6)\n(load "%s")\n(epoch 7)\n(eval "(hayes-run-all)")\n' "$CHUNKS_SX" \
+  | timeout 180 "$SX_SERVER" 2>&1)
+STATUS=$?
+
+SUMMARY=$(printf '%s\n' "$OUT" | awk '/^\{:pass / {print; exit}')
+PASS=$(printf '%s' "$SUMMARY" | sed -n 's/.*:pass \([0-9-]*\).*/\1/p')
+FAIL=$(printf '%s' "$SUMMARY" | sed -n 's/.*:fail \([0-9-]*\).*/\1/p')
+ERR=$(printf '%s' "$SUMMARY" | sed -n 's/.*:error \([0-9-]*\).*/\1/p')
+TOTAL=$(printf '%s' "$SUMMARY" | sed -n 's/.*:total \([0-9-]*\).*/\1/p')
+CHUNK_COUNT=$(grep -c '^  "' "$CHUNKS_SX" || echo 0)
+TOTAL_AVAILABLE=$(grep -c '}T' "$PREPROC" || echo 0)
+
+NOW="$(date -u +%Y-%m-%dT%H:%M:%SZ)"
+
+if [ -z "$PASS" ]; then
+  PASS=0; FAIL=0; ERR=0; TOTAL=0
+  NOTE="runner halted before completing (timeout or SX error)"
+else
+  NOTE="completed"
+fi
+
+PCT=0
+if [ "$TOTAL" -gt 0 ]; then
+  PCT=$((PASS * 100 / TOTAL))
+fi
+
+cat > "$OUT_JSON" <<JSON
+{
+  "source": "gerryjackson/forth2012-test-suite src/core.fr",
+  "generated_at": "$NOW",
+  "chunks_available": $TOTAL_AVAILABLE,
+  "chunks_fed": $CHUNK_COUNT,
+  "total": $TOTAL,
+  "pass": $PASS,
+  "fail": $FAIL,
+  "error": $ERR,
+  "percent": $PCT,
+  "note": "$NOTE"
+}
+JSON
+
+cat > "$OUT_MD" <<MD
+# Forth Hayes Core scoreboard
+
+| metric            | value |
+| ----------------- | ----: |
+| chunks available  | $TOTAL_AVAILABLE |
+| chunks fed        | $CHUNK_COUNT |
+| total             | $TOTAL |
+| pass              | $PASS |
+| fail              | $FAIL |
+| error             | $ERR |
+| percent           | ${PCT}% |
+
+- **Source**: \`gerryjackson/forth2012-test-suite\` \`src/core.fr\`
+- **Generated**: $NOW
+- **Note**: $NOTE
+
+A "chunk" is any preprocessed segment ending at a \`}T\` (every Hayes test
+is one chunk, plus the small declaration blocks between tests).
+The runner catches raised errors at chunk boundaries so one bad chunk
+does not abort the rest. \`error\` covers chunks that raised; \`fail\`
+covers tests whose \`->\` / \`}T\` comparison mismatched.
+
+### Chunk cap
+
+\`conformance.sh\` processes the first \`\$MAX_CHUNKS\` chunks (default
+**638**, i.e. the whole Hayes Core file). Lower the cap temporarily
+while iterating on primitives if a regression re-opens an infinite
+loop in later tests.
+MD
+
+echo "$SUMMARY"
+echo "Scoreboard: $OUT_JSON"
+echo "           $OUT_MD"
+
+if [ "$STATUS" -ne 0 ] && [ "$TOTAL" -eq 0 ]; then
+  exit 1
+fi

lib/forth/hayes-runner.sx

@@ -0,0 +1,158 @@
+;; Hayes conformance test runner.
+;; Installs T{ -> }T as Forth primitives that snapshot and compare dstack,
+;; plus stub TESTING / HEX / DECIMAL so the Hayes Core file can stream
+;; through the interpreter without halting on unsupported metadata words.
+
+(define hayes-pass 0)
+(define hayes-fail 0)
+(define hayes-error 0)
+(define hayes-start-depth 0)
+(define hayes-actual (list))
+(define hayes-actual-set false)
+(define hayes-failures (list))
+(define hayes-first-error "")
+(define hayes-error-hist (dict))
+
+(define
+  hayes-reset!
+  (fn
+    ()
+    (set! hayes-pass 0)
+    (set! hayes-fail 0)
+    (set! hayes-error 0)
+    (set! hayes-start-depth 0)
+    (set! hayes-actual (list))
+    (set! hayes-actual-set false)
+    (set! hayes-failures (list))
+    (set! hayes-first-error "")
+    (set! hayes-error-hist (dict))))
+
+(define
+  hayes-slice
+  (fn
+    (state base)
+    (let
+      ((n (- (forth-depth state) base)))
+      (if (<= n 0) (list) (take (get state "dstack") n)))))
+
+(define
+  hayes-truncate!
+  (fn
+    (state base)
+    (let
+      ((n (- (forth-depth state) base)))
+      (when (> n 0) (dict-set! state "dstack" (drop (get state "dstack") n))))))
+
+(define
+  hayes-install!
+  (fn
+    (state)
+    (forth-def-prim!
+      state
+      "T{"
+      (fn
+        (s)
+        (set! hayes-start-depth (forth-depth s))
+        (set! hayes-actual-set false)
+        (set! hayes-actual (list))))
+    (forth-def-prim!
+      state
+      "->"
+      (fn
+        (s)
+        (set! hayes-actual (hayes-slice s hayes-start-depth))
+        (set! hayes-actual-set true)
+        (hayes-truncate! s hayes-start-depth)))
+    (forth-def-prim!
+      state
+      "}T"
+      (fn
+        (s)
+        (let
+          ((expected (hayes-slice s hayes-start-depth)))
+          (hayes-truncate! s hayes-start-depth)
+          (if
+            (and hayes-actual-set (= expected hayes-actual))
+            (set! hayes-pass (+ hayes-pass 1))
+            (begin
+              (set! hayes-fail (+ hayes-fail 1))
+              (set!
+                hayes-failures
+                (concat
+                  hayes-failures
+                  (list
+                    (dict
+                      "kind"
+                      "fail"
+                      "expected"
+                      (str expected)
+                      "actual"
+                      (str hayes-actual))))))))))
+    (forth-def-prim! state "TESTING" (fn (s) nil))
+    ;; HEX/DECIMAL are real primitives now (runtime.sx) — no stub needed.
+    state))
+
+(define
+  hayes-boot
+  (fn () (let ((s (forth-boot))) (hayes-install! s) (hayes-reset!) s)))
+
+;; Run a single preprocessed chunk (string of Forth source) on the shared
+;; state. Catch any raised error and move on — the chunk boundary is a
+;; safe resume point.
+(define
+  hayes-bump-error-key!
+  (fn
+    (err)
+    (let
+      ((msg (str err)))
+      (let
+        ((space-idx (index-of msg " ")))
+        (let
+          ((key
+            (if
+              (> space-idx 0)
+              (substr msg 0 space-idx)
+              msg)))
+          (dict-set!
+            hayes-error-hist
+            key
+            (+ 1 (or (get hayes-error-hist key) 0))))))))
+
+(define
+  hayes-run-chunk
+  (fn
+    (state src)
+    (guard
+      (err
+        ((= 1 1)
+          (begin
+            (set! hayes-error (+ hayes-error 1))
+            (when
+              (= (len hayes-first-error) 0)
+              (set! hayes-first-error (str err)))
+            (hayes-bump-error-key! err)
+            (dict-set! state "dstack" (list))
+            (dict-set! state "rstack" (list))
+            (dict-set! state "compiling" false)
+            (dict-set! state "current-def" nil)
+            (dict-set! state "cstack" (list))
+            (dict-set! state "input" (list)))))
+      (forth-interpret state src))))
+
+(define
+  hayes-summary
+  (fn
+    ()
+    (dict
+      "pass"
+      hayes-pass
+      "fail"
+      hayes-fail
+      "error"
+      hayes-error
+      "total"
+      (+ (+ hayes-pass hayes-fail) hayes-error)
+      "first-error"
+      hayes-first-error
+      "error-hist"
+      hayes-error-hist)))

lib/forth/interpreter.sx

@@ -5,7 +5,39 @@
 
 (define
   forth-execute-word
-  (fn (state word) (let ((body (get word "body"))) (body state))))
+  (fn
+    (state word)
+    (dict-set! word "call-count" (+ 1 (or (get word "call-count") 0)))
+    (let ((body (get word "body"))) (body state))))
+
+(define
+  forth-hot-words
+  (fn
+    (state threshold)
+    (forth-hot-walk
+      (keys (get state "dict"))
+      (get state "dict")
+      threshold
+      (list))))
+
+(define
+  forth-hot-walk
+  (fn
+    (names dict threshold acc)
+    (if
+      (= (len names) 0)
+      acc
+      (let
+        ((n (first names)))
+        (let
+          ((w (get dict n)))
+          (let
+            ((c (or (get w "call-count") 0)))
+            (forth-hot-walk
+              (rest names)
+              dict
+              threshold
+              (if (>= c threshold) (cons (list n c) acc) acc))))))))
 
 (define
   forth-interpret-token
@@ -17,7 +49,7 @@
         (not (nil? w))
         (forth-execute-word state w)
         (let
-          ((n (forth-parse-number tok (get state "base"))))
+          ((n (forth-parse-number tok (get (get state "vars") "base"))))
           (if
             (not (nil? n))
             (forth-push state n)

lib/forth/runtime.sx

@@ -18,10 +18,122 @@
       (dict-set! s "output" "")
       (dict-set! s "compiling" false)
       (dict-set! s "current-def" nil)
-      (dict-set! s "base" 10)
       (dict-set! s "vars" (dict))
+      (dict-set! (get s "vars") "base" 10)
+      (dict-set! s "cstack" (list))
+      (dict-set! s "mem" (dict))
+      (dict-set! s "here" 0)
+      (dict-set! s "hold" (list))
+      (dict-set! s "files" (dict))
+      (dict-set! s "by-path" (dict))
+      (dict-set! s "next-fileid" 1)
       s)))
 
+(define
+  forth-mem-write!
+  (fn (state addr u) (dict-set! (get state "mem") (str addr) u)))
+
+(define
+  forth-mem-read
+  (fn
+    (state addr)
+    (or (get (get state "mem") (str addr)) 0)))
+
+(define
+  forth-alloc-bytes!
+  (fn
+    (state n)
+    (let
+      ((addr (get state "here")))
+      (dict-set! state "here" (+ addr n))
+      addr)))
+
+(define
+  forth-mem-write-string!
+  (fn
+    (state addr s)
+    (let
+      ((n (len s)))
+      (forth-mem-write-string-loop! state addr s 0 n))))
+
+(define
+  forth-mem-write-string-loop!
+  (fn
+    (state addr s i n)
+    (when
+      (< i n)
+      (begin
+        (forth-mem-write! state (+ addr i) (char-code (substr s i 1)))
+        (forth-mem-write-string-loop! state addr s (+ i 1) n)))))
+
+(define
+  forth-mem-read-string
+  (fn
+    (state addr n)
+    (forth-mem-read-string-loop state addr 0 n "")))
+
+(define
+  forth-mem-read-string-loop
+  (fn
+    (state addr i n acc)
+    (if
+      (>= i n)
+      acc
+      (forth-mem-read-string-loop
+        state
+        addr
+        (+ i 1)
+        n
+        (str acc (char-from-code (forth-mem-read state (+ addr i))))))))
+
+(define
+  forth-fill-loop
+  (fn
+    (state addr u char i)
+    (when
+      (< i u)
+      (begin
+        (forth-mem-write! state (+ addr i) char)
+        (forth-fill-loop state addr u char (+ i 1))))))
+
+(define
+  forth-cmove-loop
+  (fn
+    (state src dst u i)
+    (when
+      (< i u)
+      (begin
+        (forth-mem-write! state (+ dst i) (forth-mem-read state (+ src i)))
+        (forth-cmove-loop state src dst u (+ i 1))))))
+
+(define
+  forth-cmove-loop-desc
+  (fn
+    (state src dst u i)
+    (when
+      (>= i 0)
+      (begin
+        (forth-mem-write! state (+ dst i) (forth-mem-read state (+ src i)))
+        (forth-cmove-loop-desc state src dst u (- i 1))))))
+
+(define
+  forth-cpush
+  (fn (state v) (dict-set! state "cstack" (cons v (get state "cstack")))))
+
+(define
+  forth-cpop
+  (fn
+    (state)
+    (let
+      ((cs (get state "cstack")))
+      (if
+        (= (len cs) 0)
+        (forth-error state "control stack underflow")
+        (let
+          ((top (first cs)))
+          (dict-set! state "cstack" (rest cs))
+          top)))))
+
 (define
   forth-error
   (fn (state msg) (dict-set! state "error" msg) (raise msg)))
@@ -81,6 +193,12 @@
   forth-emit-str
   (fn (state s) (dict-set! state "output" (str (get state "output") s))))
 
+;; The body is always a plain SX lambda — primitives and colon-def
+;; bodies alike — which means the SX VM's JIT-on-first-call can lift
+;; the body directly into bytecode. We tag every word `:vm-eligible?
+;; true` so downstream JIT cooperation (a tracing layer, a hot-call
+;; counter) can pick out the JIT-friendly entries by metadata rather
+;; than by inspecting the body shape.
 (define
   forth-make-word
   (fn
@@ -90,6 +208,8 @@
       (dict-set! w "kind" kind)
       (dict-set! w "body" body)
       (dict-set! w "immediate?" immediate?)
+      (dict-set! w "vm-eligible?" true)
+      (dict-set! w "call-count" 0)
       w)))
 
 (define
@@ -99,7 +219,8 @@
     (dict-set!
       (get state "dict")
       (downcase name)
-      (forth-make-word "primitive" body false))))
+      (forth-make-word "primitive" body false))
+    (dict-set! state "last-defined" name)))
 
 (define
   forth-def-prim-imm!
@@ -108,7 +229,8 @@
     (dict-set!
       (get state "dict")
       (downcase name)
-      (forth-make-word "primitive" body true))))
+      (forth-make-word "primitive" body true))
+    (dict-set! state "last-defined" name)))
 
 (define
   forth-lookup
@@ -166,6 +288,220 @@
 
 (define forth-bits-width 32)
 
+;; Truncate a number to the Forth 32-bit signed range (two's-complement).
+;; Used by arithmetic primitives so wrap-around matches ANS semantics and
+;; loop idioms that rely on MSB becoming 0 after enough shifts terminate.
+(define
+  forth-clip
+  (fn
+    (n)
+    (forth-from-unsigned
+      (forth-to-unsigned n forth-bits-width)
+      forth-bits-width)))
+
+;; Double-cell helpers. Single = 32-bit signed, double = 64-bit signed
+;; represented on the data stack as (lo, hi) where hi is on top.
+;; Reassembly converts the low cell as unsigned and the high cell as
+;; signed (signed) or as unsigned (unsigned), then combines.
+(define forth-2pow32 (pow 2 32))
+(define forth-2pow64 (pow 2 64))
+
+(define
+  forth-double-from-cells-u
+  (fn
+    (lo hi)
+    (+ (forth-to-unsigned lo 32) (* (forth-to-unsigned hi 32) forth-2pow32))))
+
+(define
+  forth-double-from-cells-s
+  (fn (lo hi) (+ (forth-to-unsigned lo 32) (* hi forth-2pow32))))
+
+(define
+  forth-double-push-u
+  (fn
+    (state d)
+    (let
+      ((lo (mod d forth-2pow32)) (hi (floor (/ d forth-2pow32))))
+      (forth-push state (forth-from-unsigned lo 32))
+      (forth-push state (forth-from-unsigned hi 32)))))
+
+(define
+  forth-num-to-string-loop
+  (fn
+    (u base acc)
+    (if
+      (= u 0)
+      acc
+      (let
+        ((dig (mod u base)) (rest (floor (/ u base))))
+        (let
+          ((ch
+            (if
+              (< dig 10)
+              (char-from-code (+ 48 dig))
+              (char-from-code (+ 55 dig)))))
+          (forth-num-to-string-loop rest base (str ch acc)))))))
+
+(define
+  forth-num-to-string
+  (fn
+    (u base)
+    (if (= u 0) "0" (forth-num-to-string-loop u base ""))))
+
+(define
+  forth-spaces-str
+  (fn
+    (n)
+    (if (<= n 0) "" (str " " (forth-spaces-str (- n 1))))))
+
+(define
+  forth-join-hold
+  (fn
+    (parts)
+    (forth-join-hold-loop parts "")))
+
+(define
+  forth-join-hold-loop
+  (fn
+    (parts acc)
+    (if
+      (= (len parts) 0)
+      acc
+      (forth-join-hold-loop (rest parts) (str acc (first parts))))))
+
+(define
+  forth-pic-step
+  (fn
+    (state)
+    (let
+      ((hi (forth-pop state)) (lo (forth-pop state)))
+      (let
+        ((d (forth-double-from-cells-u lo hi))
+          (b (get (get state "vars") "base")))
+        (let
+          ((dig (mod d b)) (rest (floor (/ d b))))
+          (let
+            ((ch
+              (if
+                (< dig 10)
+                (char-from-code (+ 48 dig))
+                (char-from-code (+ 55 dig)))))
+            (dict-set! state "hold" (cons ch (get state "hold")))
+            (forth-double-push-u state rest)))))))
+
+(define
+  forth-compare-bytes-loop
+  (fn
+    (state a1 u1 a2 u2 i)
+    (cond
+      ((and (= i u1) (= i u2)) 0)
+      ((= i u1) -1)
+      ((= i u2) 1)
+      (else
+        (let
+          ((b1 (forth-mem-read state (+ a1 i)))
+            (b2 (forth-mem-read state (+ a2 i))))
+          (cond
+            ((< b1 b2) -1)
+            ((> b1 b2) 1)
+            (else (forth-compare-bytes-loop state a1 u1 a2 u2 (+ i 1)))))))))
+
+(define
+  forth-match-at
+  (fn
+    (state a1 start a2 u2 j)
+    (cond
+      ((= j u2) true)
+      ((not
+        (=
+          (forth-mem-read state (+ a1 (+ start j)))
+          (forth-mem-read state (+ a2 j))))
+        false)
+      (else (forth-match-at state a1 start a2 u2 (+ j 1))))))
+
+(define
+  forth-search-bytes
+  (fn
+    (state a1 u1 a2 u2 i)
+    (cond
+      ((= u2 0) 0)
+      ((> (+ i u2) u1) -1)
+      ((forth-match-at state a1 i a2 u2 0) i)
+      (else (forth-search-bytes state a1 u1 a2 u2 (+ i 1))))))
+
+(define
+  forth-digit-of-byte
+  (fn
+    (c base)
+    (let
+      ((v
+        (cond
+          ((and (>= c 48) (<= c 57)) (- c 48))
+          ((and (>= c 65) (<= c 90)) (- c 55))
+          ((and (>= c 97) (<= c 122)) (- c 87))
+          (else -1))))
+      (if (or (< v 0) (>= v base)) -1 v))))
+
+(define
+  forth-numparse-loop
+  (fn
+    (state addr u acc base)
+    (if
+      (= u 0)
+      (list acc addr u)
+      (let
+        ((c (forth-mem-read state addr)))
+        (let
+          ((dig (forth-digit-of-byte c base)))
+          (if
+            (< dig 0)
+            (list acc addr u)
+            (forth-numparse-loop
+              state
+              (+ addr 1)
+              (- u 1)
+              (+ (* acc base) dig)
+              base)))))))
+
+(define
+  forth-pic-S-loop
+  (fn
+    (state)
+    (forth-pic-step state)
+    (let
+      ((hi (forth-pop state)) (lo (forth-pop state)))
+      (if
+        (and (= lo 0) (= hi 0))
+        (begin (forth-push state 0) (forth-push state 0))
+        (begin
+          (forth-push state lo)
+          (forth-push state hi)
+          (forth-pic-S-loop state))))))
+
+(define
+  forth-double-push-s
+  (fn
+    (state d)
+    (if
+      (>= d 0)
+      (forth-double-push-u state d)
+      (let
+        ((q (- 0 d)))
+        (let
+          ((qlo (mod q forth-2pow32)) (qhi (floor (/ q forth-2pow32))))
+          (if
+            (= qlo 0)
+            (begin
+              (forth-push state 0)
+              (forth-push state (forth-from-unsigned (- forth-2pow32 qhi) 32)))
+            (begin
+              (forth-push
+                state
+                (forth-from-unsigned (- forth-2pow32 qlo) 32))
+              (forth-push
+                state
+                (forth-from-unsigned (- (- forth-2pow32 qhi) 1) 32)))))))))
+
 (define
   forth-to-unsigned
   (fn (n w) (let ((m (pow 2 w))) (mod (+ (mod n m) m) m))))
@@ -285,6 +621,19 @@
         (s)
         (let ((a (forth-peek s))) (when (not (= a 0)) (forth-push s a)))))
     (forth-def-prim! state "DEPTH" (fn (s) (forth-push s (forth-depth s))))
+    (forth-def-prim! state "SP@" (fn (s) (forth-push s (forth-depth s))))
+    (forth-def-prim!
+      state
+      "SP!"
+      (fn
+        (s)
+        (let
+          ((n (forth-pop s)))
+          (let
+            ((cur (forth-depth s)))
+            (when
+              (> cur n)
+              (dict-set! s "dstack" (drop (get s "dstack") (- cur n))))))))
     (forth-def-prim!
       state
       "PICK"
@@ -354,11 +703,17 @@
           (forth-push s d)
           (forth-push s a)
           (forth-push s b))))
-    (forth-def-prim! state "+" (forth-binop (fn (a b) (+ a b))))
-    (forth-def-prim! state "-" (forth-binop (fn (a b) (- a b))))
-    (forth-def-prim! state "*" (forth-binop (fn (a b) (* a b))))
-    (forth-def-prim! state "/" (forth-binop forth-div))
-    (forth-def-prim! state "MOD" (forth-binop forth-mod))
+    (forth-def-prim! state "+" (forth-binop (fn (a b) (forth-clip (+ a b)))))
+    (forth-def-prim! state "-" (forth-binop (fn (a b) (forth-clip (- a b)))))
+    (forth-def-prim! state "*" (forth-binop (fn (a b) (forth-clip (* a b)))))
+    (forth-def-prim!
+      state
+      "/"
+      (forth-binop (fn (a b) (forth-clip (forth-div a b)))))
+    (forth-def-prim!
+      state
+      "MOD"
+      (forth-binop (fn (a b) (forth-clip (forth-mod a b)))))
     (forth-def-prim!
       state
       "/MOD"
@@ -368,8 +723,8 @@
           ((b (forth-pop s)) (a (forth-pop s)))
           (forth-push s (forth-mod a b))
           (forth-push s (forth-div a b)))))
-    (forth-def-prim! state "NEGATE" (forth-unop (fn (a) (- 0 a))))
-    (forth-def-prim! state "ABS" (forth-unop abs))
+    (forth-def-prim! state "NEGATE" (forth-unop (fn (a) (forth-clip (- 0 a)))))
+    (forth-def-prim! state "ABS" (forth-unop (fn (a) (forth-clip (abs a)))))
     (forth-def-prim!
       state
       "MIN"
@@ -378,12 +733,15 @@
       state
       "MAX"
       (forth-binop (fn (a b) (if (> a b) a b))))
-    (forth-def-prim! state "1+" (forth-unop (fn (a) (+ a 1))))
-    (forth-def-prim! state "1-" (forth-unop (fn (a) (- a 1))))
-    (forth-def-prim! state "2+" (forth-unop (fn (a) (+ a 2))))
-    (forth-def-prim! state "2-" (forth-unop (fn (a) (- a 2))))
-    (forth-def-prim! state "2*" (forth-unop (fn (a) (* a 2))))
-    (forth-def-prim! state "2/" (forth-unop (fn (a) (floor (/ a 2)))))
+    (forth-def-prim! state "1+" (forth-unop (fn (a) (forth-clip (+ a 1)))))
+    (forth-def-prim! state "1-" (forth-unop (fn (a) (forth-clip (- a 1)))))
+    (forth-def-prim! state "2+" (forth-unop (fn (a) (forth-clip (+ a 2)))))
+    (forth-def-prim! state "2-" (forth-unop (fn (a) (forth-clip (- a 2)))))
+    (forth-def-prim! state "2*" (forth-unop (fn (a) (forth-clip (* a 2)))))
+    (forth-def-prim!
+      state
+      "2/"
+      (forth-unop (fn (a) (forth-clip (floor (/ a 2))))))
     (forth-def-prim! state "=" (forth-cmp (fn (a b) (= a b))))
     (forth-def-prim! state "<>" (forth-cmp (fn (a b) (not (= a b)))))
     (forth-def-prim! state "<" (forth-cmp (fn (a b) (< a b))))
@@ -398,6 +756,30 @@
     (forth-def-prim! state "OR" (forth-binop forth-bit-or))
     (forth-def-prim! state "XOR" (forth-binop forth-bit-xor))
     (forth-def-prim! state "INVERT" (forth-unop forth-bit-invert))
+    (forth-def-prim!
+      state
+      "LSHIFT"
+      (fn
+        (s)
+        (let
+          ((u (forth-pop s)) (x (forth-pop s)))
+          (let
+            ((ux (forth-to-unsigned x forth-bits-width)))
+            (let
+              ((res (mod (* ux (pow 2 u)) (pow 2 forth-bits-width))))
+              (forth-push s (forth-from-unsigned res forth-bits-width)))))))
+    (forth-def-prim!
+      state
+      "RSHIFT"
+      (fn
+        (s)
+        (let
+          ((u (forth-pop s)) (x (forth-pop s)))
+          (let
+            ((ux (forth-to-unsigned x forth-bits-width)))
+            (let
+              ((res (floor (/ ux (pow 2 u)))))
+              (forth-push s (forth-from-unsigned res forth-bits-width)))))))
     (forth-def-prim!
       state
       "."
@@ -416,7 +798,7 @@
     (forth-def-prim!
       state
       "EMIT"
-      (fn (s) (forth-emit-str s (code-char (forth-pop s)))))
+      (fn (s) (forth-emit-str s (char-from-code (forth-pop s)))))
     (forth-def-prim! state "CR" (fn (s) (forth-emit-str s "\n")))
     (forth-def-prim! state "SPACE" (fn (s) (forth-emit-str s " ")))
     (forth-def-prim!
@@ -430,4 +812,459 @@
             (> n 0)
             (for-each (fn (_) (forth-emit-str s " ")) (range 0 n))))))
     (forth-def-prim! state "BL" (fn (s) (forth-push s 32)))
+    (forth-def-prim!
+      state
+      "DECIMAL"
+      (fn (s) (dict-set! (get s "vars") "base" 10)))
+    (forth-def-prim!
+      state
+      "HEX"
+      (fn (s) (dict-set! (get s "vars") "base" 16)))
+    (forth-def-prim!
+      state
+      "OCTAL"
+      (fn (s) (dict-set! (get s "vars") "base" 8)))
+    (forth-def-prim! state "BASE" (fn (s) (forth-push s "base")))
+    (forth-def-prim! state "I" (fn (s) (forth-push s (forth-rpeek s))))
+    (forth-def-prim!
+      state
+      "J"
+      (fn (s) (forth-push s (nth (get s "rstack") 2))))
+    (forth-def-prim! state ">R" (fn (s) (forth-rpush s (forth-pop s))))
+    (forth-def-prim! state "R>" (fn (s) (forth-push s (forth-rpop s))))
+    (forth-def-prim! state "R@" (fn (s) (forth-push s (forth-rpeek s))))
+    (forth-def-prim!
+      state
+      "2>R"
+      (fn
+        (s)
+        (let
+          ((b (forth-pop s)) (a (forth-pop s)))
+          (forth-rpush s a)
+          (forth-rpush s b))))
+    (forth-def-prim!
+      state
+      "2R>"
+      (fn
+        (s)
+        (let
+          ((b (forth-rpop s)) (a (forth-rpop s)))
+          (forth-push s a)
+          (forth-push s b))))
+    (forth-def-prim!
+      state
+      "2R@"
+      (fn
+        (s)
+        (let
+          ((rs (get s "rstack")))
+          (when
+            (< (len rs) 2)
+            (forth-error s "return stack underflow"))
+          (forth-push s (nth rs 1))
+          (forth-push s (nth rs 0)))))
+    (forth-def-prim!
+      state
+      "C@"
+      (fn
+        (s)
+        (let ((addr (forth-pop s))) (forth-push s (forth-mem-read s addr)))))
+    (forth-def-prim!
+      state
+      "C!"
+      (fn
+        (s)
+        (let
+          ((addr (forth-pop s)) (v (forth-pop s)))
+          (forth-mem-write! s addr v))))
+    (forth-def-prim! state "CHAR+" (fn (s) (forth-push s (+ (forth-pop s) 1))))
+    (forth-def-prim! state "CHARS" (fn (s) nil))
+    (forth-def-prim!
+      state
+      "TYPE"
+      (fn
+        (s)
+        (let
+          ((u (forth-pop s)) (addr (forth-pop s)))
+          (forth-emit-str s (forth-mem-read-string s addr u)))))
+    (forth-def-prim!
+      state
+      "COUNT"
+      (fn
+        (s)
+        (let
+          ((addr (forth-pop s)))
+          (let
+            ((u (forth-mem-read s addr)))
+            (forth-push s (+ addr 1))
+            (forth-push s u)))))
+    (forth-def-prim!
+      state
+      "FILL"
+      (fn
+        (s)
+        (let
+          ((char (forth-pop s)) (u (forth-pop s)) (addr (forth-pop s)))
+          (forth-fill-loop s addr u char 0))))
+    (forth-def-prim!
+      state
+      "BLANK"
+      (fn
+        (s)
+        (let
+          ((u (forth-pop s)) (addr (forth-pop s)))
+          (forth-fill-loop s addr u 32 0))))
+    (forth-def-prim!
+      state
+      "CMOVE"
+      (fn
+        (s)
+        (let
+          ((u (forth-pop s)) (dst (forth-pop s)) (src (forth-pop s)))
+          (forth-cmove-loop s src dst u 0))))
+    (forth-def-prim!
+      state
+      "CMOVE>"
+      (fn
+        (s)
+        (let
+          ((u (forth-pop s)) (dst (forth-pop s)) (src (forth-pop s)))
+          (forth-cmove-loop-desc s src dst u (- u 1)))))
+    (forth-def-prim!
+      state
+      "MOVE"
+      (fn
+        (s)
+        (let
+          ((u (forth-pop s)) (dst (forth-pop s)) (src (forth-pop s)))
+          (if
+            (or (<= dst src) (>= dst (+ src u)))
+            (forth-cmove-loop s src dst u 0)
+            (forth-cmove-loop-desc s src dst u (- u 1))))))
+    (forth-def-prim!
+      state
+      "S>D"
+      (fn
+        (s)
+        (let
+          ((n (forth-pop s)))
+          (forth-push s n)
+          (forth-push s (if (< n 0) -1 0)))))
+    (forth-def-prim! state "D>S" (fn (s) (forth-pop s)))
+    (forth-def-prim!
+      state
+      "M*"
+      (fn
+        (s)
+        (let
+          ((b (forth-pop s)) (a (forth-pop s)))
+          (forth-double-push-s s (* a b)))))
+    (forth-def-prim!
+      state
+      "UM*"
+      (fn
+        (s)
+        (let
+          ((b (forth-pop s)) (a (forth-pop s)))
+          (forth-double-push-u
+            s
+            (* (forth-to-unsigned a 32) (forth-to-unsigned b 32))))))
+    (forth-def-prim!
+      state
+      "UM/MOD"
+      (fn
+        (s)
+        (let
+          ((u1 (forth-pop s)) (hi (forth-pop s)) (lo (forth-pop s)))
+          (let
+            ((d (forth-double-from-cells-u lo hi))
+              (divisor (forth-to-unsigned u1 32)))
+            (when (= divisor 0) (forth-error s "division by zero"))
+            (let
+              ((q (floor (/ d divisor))) (r (mod d divisor)))
+              (forth-push s (forth-from-unsigned r 32))
+              (forth-push s (forth-from-unsigned q 32)))))))
+    (forth-def-prim!
+      state
+      "FM/MOD"
+      (fn
+        (s)
+        (let
+          ((n (forth-pop s)) (hi (forth-pop s)) (lo (forth-pop s)))
+          (let
+            ((d (forth-double-from-cells-s lo hi)))
+            (when (= n 0) (forth-error s "division by zero"))
+            (let
+              ((q (floor (/ d n))))
+              (let
+                ((r (- d (* q n))))
+                (forth-push s (forth-clip r))
+                (forth-push s (forth-clip q))))))))
+    (forth-def-prim!
+      state
+      "SM/REM"
+      (fn
+        (s)
+        (let
+          ((n (forth-pop s)) (hi (forth-pop s)) (lo (forth-pop s)))
+          (let
+            ((d (forth-double-from-cells-s lo hi)))
+            (when (= n 0) (forth-error s "division by zero"))
+            (let
+              ((q (forth-trunc (/ d n))))
+              (let
+                ((r (- d (* q n))))
+                (forth-push s (forth-clip r))
+                (forth-push s (forth-clip q))))))))
+    (forth-def-prim!
+      state
+      "*/"
+      (fn
+        (s)
+        (let
+          ((n3 (forth-pop s)) (n2 (forth-pop s)) (n1 (forth-pop s)))
+          (when (= n3 0) (forth-error s "division by zero"))
+          (forth-push s (forth-clip (forth-trunc (/ (* n1 n2) n3)))))))
+    (forth-def-prim!
+      state
+      "*/MOD"
+      (fn
+        (s)
+        (let
+          ((n3 (forth-pop s)) (n2 (forth-pop s)) (n1 (forth-pop s)))
+          (when (= n3 0) (forth-error s "division by zero"))
+          (let
+            ((d (* n1 n2)))
+            (let
+              ((q (forth-trunc (/ d n3))))
+              (let
+                ((r (- d (* q n3))))
+                (forth-push s (forth-clip r))
+                (forth-push s (forth-clip q))))))))
+    (forth-def-prim!
+      state
+      "D+"
+      (fn
+        (s)
+        (let
+          ((hi2 (forth-pop s))
+            (lo2 (forth-pop s))
+            (hi1 (forth-pop s))
+            (lo1 (forth-pop s)))
+          (forth-double-push-s
+            s
+            (+
+              (forth-double-from-cells-s lo1 hi1)
+              (forth-double-from-cells-s lo2 hi2))))))
+    (forth-def-prim!
+      state
+      "D-"
+      (fn
+        (s)
+        (let
+          ((hi2 (forth-pop s))
+            (lo2 (forth-pop s))
+            (hi1 (forth-pop s))
+            (lo1 (forth-pop s)))
+          (forth-double-push-s
+            s
+            (-
+              (forth-double-from-cells-s lo1 hi1)
+              (forth-double-from-cells-s lo2 hi2))))))
+    (forth-def-prim!
+      state
+      "DNEGATE"
+      (fn
+        (s)
+        (let
+          ((hi (forth-pop s)) (lo (forth-pop s)))
+          (forth-double-push-s
+            s
+            (- 0 (forth-double-from-cells-s lo hi))))))
+    (forth-def-prim!
+      state
+      "DABS"
+      (fn
+        (s)
+        (let
+          ((hi (forth-pop s)) (lo (forth-pop s)))
+          (forth-double-push-s s (abs (forth-double-from-cells-s lo hi))))))
+    (forth-def-prim!
+      state
+      "D="
+      (fn
+        (s)
+        (let
+          ((hi2 (forth-pop s))
+            (lo2 (forth-pop s))
+            (hi1 (forth-pop s))
+            (lo1 (forth-pop s)))
+          (forth-push s (if (and (= lo1 lo2) (= hi1 hi2)) -1 0)))))
+    (forth-def-prim!
+      state
+      "D<"
+      (fn
+        (s)
+        (let
+          ((hi2 (forth-pop s))
+            (lo2 (forth-pop s))
+            (hi1 (forth-pop s))
+            (lo1 (forth-pop s)))
+          (forth-push
+            s
+            (if
+              (<
+                (forth-double-from-cells-s lo1 hi1)
+                (forth-double-from-cells-s lo2 hi2))
+              -1
+              0)))))
+    (forth-def-prim!
+      state
+      "D0="
+      (fn
+        (s)
+        (let
+          ((hi (forth-pop s)) (lo (forth-pop s)))
+          (forth-push s (if (and (= lo 0) (= hi 0)) -1 0)))))
+    (forth-def-prim!
+      state
+      "D0<"
+      (fn
+        (s)
+        (let
+          ((hi (forth-pop s)) (lo (forth-pop s)))
+          (forth-push s (if (< hi 0) -1 0)))))
+    (forth-def-prim!
+      state
+      "DMAX"
+      (fn
+        (s)
+        (let
+          ((hi2 (forth-pop s))
+            (lo2 (forth-pop s))
+            (hi1 (forth-pop s))
+            (lo1 (forth-pop s)))
+          (let
+            ((d1 (forth-double-from-cells-s lo1 hi1))
+              (d2 (forth-double-from-cells-s lo2 hi2)))
+            (forth-double-push-s s (if (> d1 d2) d1 d2))))))
+    (forth-def-prim!
+      state
+      "DMIN"
+      (fn
+        (s)
+        (let
+          ((hi2 (forth-pop s))
+            (lo2 (forth-pop s))
+            (hi1 (forth-pop s))
+            (lo1 (forth-pop s)))
+          (let
+            ((d1 (forth-double-from-cells-s lo1 hi1))
+              (d2 (forth-double-from-cells-s lo2 hi2)))
+            (forth-double-push-s s (if (< d1 d2) d1 d2))))))
+    (forth-def-prim! state "<#" (fn (s) (dict-set! s "hold" (list))))
+    (forth-def-prim!
+      state
+      "HOLD"
+      (fn
+        (s)
+        (let
+          ((c (forth-pop s)))
+          (dict-set!
+            s
+            "hold"
+            (cons (char-from-code c) (get s "hold"))))))
+    (forth-def-prim!
+      state
+      "SIGN"
+      (fn
+        (s)
+        (let
+          ((n (forth-pop s)))
+          (when
+            (< n 0)
+            (dict-set! s "hold" (cons "-" (get s "hold")))))))
+    (forth-def-prim!
+      state
+      "#"
+      (fn
+        (s)
+        (let
+          ((hi (forth-pop s)) (lo (forth-pop s)))
+          (let
+            ((d (forth-double-from-cells-u lo hi))
+              (b (get (get s "vars") "base")))
+            (let
+              ((dig (mod d b)) (rest (floor (/ d b))))
+              (let
+                ((ch
+                  (if
+                    (< dig 10)
+                    (char-from-code (+ 48 dig))
+                    (char-from-code (+ 55 dig)))))
+                (dict-set! s "hold" (cons ch (get s "hold")))
+                (forth-double-push-u s rest)))))))
+    (forth-def-prim!
+      state
+      "#S"
+      (fn
+        (s)
+        (forth-pic-S-loop s)))
+    (forth-def-prim!
+      state
+      "#>"
+      (fn
+        (s)
+        (forth-pop s)
+        (forth-pop s)
+        (let
+          ((str-out (forth-join-hold (get s "hold"))))
+          (let
+            ((addr (forth-alloc-bytes! s (len str-out))))
+            (forth-mem-write-string! s addr str-out)
+            (forth-push s addr)
+            (forth-push s (len str-out))))))
+    (forth-def-prim!
+      state
+      "U."
+      (fn
+        (s)
+        (let
+          ((u (forth-to-unsigned (forth-pop s) 32))
+            (b (get (get s "vars") "base")))
+          (forth-emit-str s (str (forth-num-to-string u b) " ")))))
+    (forth-def-prim!
+      state
+      "U.R"
+      (fn
+        (s)
+        (let
+          ((width (forth-pop s))
+            (u (forth-to-unsigned (forth-pop s) 32))
+            (b (get (get s "vars") "base")))
+          (let
+            ((digits (forth-num-to-string u b)))
+            (forth-emit-str
+              s
+              (forth-spaces-str (- width (len digits))))
+            (forth-emit-str s digits)))))
+    (forth-def-prim!
+      state
+      ".R"
+      (fn
+        (s)
+        (let
+          ((width (forth-pop s))
+            (n (forth-pop s))
+            (b (get (get s "vars") "base")))
+          (let
+            ((sign-prefix (if (< n 0) "-" ""))
+              (abs-digits
+                (forth-num-to-string (forth-to-unsigned (abs n) 32) b)))
+            (let
+              ((digits (str sign-prefix abs-digits)))
+              (forth-emit-str
+                s
+                (forth-spaces-str (- width (len digits))))
+              (forth-emit-str s digits))))))
     state))

lib/forth/scoreboard.json

@@ -0,0 +1,12 @@
+{
+  "source": "gerryjackson/forth2012-test-suite src/core.fr",
+  "generated_at": "2026-04-25T04:57:22Z",
+  "chunks_available": 638,
+  "chunks_fed": 638,
+  "total": 638,
+  "pass": 618,
+  "fail": 14,
+  "error": 6,
+  "percent": 96,
+  "note": "completed"
+}

lib/forth/scoreboard.md

@@ -0,0 +1,28 @@
+# Forth Hayes Core scoreboard
+
+| metric            | value |
+| ----------------- | ----: |
+| chunks available  | 638 |
+| chunks fed        | 638 |
+| total             | 638 |
+| pass              | 618 |
+| fail              | 14 |
+| error             | 6 |
+| percent           | 96% |
+
+- **Source**: `gerryjackson/forth2012-test-suite` `src/core.fr`
+- **Generated**: 2026-04-25T04:57:22Z
+- **Note**: completed
+
+A "chunk" is any preprocessed segment ending at a `}T` (every Hayes test
+is one chunk, plus the small declaration blocks between tests).
+The runner catches raised errors at chunk boundaries so one bad chunk
+does not abort the rest. `error` covers chunks that raised; `fail`
+covers tests whose `->` / `}T` comparison mismatched.
+
+### Chunk cap
+
+`conformance.sh` processes the first `$MAX_CHUNKS` chunks (default
+**638**, i.e. the whole Hayes Core file). Lower the cap temporarily
+while iterating on primitives if a regression re-opens an infinite
+loop in later tests.

lib/forth/tests/test-phase3.sx

@@ -0,0 +1,239 @@
+;; Phase 3 — control flow (IF/ELSE/THEN, BEGIN/UNTIL/WHILE/REPEAT/AGAIN,
+;; DO/LOOP, return stack). Grows as each control construct lands.
+
+(define forth-p3-passed 0)
+(define forth-p3-failed 0)
+(define forth-p3-failures (list))
+
+(define
+  forth-p3-assert
+  (fn
+    (label expected actual)
+    (if
+      (= expected actual)
+      (set! forth-p3-passed (+ forth-p3-passed 1))
+      (begin
+        (set! forth-p3-failed (+ forth-p3-failed 1))
+        (set!
+          forth-p3-failures
+          (concat
+            forth-p3-failures
+            (list
+              (str label ": expected " (str expected) " got " (str actual)))))))))
+
+(define
+  forth-p3-check-stack
+  (fn
+    (label src expected)
+    (let ((r (forth-run src))) (forth-p3-assert label expected (nth r 2)))))
+
+(define
+  forth-p3-if-tests
+  (fn
+    ()
+    (forth-p3-check-stack
+      "IF taken (-1)"
+      ": Q -1 IF 10 THEN ; Q"
+      (list 10))
+    (forth-p3-check-stack
+      "IF not taken (0)"
+      ": Q 0 IF 10 THEN ; Q"
+      (list))
+    (forth-p3-check-stack
+      "IF with non-zero truthy"
+      ": Q 42 IF 10 THEN ; Q"
+      (list 10))
+    (forth-p3-check-stack
+      "IF ELSE — true branch"
+      ": Q -1 IF 10 ELSE 20 THEN ; Q"
+      (list 10))
+    (forth-p3-check-stack
+      "IF ELSE — false branch"
+      ": Q 0 IF 10 ELSE 20 THEN ; Q"
+      (list 20))
+    (forth-p3-check-stack
+      "IF consumes flag"
+      ": Q IF 1 ELSE 2 THEN ; 0 Q"
+      (list 2))
+    (forth-p3-check-stack
+      "absolute value via IF"
+      ": ABS2 DUP 0 < IF NEGATE THEN ; -7 ABS2"
+      (list 7))
+    (forth-p3-check-stack
+      "abs leaves positive alone"
+      ": ABS2 DUP 0 < IF NEGATE THEN ; 7 ABS2"
+      (list 7))
+    (forth-p3-check-stack
+      "sign: negative"
+      ": SIGN DUP 0 < IF DROP -1 ELSE DROP 1 THEN ; -3 SIGN"
+      (list -1))
+    (forth-p3-check-stack
+      "sign: positive"
+      ": SIGN DUP 0 < IF DROP -1 ELSE DROP 1 THEN ; 3 SIGN"
+      (list 1))
+    (forth-p3-check-stack
+      "nested IF (both true)"
+      ": Q 1 IF 1 IF 10 ELSE 20 THEN ELSE 30 THEN ; Q"
+      (list 10))
+    (forth-p3-check-stack
+      "nested IF (inner false)"
+      ": Q 1 IF 0 IF 10 ELSE 20 THEN ELSE 30 THEN ; Q"
+      (list 20))
+    (forth-p3-check-stack
+      "nested IF (outer false)"
+      ": Q 0 IF 0 IF 10 ELSE 20 THEN ELSE 30 THEN ; Q"
+      (list 30))
+    (forth-p3-check-stack
+      "IF before other ops"
+      ": Q 1 IF 5 ELSE 6 THEN 2 * ; Q"
+      (list 10))
+    (forth-p3-check-stack
+      "IF in chained def"
+      ": POS? 0 > ;
+       : DOUBLE-IF-POS DUP POS? IF 2 * THEN ;
+       3 DOUBLE-IF-POS"
+      (list 6))
+    (forth-p3-check-stack
+      "empty then branch"
+      ": Q 1 IF THEN 99 ; Q"
+      (list 99))
+    (forth-p3-check-stack
+      "empty else branch"
+      ": Q 0 IF 99 ELSE THEN ; Q"
+      (list))
+    (forth-p3-check-stack
+      "sequential IF blocks"
+      ": Q -1 IF 1 THEN -1 IF 2 THEN ; Q"
+      (list 1 2))))
+
+(define
+  forth-p3-loop-tests
+  (fn
+    ()
+    (forth-p3-check-stack
+      "BEGIN UNTIL (countdown to zero)"
+      ": CD BEGIN 1- DUP 0 = UNTIL ; 3 CD"
+      (list 0))
+    (forth-p3-check-stack
+      "BEGIN UNTIL — single pass (UNTIL true immediately)"
+      ": Q BEGIN -1 UNTIL 42 ; Q"
+      (list 42))
+    (forth-p3-check-stack
+      "BEGIN UNTIL — accumulate sum 1+2+3"
+      ": SUM3 0 3 BEGIN TUCK + SWAP 1- DUP 0 = UNTIL DROP ; SUM3"
+      (list 6))
+    (forth-p3-check-stack
+      "BEGIN WHILE REPEAT — triangular sum 5"
+      ": TRI 0 5 BEGIN DUP 0 > WHILE TUCK + SWAP 1- REPEAT DROP ; TRI"
+      (list 15))
+    (forth-p3-check-stack
+      "BEGIN WHILE REPEAT — zero iterations"
+      ": TRI 0 0 BEGIN DUP 0 > WHILE TUCK + SWAP 1- REPEAT DROP ; TRI"
+      (list 0))
+    (forth-p3-check-stack
+      "BEGIN WHILE REPEAT — one iteration"
+      ": TRI 0 1 BEGIN DUP 0 > WHILE TUCK + SWAP 1- REPEAT DROP ; TRI"
+      (list 1))
+    (forth-p3-check-stack
+      "nested BEGIN UNTIL"
+      ": INNER BEGIN 1- DUP 0 = UNTIL DROP ;
+       : OUTER BEGIN 3 INNER 1- DUP 0 = UNTIL ;
+       2 OUTER"
+      (list 0))
+    (forth-p3-check-stack
+      "BEGIN UNTIL after colon prefix"
+      ": TEN 10 ;
+       : CD TEN BEGIN 1- DUP 0 = UNTIL ;
+       CD"
+      (list 0))
+    (forth-p3-check-stack
+      "WHILE inside IF branch"
+      ": Q 1 IF 0 3 BEGIN DUP 0 > WHILE TUCK + SWAP 1- REPEAT DROP ELSE 99 THEN ; Q"
+      (list 6))))
+
+(define
+  forth-p3-do-tests
+  (fn
+    ()
+    (forth-p3-check-stack
+      "DO LOOP — simple sum 0..4"
+      ": SUM 0 5 0 DO I + LOOP ; SUM"
+      (list 10))
+    (forth-p3-check-stack
+      "DO LOOP — 10..14 sum using I"
+      ": SUM 0 15 10 DO I + LOOP ; SUM"
+      (list 60))
+    (forth-p3-check-stack
+      "DO LOOP — limit = start runs one pass"
+      ": SUM 0 5 5 DO I + LOOP ; SUM"
+      (list 5))
+    (forth-p3-check-stack
+      "DO LOOP — count iterations"
+      ": COUNT 0 4 0 DO 1+ LOOP ; COUNT"
+      (list 4))
+    (forth-p3-check-stack
+      "DO LOOP — nested, I inner / J outer"
+      ": MATRIX 0 3 0 DO 3 0 DO I J + + LOOP LOOP ; MATRIX"
+      (list 18))
+    (forth-p3-check-stack
+      "DO LOOP — I used in arithmetic"
+      ": DBL 0 5 1 DO I 2 * + LOOP ; DBL"
+      (list 20))
+    (forth-p3-check-stack
+      "+LOOP — count by 2"
+      ": Q 0 10 0 DO I + 2 +LOOP ; Q"
+      (list 20))
+    (forth-p3-check-stack
+      "+LOOP — count by 3"
+      ": Q 0 10 0 DO I + 3 +LOOP ; Q"
+      (list 18))
+    (forth-p3-check-stack
+      "+LOOP — negative step"
+      ": Q 0 0 10 DO I + -1 +LOOP ; Q"
+      (list 55))
+    (forth-p3-check-stack
+      "LEAVE — early exit at I=3"
+      ": Q 0 10 0 DO I 3 = IF LEAVE THEN I + LOOP ; Q"
+      (list 3))
+    (forth-p3-check-stack
+      "LEAVE — in nested loop exits only inner"
+      ": Q 0 3 0 DO 5 0 DO I 2 = IF LEAVE THEN I + LOOP LOOP ; Q"
+      (list 3))
+    (forth-p3-check-stack
+      "DO LOOP preserves outer stack"
+      ": Q 99 5 0 DO I + LOOP ; Q"
+      (list 109))
+    (forth-p3-check-stack
+      ">R R>"
+      ": Q 7 >R 11 R> ; Q"
+      (list 11 7))
+    (forth-p3-check-stack
+      ">R R@ R>"
+      ": Q 7 >R R@ R> ; Q"
+      (list 7 7))
+    (forth-p3-check-stack
+      "2>R 2R>"
+      ": Q 1 2 2>R 99 2R> ; Q"
+      (list 99 1 2))
+    (forth-p3-check-stack
+      "2>R 2R@ 2R>"
+      ": Q 3 4 2>R 2R@ 2R> ; Q"
+      (list 3 4 3 4))))
+
+(define
+  forth-p3-run-all
+  (fn
+    ()
+    (set! forth-p3-passed 0)
+    (set! forth-p3-failed 0)
+    (set! forth-p3-failures (list))
+    (forth-p3-if-tests)
+    (forth-p3-loop-tests)
+    (forth-p3-do-tests)
+    (dict
+      "passed"
+      forth-p3-passed
+      "failed"
+      forth-p3-failed
+      "failures"
+      forth-p3-failures)))

lib/forth/tests/test-phase4.sx

@@ -0,0 +1,268 @@
+;; Phase 4 — strings + more Core.
+;; Uses the byte-memory model on state ("mem" dict + "here" cursor).
+
+(define forth-p4-passed 0)
+(define forth-p4-failed 0)
+(define forth-p4-failures (list))
+
+(define
+  forth-p4-assert
+  (fn
+    (label expected actual)
+    (if
+      (= expected actual)
+      (set! forth-p4-passed (+ forth-p4-passed 1))
+      (begin
+        (set! forth-p4-failed (+ forth-p4-failed 1))
+        (set!
+          forth-p4-failures
+          (concat
+            forth-p4-failures
+            (list
+              (str label ": expected " (str expected) " got " (str actual)))))))))
+
+(define
+  forth-p4-check-output
+  (fn
+    (label src expected)
+    (let ((r (forth-run src))) (forth-p4-assert label expected (nth r 1)))))
+
+(define
+  forth-p4-check-stack-size
+  (fn
+    (label src expected-n)
+    (let
+      ((r (forth-run src)))
+      (forth-p4-assert label expected-n (len (nth r 2))))))
+
+(define
+  forth-p4-check-top
+  (fn
+    (label src expected)
+    (let
+      ((r (forth-run src)))
+      (let
+        ((stk (nth r 2)))
+        (forth-p4-assert label expected (nth stk (- (len stk) 1)))))))
+
+(define
+  forth-p4-check-typed
+  (fn
+    (label src expected)
+    (forth-p4-check-output label (str src " TYPE") expected)))
+
+(define
+  forth-p4-string-tests
+  (fn
+    ()
+    (forth-p4-check-typed
+      "S\" + TYPE — hello"
+      "S\" HELLO\""
+      "HELLO")
+    (forth-p4-check-typed
+      "S\" + TYPE — two words"
+      "S\" HELLO WORLD\""
+      "HELLO WORLD")
+    (forth-p4-check-typed
+      "S\" + TYPE — empty"
+      "S\" \""
+      "")
+    (forth-p4-check-typed
+      "S\" + TYPE — single char"
+      "S\" X\""
+      "X")
+    (forth-p4-check-stack-size
+      "S\" pushes (addr len)"
+      "S\" HI\""
+      2)
+    (forth-p4-check-top
+      "S\" length is correct"
+      "S\" HELLO\""
+      5)
+    (forth-p4-check-output
+      ".\" prints at interpret time"
+      ".\" HELLO\""
+      "HELLO")
+    (forth-p4-check-output
+      ".\" in colon def"
+      ": GREET .\" HI \" ; GREET GREET"
+      "HI HI ")))
+
+(define
+  forth-p4-count-tests
+  (fn
+    ()
+    (forth-p4-check-typed
+      "C\" + COUNT + TYPE"
+      "C\" ABC\" COUNT"
+      "ABC")
+    (forth-p4-check-typed
+      "C\" then COUNT leaves right len"
+      "C\" HI THERE\" COUNT"
+      "HI THERE")))
+
+(define
+  forth-p4-fill-tests
+  (fn
+    ()
+    (forth-p4-check-typed
+      "FILL overwrites prefix bytes"
+      "S\" ABCDE\" 2DUP DROP 3 65 FILL"
+      "AAADE")
+    (forth-p4-check-typed
+      "BLANK sets spaces"
+      "S\" XYZAB\" 2DUP DROP 3 BLANK"
+      "   AB")))
+
+(define
+  forth-p4-cmove-tests
+  (fn
+    ()
+    (forth-p4-check-output
+      "CMOVE copies HELLO forward"
+      ": MKH 72 0 C! 69 1 C! 76 2 C! 76 3 C! 79 4 C! ;
+       : T MKH 0 10 5 CMOVE 10 5 TYPE ; T"
+      "HELLO")
+    (forth-p4-check-output
+      "CMOVE> copies overlapping backward"
+      ": MKA 65 0 C! 66 1 C! 67 2 C! ;
+       : T MKA 0 1 2 CMOVE> 0 3 TYPE ; T"
+      "AAB")
+    (forth-p4-check-output
+      "MOVE picks direction for overlap"
+      ": MKA 65 0 C! 66 1 C! 67 2 C! ;
+       : T MKA 0 1 2 MOVE 0 3 TYPE ; T"
+      "AAB")))
+
+(define
+  forth-p4-charplus-tests
+  (fn
+    ()
+    (forth-p4-check-top
+      "CHAR+ increments"
+      "5 CHAR+"
+      6)))
+
+(define
+  forth-p4-char-tests
+  (fn
+    ()
+    (forth-p4-check-top "CHAR A -> 65" "CHAR A" 65)
+    (forth-p4-check-top "CHAR x -> 120" "CHAR x" 120)
+    (forth-p4-check-top "CHAR takes only first char" "CHAR HELLO" 72)
+    (forth-p4-check-top
+      "[CHAR] compiles literal"
+      ": AA [CHAR] A ; AA"
+      65)
+    (forth-p4-check-top
+      "[CHAR] reads past IMMEDIATE"
+      ": ZZ [CHAR] Z ; ZZ"
+      90)
+    (forth-p4-check-stack-size
+      "[CHAR] doesn't leak at compile time"
+      ": FOO [CHAR] A ; "
+      0)))
+
+(define
+  forth-p4-key-accept-tests
+  (fn
+    ()
+    (let
+      ((r (forth-run "1000 2 ACCEPT")))
+      (let ((stk (nth r 2))) (forth-p4-assert "ACCEPT empty buf -> 0" (list 0) stk)))))
+
+(define
+  forth-p4-shift-tests
+  (fn
+    ()
+    (forth-p4-check-top "1 0 LSHIFT" "1 0 LSHIFT" 1)
+    (forth-p4-check-top "1 1 LSHIFT" "1 1 LSHIFT" 2)
+    (forth-p4-check-top "1 2 LSHIFT" "1 2 LSHIFT" 4)
+    (forth-p4-check-top "1 15 LSHIFT" "1 15 LSHIFT" 32768)
+    (forth-p4-check-top "1 31 LSHIFT" "1 31 LSHIFT" -2147483648)
+    (forth-p4-check-top "1 0 RSHIFT" "1 0 RSHIFT" 1)
+    (forth-p4-check-top "1 1 RSHIFT" "1 1 RSHIFT" 0)
+    (forth-p4-check-top "2 1 RSHIFT" "2 1 RSHIFT" 1)
+    (forth-p4-check-top "4 2 RSHIFT" "4 2 RSHIFT" 1)
+    (forth-p4-check-top "-1 1 RSHIFT (logical, not arithmetic)" "-1 1 RSHIFT" 2147483647)
+    (forth-p4-check-top "MSB via 1S 1 RSHIFT INVERT" "0 INVERT 1 RSHIFT INVERT" -2147483648)))
+
+(define
+  forth-p4-sp-tests
+  (fn
+    ()
+    (forth-p4-check-top "SP@ returns depth (0)" "SP@" 0)
+    (forth-p4-check-top
+      "SP@ after pushes"
+      "1 2 3 SP@ SWAP DROP SWAP DROP SWAP DROP"
+      3)
+    (forth-p4-check-stack-size
+      "SP! truncates"
+      "1 2 3 4 5 2 SP!"
+      2)
+    (forth-p4-check-top
+      "SP! leaves base items intact"
+      "1 2 3 4 5 2 SP!"
+      2)))
+
+(define
+  forth-p4-base-tests
+  (fn
+    ()
+    (forth-p4-check-top
+      "BASE default is 10"
+      "BASE @"
+      10)
+    (forth-p4-check-top
+      "HEX switches base to 16"
+      "HEX BASE @"
+      16)
+    (forth-p4-check-top
+      "DECIMAL resets to 10"
+      "HEX DECIMAL BASE @"
+      10)
+    (forth-p4-check-top
+      "HEX parses 10 as 16"
+      "HEX 10"
+      16)
+    (forth-p4-check-top
+      "HEX parses FF as 255"
+      "HEX FF"
+      255)
+    (forth-p4-check-top
+      "DECIMAL parses 10 as 10"
+      "HEX DECIMAL 10"
+      10)
+    (forth-p4-check-top
+      "OCTAL parses 17 as 15"
+      "OCTAL 17"
+      15)
+    (forth-p4-check-top
+      "BASE @ ; 16 BASE ! ; BASE @"
+      "BASE @ 16 BASE ! BASE @ SWAP DROP"
+      16)))
+
+(define
+  forth-p4-run-all
+  (fn
+    ()
+    (set! forth-p4-passed 0)
+    (set! forth-p4-failed 0)
+    (set! forth-p4-failures (list))
+    (forth-p4-string-tests)
+    (forth-p4-count-tests)
+    (forth-p4-fill-tests)
+    (forth-p4-cmove-tests)
+    (forth-p4-charplus-tests)
+    (forth-p4-char-tests)
+    (forth-p4-key-accept-tests)
+    (forth-p4-base-tests)
+    (forth-p4-shift-tests)
+    (forth-p4-sp-tests)
+    (dict
+      "passed"
+      forth-p4-passed
+      "failed"
+      forth-p4-failed
+      "failures"
+      forth-p4-failures)))

lib/forth/tests/test-phase5.sx

@@ -0,0 +1,333 @@
+;; Phase 5 — Core Extension + memory primitives.
+
+(define forth-p5-passed 0)
+(define forth-p5-failed 0)
+(define forth-p5-failures (list))
+
+(define
+  forth-p5-assert
+  (fn
+    (label expected actual)
+    (if
+      (= expected actual)
+      (set! forth-p5-passed (+ forth-p5-passed 1))
+      (begin
+        (set! forth-p5-failed (+ forth-p5-failed 1))
+        (set!
+          forth-p5-failures
+          (concat
+            forth-p5-failures
+            (list
+              (str label ": expected " (str expected) " got " (str actual)))))))))
+
+(define
+  forth-p5-check-stack
+  (fn
+    (label src expected)
+    (let ((r (forth-run src))) (forth-p5-assert label expected (nth r 2)))))
+
+(define
+  forth-p5-check-top
+  (fn
+    (label src expected)
+    (let
+      ((r (forth-run src)))
+      (let
+        ((stk (nth r 2)))
+        (forth-p5-assert label expected (nth stk (- (len stk) 1)))))))
+
+(define
+  forth-p5-create-tests
+  (fn
+    ()
+    (forth-p5-check-top
+      "CREATE pushes HERE-at-creation"
+      "HERE CREATE FOO FOO ="
+      -1)
+    (forth-p5-check-top
+      "CREATE + ALLOT advances HERE"
+      "HERE 5 ALLOT HERE SWAP -"
+      5)
+    (forth-p5-check-top
+      "CREATE + , stores cell"
+      "CREATE FOO 42 , FOO @"
+      42)
+    (forth-p5-check-stack
+      "CREATE multiple ,"
+      "CREATE TBL 1 , 2 , 3 , TBL @ TBL CELL+ @ TBL CELL+ CELL+ @"
+      (list 1 2 3))
+    (forth-p5-check-top
+      "C, stores byte"
+      "CREATE B 65 C, 66 C, B C@"
+      65)))
+
+(define
+  forth-p5-unsigned-tests
+  (fn
+    ()
+    (forth-p5-check-top "1 2 U<" "1 2 U<" -1)
+    (forth-p5-check-top "2 1 U<" "2 1 U<" 0)
+    (forth-p5-check-top "0 1 U<" "0 1 U<" -1)
+    (forth-p5-check-top "-1 1 U< (since -1 unsigned is huge)" "-1 1 U<" 0)
+    (forth-p5-check-top "1 -1 U<" "1 -1 U<" -1)
+    (forth-p5-check-top "1 2 U>" "1 2 U>" 0)
+    (forth-p5-check-top "-1 1 U>" "-1 1 U>" -1)))
+
+(define
+  forth-p5-2bang-tests
+  (fn
+    ()
+    (forth-p5-check-stack
+      "2! / 2@"
+      "CREATE X 0 , 0 , 11 22 X 2! X 2@"
+      (list 11 22))))
+
+(define
+  forth-p5-mixed-tests
+  (fn
+    ()
+    (forth-p5-check-stack "S>D positive" "5 S>D" (list 5 0))
+    (forth-p5-check-stack "S>D negative" "-5 S>D" (list -5 -1))
+    (forth-p5-check-stack "S>D zero" "0 S>D" (list 0 0))
+    (forth-p5-check-top "D>S keeps low" "5 0 D>S" 5)
+    (forth-p5-check-stack "M* small positive" "3 4 M*" (list 12 0))
+    (forth-p5-check-stack "M* negative" "-3 4 M*" (list -12 -1))
+    (forth-p5-check-stack
+      "M* negative * negative"
+      "-3 -4 M*"
+      (list 12 0))
+    (forth-p5-check-stack "UM* small" "3 4 UM*" (list 12 0))
+    (forth-p5-check-stack
+      "UM/MOD: 100 0 / 5"
+      "100 0 5 UM/MOD"
+      (list 0 20))
+    (forth-p5-check-stack
+      "FM/MOD: -7 / 2 floored"
+      "-7 -1 2 FM/MOD"
+      (list 1 -4))
+    (forth-p5-check-stack
+      "SM/REM: -7 / 2 truncated"
+      "-7 -1 2 SM/REM"
+      (list -1 -3))
+    (forth-p5-check-top "*/ truncated" "7 11 13 */" 5)
+    (forth-p5-check-stack "*/MOD" "7 11 13 */MOD" (list 12 5))))
+
+(define
+  forth-p5-double-tests
+  (fn
+    ()
+    (forth-p5-check-stack "D+ small" "5 0 7 0 D+" (list 12 0))
+    (forth-p5-check-stack "D+ negative" "-5 -1 -3 -1 D+" (list -8 -1))
+    (forth-p5-check-stack "D- small" "10 0 3 0 D-" (list 7 0))
+    (forth-p5-check-stack "DNEGATE positive" "5 0 DNEGATE" (list -5 -1))
+    (forth-p5-check-stack "DNEGATE negative" "-5 -1 DNEGATE" (list 5 0))
+    (forth-p5-check-stack "DABS negative" "-7 -1 DABS" (list 7 0))
+    (forth-p5-check-stack "DABS positive" "7 0 DABS" (list 7 0))
+    (forth-p5-check-top "D= equal" "5 0 5 0 D=" -1)
+    (forth-p5-check-top "D= unequal lo" "5 0 7 0 D=" 0)
+    (forth-p5-check-top "D= unequal hi" "5 0 5 1 D=" 0)
+    (forth-p5-check-top "D< lt" "5 0 7 0 D<" -1)
+    (forth-p5-check-top "D< gt" "7 0 5 0 D<" 0)
+    (forth-p5-check-top "D0= zero" "0 0 D0=" -1)
+    (forth-p5-check-top "D0= nonzero" "5 0 D0=" 0)
+    (forth-p5-check-top "D0< neg" "-5 -1 D0<" -1)
+    (forth-p5-check-top "D0< pos" "5 0 D0<" 0)
+    (forth-p5-check-stack "DMAX" "5 0 7 0 DMAX" (list 7 0))
+    (forth-p5-check-stack "DMIN" "5 0 7 0 DMIN" (list 5 0))))
+
+(define
+  forth-p5-format-tests
+  (fn
+    ()
+    (forth-p4-check-output-passthrough
+      "U. prints with trailing space"
+      "123 U."
+      "123 ")
+    (forth-p4-check-output-passthrough
+      "<# #S #> TYPE — decimal"
+      "123 0 <# #S #> TYPE"
+      "123")
+    (forth-p4-check-output-passthrough
+      "<# #S #> TYPE — hex"
+      "255 HEX 0 <# #S #> TYPE"
+      "FF")
+    (forth-p4-check-output-passthrough
+      "<# # # #> partial"
+      "1234 0 <# # # #> TYPE"
+      "34")
+    (forth-p4-check-output-passthrough
+      "SIGN holds minus"
+      "<# -1 SIGN -1 SIGN 0 0 #> TYPE"
+      "--")
+    (forth-p4-check-output-passthrough
+      ".R right-justifies"
+      "42 5 .R"
+      "   42")
+    (forth-p4-check-output-passthrough
+      ".R negative"
+      "-42 5 .R"
+      "  -42")
+    (forth-p4-check-output-passthrough
+      "U.R"
+      "42 5 U.R"
+      "   42")
+    (forth-p4-check-output-passthrough
+      "HOLD char"
+      "<# 0 0 65 HOLD #> TYPE"
+      "A")))
+
+(define
+  forth-p5-dict-tests
+  (fn
+    ()
+    (forth-p5-check-top
+      "EXECUTE via tick"
+      ": INC 1+ ; 9 ' INC EXECUTE"
+      10)
+    (forth-p5-check-top
+      "['] inside def"
+      ": DUB 2* ; : APPLY ['] DUB EXECUTE ; 5 APPLY"
+      10)
+    (forth-p5-check-top
+      ">BODY of CREATE word"
+      "CREATE C 99 , ' C >BODY @"
+      99)
+    (forth-p5-check-stack
+      "WORD parses next token to counted-string"
+      ": A 5 ; BL WORD A COUNT TYPE"
+      (list))
+    (forth-p5-check-top
+      "FIND on known word -> non-zero"
+      ": A 5 ; BL WORD A FIND SWAP DROP"
+      -1)))
+
+(define
+  forth-p5-state-tests
+  (fn
+    ()
+    (forth-p5-check-top
+      "STATE @ in interpret mode"
+      "STATE @"
+      0)
+    (forth-p5-check-top
+      "STATE @ via IMMEDIATE inside compile"
+      ": GT8 STATE @ ; IMMEDIATE : T GT8 LITERAL ; T"
+      -1)
+    (forth-p5-check-top
+      "[ ] LITERAL captures"
+      ": SEVEN [ 7 ] LITERAL ; SEVEN"
+      7)
+    (forth-p5-check-top
+      "EVALUATE in interpret mode"
+      "S\" 5 7 +\" EVALUATE"
+      12)
+    (forth-p5-check-top
+      "EVALUATE inside def"
+      ": A 100 ; : B S\" A\" EVALUATE ; B"
+      100)))
+
+(define
+  forth-p5-misc-tests
+  (fn
+    ()
+    (forth-p5-check-top "WITHIN inclusive lower" "3 2 10 WITHIN" -1)
+    (forth-p5-check-top "WITHIN exclusive upper" "10 2 10 WITHIN" 0)
+    (forth-p5-check-top "WITHIN below range" "1 2 10 WITHIN" 0)
+    (forth-p5-check-top "WITHIN at lower" "2 2 10 WITHIN" -1)
+    (forth-p5-check-top
+      "EXIT leaves colon-def early"
+      ": F 5 EXIT 99 ; F"
+      5)
+    (forth-p5-check-stack
+      "EXIT in IF branch"
+      ": F 5 0 IF DROP 99 EXIT THEN ; F"
+      (list 5))
+    (forth-p5-check-top
+      "UNLOOP + EXIT in DO"
+      ": SUM 0 10 0 DO I 5 = IF I UNLOOP EXIT THEN LOOP ; SUM"
+      5)))
+
+(define
+  forth-p5-fa-tests
+  (fn
+    ()
+    (forth-p5-check-top
+      "R/O R/W W/O constants"
+      "R/O R/W W/O + +"
+      3)
+    (forth-p5-check-top
+      "CREATE-FILE returns ior=0"
+      "CREATE PAD 50 ALLOT PAD S\" /tmp/test.fxf\" ROT SWAP CMOVE S\" /tmp/test.fxf\" R/W CREATE-FILE SWAP DROP"
+      0)
+    (forth-p5-check-top
+      "WRITE-FILE then CLOSE"
+      "S\" /tmp/t2.fxf\" R/W CREATE-FILE DROP >R S\" HI\" R@ WRITE-FILE R> CLOSE-FILE +"
+      0)
+    (forth-p5-check-top
+      "OPEN-FILE on unknown path returns ior!=0"
+      "S\" /tmp/nope.fxf\" R/O OPEN-FILE SWAP DROP 0 ="
+      0)))
+
+(define
+  forth-p5-string-tests
+  (fn
+    ()
+    (forth-p5-check-top "COMPARE equal" "S\" ABC\" S\" ABC\" COMPARE" 0)
+    (forth-p5-check-top "COMPARE less" "S\" ABC\" S\" ABD\" COMPARE" -1)
+    (forth-p5-check-top "COMPARE greater" "S\" ABD\" S\" ABC\" COMPARE" 1)
+    (forth-p5-check-top
+      "COMPARE prefix less"
+      "S\" AB\" S\" ABC\" COMPARE"
+      -1)
+    (forth-p5-check-top
+      "COMPARE prefix greater"
+      "S\" ABC\" S\" AB\" COMPARE"
+      1)
+    (forth-p5-check-top
+      "SEARCH found flag"
+      "S\" HELLO WORLD\" S\" WORLD\" SEARCH"
+      -1)
+    (forth-p5-check-top
+      "SEARCH not found flag"
+      "S\" HELLO\" S\" XYZ\" SEARCH"
+      0)
+    (forth-p5-check-top
+      "SEARCH empty needle flag"
+      "S\" HELLO\" S\" \" SEARCH"
+      -1)
+    (forth-p5-check-top
+      "SLITERAL via [ S\" ... \" ]"
+      ": A [ S\" HI\" ] SLITERAL ; A SWAP DROP"
+      2)))
+
+(define
+  forth-p4-check-output-passthrough
+  (fn
+    (label src expected)
+    (let ((r (forth-run src))) (forth-p5-assert label expected (nth r 1)))))
+
+(define
+  forth-p5-run-all
+  (fn
+    ()
+    (set! forth-p5-passed 0)
+    (set! forth-p5-failed 0)
+    (set! forth-p5-failures (list))
+    (forth-p5-create-tests)
+    (forth-p5-unsigned-tests)
+    (forth-p5-2bang-tests)
+    (forth-p5-mixed-tests)
+    (forth-p5-double-tests)
+    (forth-p5-format-tests)
+    (forth-p5-dict-tests)
+    (forth-p5-state-tests)
+    (forth-p5-misc-tests)
+    (forth-p5-fa-tests)
+    (forth-p5-string-tests)
+    (dict
+      "passed"
+      forth-p5-passed
+      "failed"
+      forth-p5-failed
+      "failures"
+      forth-p5-failures)))

lib/prolog/conformance.sh

@@ -1,108 +0,0 @@
-#!/usr/bin/env bash
-# Run every Prolog test suite via sx_server and refresh scoreboard.{json,md}.
-# Exit 0 if all green, 1 if any failures.
-set -euo pipefail
-
-HERE="$(cd "$(dirname "$0")" && pwd)"
-ROOT="$(cd "$HERE/../.." && pwd)"
-SX="${SX_SERVER:-/root/rose-ash/hosts/ocaml/_build/default/bin/sx_server.exe}"
-
-if [[ ! -x "$SX" ]]; then
-  echo "sx_server not found at $SX (set SX_SERVER env to override)" >&2
-  exit 2
-fi
-
-cd "$ROOT"
-
-# name : test-file : runner-fn
-SUITES=(
-  "parse:lib/prolog/tests/parse.sx:pl-parse-tests-run!"
-  "unify:lib/prolog/tests/unify.sx:pl-unify-tests-run!"
-  "clausedb:lib/prolog/tests/clausedb.sx:pl-clausedb-tests-run!"
-  "solve:lib/prolog/tests/solve.sx:pl-solve-tests-run!"
-  "operators:lib/prolog/tests/operators.sx:pl-operators-tests-run!"
-  "dynamic:lib/prolog/tests/dynamic.sx:pl-dynamic-tests-run!"
-  "append:lib/prolog/tests/programs/append.sx:pl-append-tests-run!"
-  "reverse:lib/prolog/tests/programs/reverse.sx:pl-reverse-tests-run!"
-  "member:lib/prolog/tests/programs/member.sx:pl-member-tests-run!"
-  "nqueens:lib/prolog/tests/programs/nqueens.sx:pl-nqueens-tests-run!"
-  "family:lib/prolog/tests/programs/family.sx:pl-family-tests-run!"
-)
-
-SCRIPT='(epoch 1)
-(load "lib/prolog/tokenizer.sx")
-(load "lib/prolog/parser.sx")
-(load "lib/prolog/runtime.sx")'
-for entry in "${SUITES[@]}"; do
-  IFS=: read -r _ file _ <<< "$entry"
-  SCRIPT+=$'\n(load "'"$file"$'")'
-done
-for entry in "${SUITES[@]}"; do
-  IFS=: read -r _ _ fn <<< "$entry"
-  SCRIPT+=$'\n(eval "('"$fn"$')")'
-done
-
-OUTPUT="$(printf '%s\n' "$SCRIPT" | "$SX" 2>&1)"
-
-mapfile -t LINES < <(printf '%s\n' "$OUTPUT" | grep -E '^\{:failed')
-
-if [[ ${#LINES[@]} -ne ${#SUITES[@]} ]]; then
-  echo "Expected ${#SUITES[@]} suite results, got ${#LINES[@]}" >&2
-  echo "---- raw output ----" >&2
-  printf '%s\n' "$OUTPUT" >&2
-  exit 3
-fi
-
-TOTAL_PASS=0
-TOTAL_FAIL=0
-TOTAL=0
-JSON_SUITES=""
-MD_ROWS=""
-
-for i in "${!SUITES[@]}"; do
-  IFS=: read -r name _ _ <<< "${SUITES[$i]}"
-  line="${LINES[$i]}"
-  passed=$(grep -oE ':passed [0-9]+' <<< "$line" | grep -oE '[0-9]+')
-  total=$(grep -oE ':total [0-9]+' <<< "$line" | grep -oE '[0-9]+')
-  failed=$(grep -oE ':failed [0-9]+' <<< "$line" | grep -oE '[0-9]+')
-  TOTAL_PASS=$((TOTAL_PASS + passed))
-  TOTAL_FAIL=$((TOTAL_FAIL + failed))
-  TOTAL=$((TOTAL + total))
-  status="ok"
-  [[ "$failed" -gt 0 ]] && status="FAIL"
-  [[ -n "$JSON_SUITES" ]] && JSON_SUITES+=","
-  JSON_SUITES+="\"$name\":{\"passed\":$passed,\"total\":$total,\"failed\":$failed}"
-  MD_ROWS+="| $name | $passed | $total | $status |"$'\n'
-done
-
-WHEN="$(date -Iseconds 2>/dev/null || date)"
-
-cat > "$HERE/scoreboard.json" <<JSON
-{
-  "total_passed": $TOTAL_PASS,
-  "total_failed": $TOTAL_FAIL,
-  "total": $TOTAL,
-  "suites": {$JSON_SUITES},
-  "generated": "$WHEN"
-}
-JSON
-
-cat > "$HERE/scoreboard.md" <<MD
-# Prolog scoreboard
-
-**$TOTAL_PASS / $TOTAL passing** ($TOTAL_FAIL failure(s)).
-Generated $WHEN.
-
-| Suite | Passed | Total | Status |
-|-------|--------|-------|--------|
-$MD_ROWS
-Run \`bash lib/prolog/conformance.sh\` to refresh. Override the binary
-with \`SX_SERVER=path/to/sx_server.exe bash …\`.
-MD
-
-if [[ "$TOTAL_FAIL" -gt 0 ]]; then
-  echo "$TOTAL_FAIL failure(s) across $TOTAL tests" >&2
-  exit 1
-fi
-
-echo "All $TOTAL tests pass."

lib/prolog/parser.sx

@@ -1,20 +1,28 @@
 ;; lib/prolog/parser.sx — tokens → Prolog AST
 ;;
-;; Phase 4 grammar (with operator table):
+;; Phase 1 grammar (NO operator table yet):
 ;;   Program      := Clause* EOF
-;;   Clause       := Term[999] "." | Term[999] ":-" Term[1200] "."
-;;   Term[Pmax] uses precedence climbing on the operator table:
-;;     primary = Atom | Var | Number | String | Compound | List | "(" Term[1200] ")"
-;;     while next token is infix op `op` with prec(op) ≤ Pmax:
-;;       consume op; parse rhs at right-prec(op); fold into compound(op-name,[lhs,rhs])
+;;   Clause       := Term "." | Term ":-" Term "."
+;;   Term         := Atom | Var | Number | String | Compound | List
+;;   Compound     := atom "(" ArgList ")"
+;;   ArgList      := Term ("," Term)*
+;;   List         := "[" "]" | "[" Term ("," Term)* ("|" Term)? "]"
 ;;
-;; Op type → right-prec for op at precedence P:
-;;   xfx → P-1   strict-both
-;;   xfy → P     right-associative
-;;   yfx → P-1   left-associative
+;; Term AST shapes (all tagged lists for uniform dispatch):
+;;   ("atom"  name)            — atom
+;;   ("var"   name)            — variable template (parser-time only)
+;;   ("num"   value)           — integer or float
+;;   ("str"   value)           — string literal
+;;   ("compound" functor args) — compound term, args is list of term-ASTs
+;;   ("cut")                   — the cut atom !
 ;;
-;; AST shapes are unchanged — operators just become compound terms.
+;; A clause is (list "clause" head body). A fact is head with body = ("atom" "true").
+;;
+;; The empty list is (atom "[]"). Cons is compound "." with two args:
+;; [1, 2, 3]   → .(1, .(2, .(3, [])))
+;; [H|T]       → .(H, T)
 
+;; ── Parser state helpers ────────────────────────────────────────────
 (define
   pp-peek
   (fn
@@ -58,6 +66,7 @@
             (if (= (get t :value) nil) "" (get t :value))
             "'"))))))
 
+;; ── AST constructors ────────────────────────────────────────────────
 (define pl-mk-atom (fn (name) (list "atom" name)))
 (define pl-mk-var (fn (name) (list "var" name)))
 (define pl-mk-num (fn (n) (list "num" n)))
@@ -65,14 +74,18 @@
 (define pl-mk-compound (fn (f args) (list "compound" f args)))
 (define pl-mk-cut (fn () (list "cut")))
 
+;; Term tag extractors
 (define pl-term-tag (fn (t) (if (list? t) (first t) nil)))
 (define pl-term-val (fn (t) (nth t 1)))
 (define pl-compound-functor (fn (t) (nth t 1)))
 (define pl-compound-args (fn (t) (nth t 2)))
 
+;; Empty-list atom and cons helpers
 (define pl-nil-term (fn () (pl-mk-atom "[]")))
+
 (define pl-mk-cons (fn (h t) (pl-mk-compound "." (list h t))))
 
+;; Build cons list from a list of terms + optional tail
 (define
   pl-mk-list-term
   (fn
@@ -82,60 +95,9 @@
       tail
       (pl-mk-cons (first items) (pl-mk-list-term (rest items) tail)))))
 
-;; ── Operator table (Phase 4) ──────────────────────────────────────
-;; Each entry: (name precedence type). Type ∈ "xfx" "xfy" "yfx".
-(define
-  pl-op-table
-  (list
-    (list "," 1000 "xfy")
-    (list ";" 1100 "xfy")
-    (list "->" 1050 "xfy")
-    (list "=" 700 "xfx")
-    (list "\\=" 700 "xfx")
-    (list "is" 700 "xfx")
-    (list "<" 700 "xfx")
-    (list ">" 700 "xfx")
-    (list "=<" 700 "xfx")
-    (list ">=" 700 "xfx")
-    (list "+" 500 "yfx")
-    (list "-" 500 "yfx")
-    (list "*" 400 "yfx")
-    (list "/" 400 "yfx")
-    (list "mod" 400 "yfx")))
-
-(define
-  pl-op-find
-  (fn
-    (name table)
-    (cond
-      ((empty? table) nil)
-      ((= (first (first table)) name) (rest (first table)))
-      (true (pl-op-find name (rest table))))))
-
-(define pl-op-lookup (fn (name) (pl-op-find name pl-op-table)))
-
-;; Token → (name prec type) for known infix ops, else nil.
-(define
-  pl-token-op
-  (fn
-    (t)
-    (let
-      ((ty (get t :type)) (vv (get t :value)))
-      (cond
-        ((and (= ty "punct") (= vv ","))
-          (let
-            ((info (pl-op-lookup ",")))
-            (if (nil? info) nil (cons "," info))))
-        ((= ty "atom")
-          (let
-            ((info (pl-op-lookup vv)))
-            (if (nil? info) nil (cons vv info))))
-        (true nil)))))
-
 ;; ── Term parser ─────────────────────────────────────────────────────
-;; Primary term: atom, var, num, str, compound (atom + paren), list, cut, parens.
 (define
-  pp-parse-primary
+  pp-parse-term
   (fn
     (st)
     (let
@@ -149,12 +111,6 @@
           ((and (= ty "op") (= vv "!"))
             (do (pp-advance! st) (pl-mk-cut)))
           ((and (= ty "punct") (= vv "[")) (pp-parse-list st))
-          ((and (= ty "punct") (= vv "("))
-            (do
-              (pp-advance! st)
-              (let
-                ((inner (pp-parse-term-prec st 1200)))
-                (do (pp-expect! st "punct" ")") inner))))
           ((= ty "atom")
             (do
               (pp-advance! st)
@@ -177,51 +133,13 @@
                 (if (= vv nil) "" vv)
                 "'"))))))))
 
-;; Operator-aware term parser: precedence climbing.
-(define
-  pp-parse-term-prec
-  (fn
-    (st max-prec)
-    (let ((left (pp-parse-primary st))) (pp-parse-op-rhs st left max-prec))))
-
-(define
-  pp-parse-op-rhs
-  (fn
-    (st left max-prec)
-    (let
-      ((op-info (pl-token-op (pp-peek st))))
-      (cond
-        ((nil? op-info) left)
-        (true
-          (let
-            ((name (first op-info))
-              (prec (nth op-info 1))
-              (ty (nth op-info 2)))
-            (cond
-              ((> prec max-prec) left)
-              (true
-                (let
-                  ((right-prec (if (= ty "xfy") prec (- prec 1))))
-                  (do
-                    (pp-advance! st)
-                    (let
-                      ((right (pp-parse-term-prec st right-prec)))
-                      (pp-parse-op-rhs
-                        st
-                        (pl-mk-compound name (list left right))
-                        max-prec))))))))))))
-
-;; Backwards-compat alias.
-(define pp-parse-term (fn (st) (pp-parse-term-prec st 999)))
-
-;; Args inside parens: parse at prec 999 so comma-as-operator (1000)
-;; is not consumed; the explicit comma loop handles separation.
+;; Parse one or more comma-separated terms (arguments).
 (define
   pp-parse-arg-list
   (fn
     (st)
     (let
-      ((first-arg (pp-parse-term-prec st 999)) (args (list)))
+      ((first-arg (pp-parse-term st)) (args (list)))
       (do
         (append! args first-arg)
         (define
@@ -232,12 +150,12 @@
               (pp-at? st "punct" ",")
               (do
                 (pp-advance! st)
-                (append! args (pp-parse-term-prec st 999))
+                (append! args (pp-parse-term st))
                 (loop)))))
         (loop)
         args))))
 
-;; List literal.
+;; Parse a [ ... ] list literal. Consumes the "[".
 (define
   pp-parse-list
   (fn
@@ -250,7 +168,7 @@
         (let
           ((items (list)))
           (do
-            (append! items (pp-parse-term-prec st 999))
+            (append! items (pp-parse-term st))
             (define
               comma-loop
               (fn
@@ -259,17 +177,52 @@
                   (pp-at? st "punct" ",")
                   (do
                     (pp-advance! st)
-                    (append! items (pp-parse-term-prec st 999))
+                    (append! items (pp-parse-term st))
                     (comma-loop)))))
             (comma-loop)
             (let
-              ((tail (if (pp-at? st "punct" "|") (do (pp-advance! st) (pp-parse-term-prec st 999)) (pl-nil-term))))
+              ((tail (if (pp-at? st "punct" "|") (do (pp-advance! st) (pp-parse-term st)) (pl-nil-term))))
               (do (pp-expect! st "punct" "]") (pl-mk-list-term items tail)))))))))
 
 ;; ── Body parsing ────────────────────────────────────────────────────
-;; A body is a single term parsed at prec 1200 — operator parser folds
-;; `,`, `;`, `->` automatically into right-associative compounds.
-(define pp-parse-body (fn (st) (pp-parse-term-prec st 1200)))
+;; A clause body is a comma-separated list of goals. We flatten into a
+;; right-associative `,` compound: (A, B, C) → ','(A, ','(B, C))
+;; If only one goal, it's that goal directly.
+(define
+  pp-parse-body
+  (fn
+    (st)
+    (let
+      ((first-goal (pp-parse-term st)) (rest-goals (list)))
+      (do
+        (define
+          gloop
+          (fn
+            ()
+            (when
+              (pp-at? st "punct" ",")
+              (do
+                (pp-advance! st)
+                (append! rest-goals (pp-parse-term st))
+                (gloop)))))
+        (gloop)
+        (if
+          (= (len rest-goals) 0)
+          first-goal
+          (pp-build-conj first-goal rest-goals))))))
+
+(define
+  pp-build-conj
+  (fn
+    (first-goal rest-goals)
+    (if
+      (= (len rest-goals) 0)
+      first-goal
+      (pl-mk-compound
+        ","
+        (list
+          first-goal
+          (pp-build-conj (first rest-goals) (rest rest-goals)))))))
 
 ;; ── Clause parsing ──────────────────────────────────────────────────
 (define
@@ -277,11 +230,12 @@
   (fn
     (st)
     (let
-      ((head (pp-parse-term-prec st 999)))
+      ((head (pp-parse-term st)))
       (let
         ((body (if (pp-at? st "op" ":-") (do (pp-advance! st) (pp-parse-body st)) (pl-mk-atom "true"))))
         (do (pp-expect! st "punct" ".") (list "clause" head body))))))
 
+;; Parse an entire program — returns list of clauses.
 (define
   pl-parse-program
   (fn
@@ -299,9 +253,13 @@
         (ploop)
         clauses))))
 
+;; Parse a single query term (no trailing "."). Returns the term.
 (define
   pl-parse-query
   (fn (tokens) (let ((st {:idx 0 :tokens tokens})) (pp-parse-body st))))
 
+;; Convenience: source → clauses
 (define pl-parse (fn (src) (pl-parse-program (pl-tokenize src))))
+
+;; Convenience: source → query term
 (define pl-parse-goal (fn (src) (pl-parse-query (pl-tokenize src))))

lib/prolog/runtime.sx

@@ -98,11 +98,6 @@
             "compound"
             fun
             (map (fn (a) (pl-instantiate a var-env)) args))))
-      ((= (first ast) "clause")
-        (let
-          ((h (pl-instantiate (nth ast 1) var-env))
-            (b (pl-instantiate (nth ast 2) var-env)))
-          (list "clause" h b)))
       (true ast))))
 
 (define pl-instantiate-fresh (fn (ast) (pl-instantiate ast {})))
@@ -235,499 +230,3 @@
         (pl-unify! t1 t2 trail)
         true
         (do (pl-trail-undo-to! trail mark) false)))))
-
-(define pl-mk-db (fn () {:clauses {}}))
-
-(define
-  pl-head-key
-  (fn
-    (head)
-    (cond
-      ((pl-compound? head) (str (pl-fun head) "/" (len (pl-args head))))
-      ((pl-atom? head) (str (pl-atom-name head) "/0"))
-      (true (error "pl-head-key: invalid head")))))
-
-(define pl-clause-key (fn (clause) (pl-head-key (nth clause 1))))
-
-(define pl-goal-key (fn (goal) (pl-head-key goal)))
-
-(define
-  pl-db-add!
-  (fn
-    (db clause)
-    (let
-      ((key (pl-clause-key clause)) (table (dict-get db :clauses)))
-      (cond
-        ((nil? (dict-get table key)) (dict-set! table key (list clause)))
-        (true (begin (append! (dict-get table key) clause) nil))))))
-
-(define
-  pl-db-load!
-  (fn
-    (db program)
-    (cond
-      ((empty? program) nil)
-      (true
-        (begin
-          (pl-db-add! db (first program))
-          (pl-db-load! db (rest program)))))))
-
-(define
-  pl-db-lookup
-  (fn
-    (db key)
-    (let
-      ((v (dict-get (dict-get db :clauses) key)))
-      (cond ((nil? v) (list)) (true v)))))
-
-(define
-  pl-db-lookup-goal
-  (fn (db goal) (pl-db-lookup db (pl-goal-key goal))))
-
-(define
-  pl-rt-walk-to-ast
-  (fn
-    (w)
-    (cond
-      ((pl-var? w) (list "var" (str "_G" (pl-var-id w))))
-      ((and (list? w) (not (empty? w)) (= (first w) "compound"))
-        (list "compound" (nth w 1) (map pl-rt-walk-to-ast (nth w 2))))
-      (true w))))
-
-(define pl-rt-to-ast (fn (t) (pl-rt-walk-to-ast (pl-walk-deep t))))
-
-(define
-  pl-build-clause
-  (fn
-    (ast)
-    (cond
-      ((and (list? ast) (= (first ast) "compound") (= (nth ast 1) ":-") (= (len (nth ast 2)) 2))
-        (list "clause" (first (nth ast 2)) (nth (nth ast 2) 1)))
-      (true (list "clause" ast (list "atom" "true"))))))
-
-(define
-  pl-db-prepend!
-  (fn
-    (db clause)
-    (let
-      ((key (pl-clause-key clause)) (table (dict-get db :clauses)))
-      (cond
-        ((nil? (dict-get table key)) (dict-set! table key (list clause)))
-        (true (dict-set! table key (cons clause (dict-get table key))))))))
-
-(define
-  pl-list-without
-  (fn
-    (lst i)
-    (cond
-      ((empty? lst) (list))
-      ((= i 0) (rest lst))
-      (true (cons (first lst) (pl-list-without (rest lst) (- i 1)))))))
-
-(define
-  pl-solve-assertz!
-  (fn
-    (db term k)
-    (begin (pl-db-add! db (pl-build-clause (pl-rt-to-ast term))) (k))))
-
-(define
-  pl-solve-asserta!
-  (fn
-    (db term k)
-    (begin (pl-db-prepend! db (pl-build-clause (pl-rt-to-ast term))) (k))))
-
-(define
-  pl-solve-retract!
-  (fn
-    (db term trail k)
-    (let
-      ((head-runtime (cond ((and (pl-compound? term) (= (pl-fun term) ":-") (= (len (pl-args term)) 2)) (first (pl-args term))) (true term)))
-        (body-runtime
-          (cond
-            ((and (pl-compound? term) (= (pl-fun term) ":-") (= (len (pl-args term)) 2))
-              (nth (pl-args term) 1))
-            (true (list "atom" "true")))))
-      (let
-        ((wh (pl-walk head-runtime)))
-        (cond
-          ((pl-var? wh) false)
-          (true
-            (let
-              ((key (pl-head-key wh)))
-              (pl-retract-try-each
-                db
-                key
-                (pl-db-lookup db key)
-                head-runtime
-                body-runtime
-                0
-                trail
-                k))))))))
-
-(define
-  pl-retract-try-each
-  (fn
-    (db key remaining head-rt body-rt idx trail k)
-    (cond
-      ((empty? remaining) false)
-      (true
-        (let
-          ((mark (pl-trail-mark trail))
-            (cl (pl-instantiate-fresh (first remaining))))
-          (cond
-            ((and (pl-unify! head-rt (nth cl 1) trail) (pl-unify! body-rt (nth cl 2) trail))
-              (begin
-                (let
-                  ((all (pl-db-lookup db key)))
-                  (dict-set!
-                    (dict-get db :clauses)
-                    key
-                    (pl-list-without all idx)))
-                (let
-                  ((r (k)))
-                  (cond
-                    (r true)
-                    (true (begin (pl-trail-undo-to! trail mark) false))))))
-            (true
-              (begin
-                (pl-trail-undo-to! trail mark)
-                (pl-retract-try-each
-                  db
-                  key
-                  (rest remaining)
-                  head-rt
-                  body-rt
-                  (+ idx 1)
-                  trail
-                  k)))))))))
-
-(define
-  pl-cut?
-  (fn (t) (and (list? t) (not (empty? t)) (= (first t) "cut"))))
-
-(define
-  pl-solve!
-  (fn
-    (db goal trail cut-box k)
-    (let
-      ((g (pl-walk goal)))
-      (cond
-        ((pl-var? g) false)
-        ((pl-cut? g) (begin (dict-set! cut-box :cut true) (k)))
-        ((and (pl-atom? g) (= (pl-atom-name g) "true")) (k))
-        ((and (pl-atom? g) (= (pl-atom-name g) "fail")) false)
-        ((and (pl-atom? g) (= (pl-atom-name g) "nl"))
-          (begin (pl-output-write! "\n") (k)))
-        ((and (pl-compound? g) (= (pl-fun g) "=") (= (len (pl-args g)) 2))
-          (pl-solve-eq! (first (pl-args g)) (nth (pl-args g) 1) trail k))
-        ((and (pl-compound? g) (= (pl-fun g) "\\=") (= (len (pl-args g)) 2))
-          (pl-solve-not-eq!
-            (first (pl-args g))
-            (nth (pl-args g) 1)
-            trail
-            k))
-        ((and (pl-compound? g) (= (pl-fun g) "is") (= (len (pl-args g)) 2))
-          (pl-solve-eq!
-            (first (pl-args g))
-            (list "num" (pl-eval-arith (nth (pl-args g) 1)))
-            trail
-            k))
-        ((and (pl-compound? g) (= (pl-fun g) "<") (= (len (pl-args g)) 2))
-          (cond
-            ((< (pl-eval-arith (first (pl-args g))) (pl-eval-arith (nth (pl-args g) 1)))
-              (k))
-            (true false)))
-        ((and (pl-compound? g) (= (pl-fun g) ">") (= (len (pl-args g)) 2))
-          (cond
-            ((> (pl-eval-arith (first (pl-args g))) (pl-eval-arith (nth (pl-args g) 1)))
-              (k))
-            (true false)))
-        ((and (pl-compound? g) (= (pl-fun g) "=<") (= (len (pl-args g)) 2))
-          (cond
-            ((<= (pl-eval-arith (first (pl-args g))) (pl-eval-arith (nth (pl-args g) 1)))
-              (k))
-            (true false)))
-        ((and (pl-compound? g) (= (pl-fun g) ">=") (= (len (pl-args g)) 2))
-          (cond
-            ((>= (pl-eval-arith (first (pl-args g))) (pl-eval-arith (nth (pl-args g) 1)))
-              (k))
-            (true false)))
-        ((and (pl-compound? g) (= (pl-fun g) ",") (= (len (pl-args g)) 2))
-          (pl-solve!
-            db
-            (first (pl-args g))
-            trail
-            cut-box
-            (fn () (pl-solve! db (nth (pl-args g) 1) trail cut-box k))))
-        ((and (pl-compound? g) (= (pl-fun g) ";") (= (len (pl-args g)) 2))
-          (pl-solve-or!
-            db
-            (first (pl-args g))
-            (nth (pl-args g) 1)
-            trail
-            cut-box
-            k))
-        ((and (pl-compound? g) (= (pl-fun g) "->") (= (len (pl-args g)) 2))
-          (pl-solve-if-then-else!
-            db
-            (first (pl-args g))
-            (nth (pl-args g) 1)
-            (list "atom" "fail")
-            trail
-            cut-box
-            k))
-        ((and (pl-compound? g) (= (pl-fun g) "call") (= (len (pl-args g)) 1))
-          (let
-            ((call-cb {:cut false}))
-            (pl-solve! db (first (pl-args g)) trail call-cb k)))
-        ((and (pl-compound? g) (= (pl-fun g) "write") (= (len (pl-args g)) 1))
-          (begin
-            (pl-output-write! (pl-format-term (first (pl-args g))))
-            (k)))
-        ((and (pl-compound? g) (= (pl-fun g) "assertz") (= (len (pl-args g)) 1))
-          (pl-solve-assertz! db (first (pl-args g)) k))
-        ((and (pl-compound? g) (= (pl-fun g) "assert") (= (len (pl-args g)) 1))
-          (pl-solve-assertz! db (first (pl-args g)) k))
-        ((and (pl-compound? g) (= (pl-fun g) "asserta") (= (len (pl-args g)) 1))
-          (pl-solve-asserta! db (first (pl-args g)) k))
-        ((and (pl-compound? g) (= (pl-fun g) "retract") (= (len (pl-args g)) 1))
-          (pl-solve-retract! db (first (pl-args g)) trail k))
-        (true (pl-solve-user! db g trail cut-box k))))))
-
-(define
-  pl-solve-or!
-  (fn
-    (db a b trail cut-box k)
-    (cond
-      ((and (pl-compound? a) (= (pl-fun a) "->") (= (len (pl-args a)) 2))
-        (pl-solve-if-then-else!
-          db
-          (first (pl-args a))
-          (nth (pl-args a) 1)
-          b
-          trail
-          cut-box
-          k))
-      (true
-        (let
-          ((mark (pl-trail-mark trail)))
-          (let
-            ((r (pl-solve! db a trail cut-box k)))
-            (cond
-              (r true)
-              ((dict-get cut-box :cut) false)
-              (true
-                (begin
-                  (pl-trail-undo-to! trail mark)
-                  (pl-solve! db b trail cut-box k))))))))))
-
-(define
-  pl-solve-if-then-else!
-  (fn
-    (db cond-goal then-goal else-goal trail cut-box k)
-    (let
-      ((mark (pl-trail-mark trail)))
-      (let
-        ((local-cb {:cut false}))
-        (let
-          ((found {:val false}))
-          (pl-solve!
-            db
-            cond-goal
-            trail
-            local-cb
-            (fn () (begin (dict-set! found :val true) true)))
-          (cond
-            ((dict-get found :val) (pl-solve! db then-goal trail cut-box k))
-            (true
-              (begin
-                (pl-trail-undo-to! trail mark)
-                (pl-solve! db else-goal trail cut-box k)))))))))
-
-(define pl-output-buffer "")
-
-(define pl-output-clear! (fn () (set! pl-output-buffer "")))
-
-(define
-  pl-output-write!
-  (fn (s) (begin (set! pl-output-buffer (str pl-output-buffer s)) nil)))
-
-(define
-  pl-format-args
-  (fn
-    (args)
-    (cond
-      ((empty? args) "")
-      ((= (len args) 1) (pl-format-term (first args)))
-      (true
-        (str
-          (pl-format-term (first args))
-          ", "
-          (pl-format-args (rest args)))))))
-
-(define
-  pl-format-term
-  (fn
-    (t)
-    (let
-      ((w (pl-walk-deep t)))
-      (cond
-        ((pl-var? w) (str "_" (pl-var-id w)))
-        ((pl-atom? w) (pl-atom-name w))
-        ((pl-num? w) (str (pl-num-val w)))
-        ((pl-str? w) (pl-str-val w))
-        ((pl-compound? w)
-          (str (pl-fun w) "(" (pl-format-args (pl-args w)) ")"))
-        (true (str w))))))
-
-(define
-  pl-eval-arith
-  (fn
-    (t)
-    (let
-      ((w (pl-walk-deep t)))
-      (cond
-        ((pl-num? w) (pl-num-val w))
-        ((pl-compound? w)
-          (let
-            ((f (pl-fun w)) (args (pl-args w)))
-            (cond
-              ((and (= f "+") (= (len args) 2))
-                (+
-                  (pl-eval-arith (first args))
-                  (pl-eval-arith (nth args 1))))
-              ((and (= f "-") (= (len args) 2))
-                (-
-                  (pl-eval-arith (first args))
-                  (pl-eval-arith (nth args 1))))
-              ((and (= f "-") (= (len args) 1))
-                (- 0 (pl-eval-arith (first args))))
-              ((and (= f "*") (= (len args) 2))
-                (*
-                  (pl-eval-arith (first args))
-                  (pl-eval-arith (nth args 1))))
-              ((and (= f "/") (= (len args) 2))
-                (/
-                  (pl-eval-arith (first args))
-                  (pl-eval-arith (nth args 1))))
-              ((and (= f "mod") (= (len args) 2))
-                (mod
-                  (pl-eval-arith (first args))
-                  (pl-eval-arith (nth args 1))))
-              ((and (= f "abs") (= (len args) 1))
-                (let
-                  ((v (pl-eval-arith (first args))))
-                  (cond ((< v 0) (- 0 v)) (true v))))
-              (true 0))))
-        (true 0)))))
-
-(define
-  pl-solve-not-eq!
-  (fn
-    (a b trail k)
-    (let
-      ((mark (pl-trail-mark trail)))
-      (let
-        ((unified (pl-unify! a b trail)))
-        (begin
-          (pl-trail-undo-to! trail mark)
-          (cond (unified false) (true (k))))))))
-
-(define
-  pl-solve-eq!
-  (fn
-    (a b trail k)
-    (let
-      ((mark (pl-trail-mark trail)))
-      (cond
-        ((pl-unify! a b trail)
-          (let
-            ((r (k)))
-            (cond
-              (r true)
-              (true (begin (pl-trail-undo-to! trail mark) false)))))
-        (true (begin (pl-trail-undo-to! trail mark) false))))))
-
-(define
-  pl-solve-user!
-  (fn
-    (db goal trail outer-cut-box k)
-    (let
-      ((inner-cut-box {:cut false}))
-      (let
-        ((outer-was-cut (dict-get outer-cut-box :cut)))
-        (pl-try-clauses!
-          db
-          goal
-          trail
-          (pl-db-lookup-goal db goal)
-          outer-cut-box
-          outer-was-cut
-          inner-cut-box
-          k)))))
-
-(define
-  pl-try-clauses!
-  (fn
-    (db goal trail clauses outer-cut-box outer-was-cut inner-cut-box k)
-    (cond
-      ((empty? clauses) false)
-      (true
-        (let
-          ((mark (pl-trail-mark trail)))
-          (let
-            ((clause (pl-instantiate-fresh (first clauses))))
-            (let
-              ((head (nth clause 1)) (body (nth clause 2)))
-              (cond
-                ((pl-unify! goal head trail)
-                  (let
-                    ((r (pl-solve! db body trail inner-cut-box k)))
-                    (cond
-                      (r true)
-                      ((dict-get inner-cut-box :cut)
-                        (begin (pl-trail-undo-to! trail mark) false))
-                      ((and (not outer-was-cut) (dict-get outer-cut-box :cut))
-                        (begin (pl-trail-undo-to! trail mark) false))
-                      (true
-                        (begin
-                          (pl-trail-undo-to! trail mark)
-                          (pl-try-clauses!
-                            db
-                            goal
-                            trail
-                            (rest clauses)
-                            outer-cut-box
-                            outer-was-cut
-                            inner-cut-box
-                            k))))))
-                (true
-                  (begin
-                    (pl-trail-undo-to! trail mark)
-                    (pl-try-clauses!
-                      db
-                      goal
-                      trail
-                      (rest clauses)
-                      outer-cut-box
-                      outer-was-cut
-                      inner-cut-box
-                      k)))))))))))
-
-(define
-  pl-solve-once!
-  (fn (db goal trail) (pl-solve! db goal trail {:cut false} (fn () true))))
-
-(define
-  pl-solve-count!
-  (fn
-    (db goal trail)
-    (let
-      ((box {:n 0}))
-      (pl-solve!
-        db
-        goal
-        trail
-        {:cut false}
-        (fn () (begin (dict-set! box :n (+ (dict-get box :n) 1)) false)))
-      (dict-get box :n))))

lib/prolog/scoreboard.json

@@ -1,7 +0,0 @@
-{
-  "total_passed": 213,
-  "total_failed": 0,
-  "total": 213,
-  "suites": {"parse":{"passed":25,"total":25,"failed":0},"unify":{"passed":47,"total":47,"failed":0},"clausedb":{"passed":14,"total":14,"failed":0},"solve":{"passed":62,"total":62,"failed":0},"operators":{"passed":19,"total":19,"failed":0},"dynamic":{"passed":11,"total":11,"failed":0},"append":{"passed":6,"total":6,"failed":0},"reverse":{"passed":6,"total":6,"failed":0},"member":{"passed":7,"total":7,"failed":0},"nqueens":{"passed":6,"total":6,"failed":0},"family":{"passed":10,"total":10,"failed":0}},
-  "generated": "2026-04-25T07:31:46+00:00"
-}

lib/prolog/scoreboard.md

@@ -1,21 +0,0 @@
-# Prolog scoreboard
-
-**213 / 213 passing** (0 failure(s)).
-Generated 2026-04-25T07:31:46+00:00.
-
-| Suite | Passed | Total | Status |
-|-------|--------|-------|--------|
-| parse | 25 | 25 | ok |
-| unify | 47 | 47 | ok |
-| clausedb | 14 | 14 | ok |
-| solve | 62 | 62 | ok |
-| operators | 19 | 19 | ok |
-| dynamic | 11 | 11 | ok |
-| append | 6 | 6 | ok |
-| reverse | 6 | 6 | ok |
-| member | 7 | 7 | ok |
-| nqueens | 6 | 6 | ok |
-| family | 10 | 10 | ok |
-
-Run `bash lib/prolog/conformance.sh` to refresh. Override the binary
-with `SX_SERVER=path/to/sx_server.exe bash …`.

lib/prolog/tests/clausedb.sx

@@ -1,99 +0,0 @@
-;; lib/prolog/tests/clausedb.sx — Clause DB unit tests
-
-(define pl-db-test-count 0)
-(define pl-db-test-pass 0)
-(define pl-db-test-fail 0)
-(define pl-db-test-failures (list))
-
-(define
-  pl-db-test!
-  (fn
-    (name got expected)
-    (begin
-      (set! pl-db-test-count (+ pl-db-test-count 1))
-      (if
-        (= got expected)
-        (set! pl-db-test-pass (+ pl-db-test-pass 1))
-        (begin
-          (set! pl-db-test-fail (+ pl-db-test-fail 1))
-          (append!
-            pl-db-test-failures
-            (str name "\n    expected: " expected "\n    got:      " got)))))))
-
-(pl-db-test!
-  "head-key atom arity 0"
-  (pl-head-key (nth (first (pl-parse "foo.")) 1))
-  "foo/0")
-
-(pl-db-test!
-  "head-key compound arity 2"
-  (pl-head-key (nth (first (pl-parse "bar(a, b).")) 1))
-  "bar/2")
-
-(pl-db-test!
-  "clause-key of :- clause"
-  (pl-clause-key (first (pl-parse "likes(mary, X) :- friendly(X).")))
-  "likes/2")
-
-(pl-db-test!
-  "empty db lookup returns empty list"
-  (len (pl-db-lookup (pl-mk-db) "parent/2"))
-  0)
-
-(define pl-db-t1 (pl-mk-db))
-(pl-db-load! pl-db-t1 (pl-parse "foo(a). foo(b). foo(c)."))
-
-(pl-db-test!
-  "three facts same functor"
-  (len (pl-db-lookup pl-db-t1 "foo/1"))
-  3)
-(pl-db-test!
-  "mismatching key returns empty"
-  (len (pl-db-lookup pl-db-t1 "foo/2"))
-  0)
-
-(pl-db-test!
-  "first clause has arg a"
-  (pl-atom-name
-    (first (pl-args (nth (first (pl-db-lookup pl-db-t1 "foo/1")) 1))))
-  "a")
-
-(pl-db-test!
-  "third clause has arg c"
-  (pl-atom-name
-    (first (pl-args (nth (nth (pl-db-lookup pl-db-t1 "foo/1") 2) 1))))
-  "c")
-
-(define pl-db-t2 (pl-mk-db))
-(pl-db-load! pl-db-t2 (pl-parse "foo. bar. foo. parent(a, b). parent(c, d)."))
-
-(pl-db-test!
-  "atom heads keyed as foo/0"
-  (len (pl-db-lookup pl-db-t2 "foo/0"))
-  2)
-(pl-db-test!
-  "atom heads keyed as bar/0"
-  (len (pl-db-lookup pl-db-t2 "bar/0"))
-  1)
-(pl-db-test!
-  "compound heads keyed as parent/2"
-  (len (pl-db-lookup pl-db-t2 "parent/2"))
-  2)
-
-(pl-db-test!
-  "lookup-goal extracts functor/arity"
-  (len
-    (pl-db-lookup-goal pl-db-t2 (nth (first (pl-parse "parent(X, Y).")) 1)))
-  2)
-
-(pl-db-test!
-  "lookup-goal on atom goal"
-  (len (pl-db-lookup-goal pl-db-t2 (nth (first (pl-parse "foo.")) 1)))
-  2)
-
-(pl-db-test!
-  "stored clause is clause form"
-  (first (first (pl-db-lookup pl-db-t2 "parent/2")))
-  "clause")
-
-(define pl-clausedb-tests-run! (fn () {:failed pl-db-test-fail :passed pl-db-test-pass :total pl-db-test-count :failures pl-db-test-failures}))

lib/prolog/tests/dynamic.sx

@@ -1,158 +0,0 @@
-;; lib/prolog/tests/dynamic.sx — assert/asserta/assertz/retract.
-
-(define pl-dy-test-count 0)
-(define pl-dy-test-pass 0)
-(define pl-dy-test-fail 0)
-(define pl-dy-test-failures (list))
-
-(define
-  pl-dy-test!
-  (fn
-    (name got expected)
-    (begin
-      (set! pl-dy-test-count (+ pl-dy-test-count 1))
-      (if
-        (= got expected)
-        (set! pl-dy-test-pass (+ pl-dy-test-pass 1))
-        (begin
-          (set! pl-dy-test-fail (+ pl-dy-test-fail 1))
-          (append!
-            pl-dy-test-failures
-            (str name "\n    expected: " expected "\n    got:      " got)))))))
-
-(define
-  pl-dy-goal
-  (fn
-    (src env)
-    (pl-instantiate (nth (first (pl-parse (str "g :- " src "."))) 2) env)))
-
-;; assertz then query
-(define pl-dy-db1 (pl-mk-db))
-(pl-solve-once!
-  pl-dy-db1
-  (pl-dy-goal "assertz(foo(1))" {})
-  (pl-mk-trail))
-(pl-dy-test!
-  "assertz(foo(1)) + foo(1)"
-  (pl-solve-once! pl-dy-db1 (pl-dy-goal "foo(1)" {}) (pl-mk-trail))
-  true)
-
-(pl-dy-test!
-  "after one assertz, foo/1 has 1 clause"
-  (pl-solve-count! pl-dy-db1 (pl-dy-goal "foo(X)" {}) (pl-mk-trail))
-  1)
-
-;; assertz appends — order preserved
-(define pl-dy-db2 (pl-mk-db))
-(pl-solve-once!
-  pl-dy-db2
-  (pl-dy-goal "assertz(p(1))" {})
-  (pl-mk-trail))
-(pl-solve-once!
-  pl-dy-db2
-  (pl-dy-goal "assertz(p(2))" {})
-  (pl-mk-trail))
-(pl-dy-test!
-  "assertz twice — count 2"
-  (pl-solve-count! pl-dy-db2 (pl-dy-goal "p(X)" {}) (pl-mk-trail))
-  2)
-
-(define pl-dy-env-a {})
-(pl-solve-once! pl-dy-db2 (pl-dy-goal "p(X)" pl-dy-env-a) (pl-mk-trail))
-(pl-dy-test!
-  "assertz: first solution is the first asserted (1)"
-  (pl-num-val (pl-walk-deep (dict-get pl-dy-env-a "X")))
-  1)
-
-;; asserta prepends
-(define pl-dy-db3 (pl-mk-db))
-(pl-solve-once!
-  pl-dy-db3
-  (pl-dy-goal "assertz(p(1))" {})
-  (pl-mk-trail))
-(pl-solve-once!
-  pl-dy-db3
-  (pl-dy-goal "asserta(p(99))" {})
-  (pl-mk-trail))
-(define pl-dy-env-b {})
-(pl-solve-once! pl-dy-db3 (pl-dy-goal "p(X)" pl-dy-env-b) (pl-mk-trail))
-(pl-dy-test!
-  "asserta: prepended clause is first solution"
-  (pl-num-val (pl-walk-deep (dict-get pl-dy-env-b "X")))
-  99)
-
-;; assert/1 = assertz/1
-(define pl-dy-db4 (pl-mk-db))
-(pl-solve-once!
-  pl-dy-db4
-  (pl-dy-goal "assert(g(7))" {})
-  (pl-mk-trail))
-(pl-dy-test!
-  "assert/1 alias"
-  (pl-solve-once! pl-dy-db4 (pl-dy-goal "g(7)" {}) (pl-mk-trail))
-  true)
-
-;; retract removes a fact
-(define pl-dy-db5 (pl-mk-db))
-(pl-solve-once!
-  pl-dy-db5
-  (pl-dy-goal "assertz(q(1))" {})
-  (pl-mk-trail))
-(pl-solve-once!
-  pl-dy-db5
-  (pl-dy-goal "assertz(q(2))" {})
-  (pl-mk-trail))
-(pl-solve-once!
-  pl-dy-db5
-  (pl-dy-goal "assertz(q(3))" {})
-  (pl-mk-trail))
-(pl-dy-test!
-  "before retract: 3 clauses"
-  (pl-solve-count! pl-dy-db5 (pl-dy-goal "q(X)" {}) (pl-mk-trail))
-  3)
-(pl-solve-once!
-  pl-dy-db5
-  (pl-dy-goal "retract(q(2))" {})
-  (pl-mk-trail))
-(pl-dy-test!
-  "after retract(q(2)): 2 clauses left"
-  (pl-solve-count! pl-dy-db5 (pl-dy-goal "q(X)" {}) (pl-mk-trail))
-  2)
-
-(define pl-dy-env-c {})
-(pl-solve-once! pl-dy-db5 (pl-dy-goal "q(X)" pl-dy-env-c) (pl-mk-trail))
-(pl-dy-test!
-  "after retract(q(2)): first remaining is 1"
-  (pl-num-val (pl-walk-deep (dict-get pl-dy-env-c "X")))
-  1)
-
-;; retract of non-existent
-(pl-dy-test!
-  "retract(missing(0)) on empty db fails"
-  (pl-solve-once!
-    (pl-mk-db)
-    (pl-dy-goal "retract(missing(0))" {})
-    (pl-mk-trail))
-  false)
-
-;; retract with unbound var matches first
-(define pl-dy-db6 (pl-mk-db))
-(pl-solve-once!
-  pl-dy-db6
-  (pl-dy-goal "assertz(r(11))" {})
-  (pl-mk-trail))
-(pl-solve-once!
-  pl-dy-db6
-  (pl-dy-goal "assertz(r(22))" {})
-  (pl-mk-trail))
-(define pl-dy-env-d {})
-(pl-solve-once!
-  pl-dy-db6
-  (pl-dy-goal "retract(r(X))" pl-dy-env-d)
-  (pl-mk-trail))
-(pl-dy-test!
-  "retract(r(X)) binds X to first match"
-  (pl-num-val (pl-walk-deep (dict-get pl-dy-env-d "X")))
-  11)
-
-(define pl-dynamic-tests-run! (fn () {:failed pl-dy-test-fail :passed pl-dy-test-pass :total pl-dy-test-count :failures pl-dy-test-failures}))

lib/prolog/tests/operators.sx

@@ -1,193 +0,0 @@
-;; lib/prolog/tests/operators.sx — operator-table parsing + comparison built-ins.
-
-(define pl-op-test-count 0)
-(define pl-op-test-pass 0)
-(define pl-op-test-fail 0)
-(define pl-op-test-failures (list))
-
-(define
-  pl-op-test!
-  (fn
-    (name got expected)
-    (begin
-      (set! pl-op-test-count (+ pl-op-test-count 1))
-      (if
-        (= got expected)
-        (set! pl-op-test-pass (+ pl-op-test-pass 1))
-        (begin
-          (set! pl-op-test-fail (+ pl-op-test-fail 1))
-          (append!
-            pl-op-test-failures
-            (str name "\n    expected: " expected "\n    got:      " got)))))))
-
-(define pl-op-empty-db (pl-mk-db))
-
-(define
-  pl-op-body
-  (fn (src) (nth (first (pl-parse (str "g :- " src "."))) 2)))
-
-(define pl-op-goal (fn (src env) (pl-instantiate (pl-op-body src) env)))
-
-;; ── parsing tests ──
-
-(pl-op-test!
-  "infix +"
-  (pl-op-body "a + b")
-  (list "compound" "+" (list (list "atom" "a") (list "atom" "b"))))
-
-(pl-op-test!
-  "infix * tighter than +"
-  (pl-op-body "a + b * c")
-  (list
-    "compound"
-    "+"
-    (list
-      (list "atom" "a")
-      (list "compound" "*" (list (list "atom" "b") (list "atom" "c"))))))
-
-(pl-op-test!
-  "parens override precedence"
-  (pl-op-body "(a + b) * c")
-  (list
-    "compound"
-    "*"
-    (list
-      (list "compound" "+" (list (list "atom" "a") (list "atom" "b")))
-      (list "atom" "c"))))
-
-(pl-op-test!
-  "+ is yfx (left-assoc)"
-  (pl-op-body "a + b + c")
-  (list
-    "compound"
-    "+"
-    (list
-      (list "compound" "+" (list (list "atom" "a") (list "atom" "b")))
-      (list "atom" "c"))))
-
-(pl-op-test!
-  "; is xfy (right-assoc)"
-  (pl-op-body "a ; b ; c")
-  (list
-    "compound"
-    ";"
-    (list
-      (list "atom" "a")
-      (list "compound" ";" (list (list "atom" "b") (list "atom" "c"))))))
-
-(pl-op-test!
-  "= folds at 700"
-  (pl-op-body "X = 5")
-  (list "compound" "=" (list (list "var" "X") (list "num" 5))))
-
-(pl-op-test!
-  "is + nests via 700>500>400"
-  (pl-op-body "X is 2 + 3 * 4")
-  (list
-    "compound"
-    "is"
-    (list
-      (list "var" "X")
-      (list
-        "compound"
-        "+"
-        (list
-          (list "num" 2)
-          (list "compound" "*" (list (list "num" 3) (list "num" 4))))))))
-
-(pl-op-test!
-  "< parses at 700"
-  (pl-op-body "2 < 3")
-  (list "compound" "<" (list (list "num" 2) (list "num" 3))))
-
-(pl-op-test!
-  "mod parses as yfx 400"
-  (pl-op-body "10 mod 3")
-  (list "compound" "mod" (list (list "num" 10) (list "num" 3))))
-
-(pl-op-test!
-  "comma in body folds right-assoc"
-  (pl-op-body "a, b, c")
-  (list
-    "compound"
-    ","
-    (list
-      (list "atom" "a")
-      (list "compound" "," (list (list "atom" "b") (list "atom" "c"))))))
-
-;; ── solver tests via infix ──
-
-(pl-op-test!
-  "X is 2 + 3 binds X = 5"
-  (let
-    ((env {}) (trail (pl-mk-trail)))
-    (begin
-      (pl-solve-once! pl-op-empty-db (pl-op-goal "X is 2 + 3" env) trail)
-      (pl-num-val (pl-walk-deep (dict-get env "X")))))
-  5)
-
-(pl-op-test!
-  "infix conjunction parses + solves"
-  (pl-solve-once!
-    pl-op-empty-db
-    (pl-op-goal "X = 5, X = 5" {})
-    (pl-mk-trail))
-  true)
-
-(pl-op-test!
-  "infix mismatch fails"
-  (pl-solve-once!
-    pl-op-empty-db
-    (pl-op-goal "X = 5, X = 6" {})
-    (pl-mk-trail))
-  false)
-
-(pl-op-test!
-  "infix disjunction picks left"
-  (pl-solve-once!
-    pl-op-empty-db
-    (pl-op-goal "true ; fail" {})
-    (pl-mk-trail))
-  true)
-
-(pl-op-test!
-  "2 < 5 succeeds"
-  (pl-solve-once!
-    pl-op-empty-db
-    (pl-op-goal "2 < 5" {})
-    (pl-mk-trail))
-  true)
-
-(pl-op-test!
-  "5 < 2 fails"
-  (pl-solve-once!
-    pl-op-empty-db
-    (pl-op-goal "5 < 2" {})
-    (pl-mk-trail))
-  false)
-
-(pl-op-test!
-  "5 >= 5 succeeds"
-  (pl-solve-once!
-    pl-op-empty-db
-    (pl-op-goal "5 >= 5" {})
-    (pl-mk-trail))
-  true)
-
-(pl-op-test!
-  "3 =< 5 succeeds"
-  (pl-solve-once!
-    pl-op-empty-db
-    (pl-op-goal "3 =< 5" {})
-    (pl-mk-trail))
-  true)
-
-(pl-op-test!
-  "infix < with arithmetic both sides"
-  (pl-solve-once!
-    pl-op-empty-db
-    (pl-op-goal "1 + 2 < 2 * 3" {})
-    (pl-mk-trail))
-  true)
-
-(define pl-operators-tests-run! (fn () {:failed pl-op-test-fail :passed pl-op-test-pass :total pl-op-test-count :failures pl-op-test-failures}))

lib/prolog/tests/programs/append.pl

@@ -1,5 +0,0 @@
-%% append/3 — list concatenation, classic Prolog
-%% Two clauses: empty-prefix base case + recursive cons-prefix.
-%% Bidirectional — works in all modes: build, check, split.
-append([], L, L).
-append([H|T], L, [H|R]) :- append(T, L, R).

lib/prolog/tests/programs/append.sx

@@ -1,114 +0,0 @@
-;; lib/prolog/tests/programs/append.sx — append/3 test runner
-;;
-;; Mirrors the Prolog source in append.pl (embedded as a string here because
-;; the SX runtime has no file-read primitive yet).
-
-(define pl-ap-test-count 0)
-(define pl-ap-test-pass 0)
-(define pl-ap-test-fail 0)
-(define pl-ap-test-failures (list))
-
-(define
-  pl-ap-test!
-  (fn
-    (name got expected)
-    (begin
-      (set! pl-ap-test-count (+ pl-ap-test-count 1))
-      (if
-        (= got expected)
-        (set! pl-ap-test-pass (+ pl-ap-test-pass 1))
-        (begin
-          (set! pl-ap-test-fail (+ pl-ap-test-fail 1))
-          (append!
-            pl-ap-test-failures
-            (str name "\n    expected: " expected "\n    got:      " got)))))))
-
-(define
-  pl-ap-term-to-sx
-  (fn
-    (t)
-    (cond
-      ((pl-num? t) (pl-num-val t))
-      ((pl-atom? t) (pl-atom-name t))
-      (true (list :complex)))))
-
-(define
-  pl-ap-list-walked
-  (fn
-    (w)
-    (cond
-      ((and (pl-atom? w) (= (pl-atom-name w) "[]")) (list))
-      ((and (pl-compound? w) (= (pl-fun w) ".") (= (len (pl-args w)) 2))
-        (cons
-          (pl-ap-term-to-sx (first (pl-args w)))
-          (pl-ap-list-walked (nth (pl-args w) 1))))
-      (true (list :not-list)))))
-
-(define pl-ap-list-to-sx (fn (t) (pl-ap-list-walked (pl-walk-deep t))))
-
-(define
-  pl-ap-goal
-  (fn
-    (src env)
-    (pl-instantiate (nth (first (pl-parse (str "g :- " src "."))) 2) env)))
-
-(define
-  pl-ap-prog-src
-  "append([], L, L). append([H|T], L, [H|R]) :- append(T, L, R).")
-
-(define pl-ap-db (pl-mk-db))
-
-(pl-db-load! pl-ap-db (pl-parse pl-ap-prog-src))
-
-(define pl-ap-env-1 {})
-(define pl-ap-goal-1 (pl-ap-goal "append([], [a, b], X)" pl-ap-env-1))
-(pl-solve-once! pl-ap-db pl-ap-goal-1 (pl-mk-trail))
-
-(pl-ap-test!
-  "append([], [a, b], X) → X = [a, b]"
-  (pl-ap-list-to-sx (dict-get pl-ap-env-1 "X"))
-  (list "a" "b"))
-
-(define pl-ap-env-2 {})
-(define pl-ap-goal-2 (pl-ap-goal "append([1, 2], [3, 4], X)" pl-ap-env-2))
-(pl-solve-once! pl-ap-db pl-ap-goal-2 (pl-mk-trail))
-
-(pl-ap-test!
-  "append([1, 2], [3, 4], X) → X = [1, 2, 3, 4]"
-  (pl-ap-list-to-sx (dict-get pl-ap-env-2 "X"))
-  (list 1 2 3 4))
-
-(pl-ap-test!
-  "append([1], [2, 3], [1, 2, 3]) succeeds"
-  (pl-solve-once!
-    pl-ap-db
-    (pl-ap-goal "append([1], [2, 3], [1, 2, 3])" {})
-    (pl-mk-trail))
-  true)
-
-(pl-ap-test!
-  "append([1, 2], [3], [1, 2, 4]) fails"
-  (pl-solve-once!
-    pl-ap-db
-    (pl-ap-goal "append([1, 2], [3], [1, 2, 4])" {})
-    (pl-mk-trail))
-  false)
-
-(pl-ap-test!
-  "append(X, Y, [1, 2, 3]) backtracks 4 times"
-  (pl-solve-count!
-    pl-ap-db
-    (pl-ap-goal "append(X, Y, [1, 2, 3])" {})
-    (pl-mk-trail))
-  4)
-
-(define pl-ap-env-6 {})
-(define pl-ap-goal-6 (pl-ap-goal "append(X, [3], [1, 2, 3])" pl-ap-env-6))
-(pl-solve-once! pl-ap-db pl-ap-goal-6 (pl-mk-trail))
-
-(pl-ap-test!
-  "append(X, [3], [1, 2, 3]) deduces X = [1, 2]"
-  (pl-ap-list-to-sx (dict-get pl-ap-env-6 "X"))
-  (list 1 2))
-
-(define pl-append-tests-run! (fn () {:failed pl-ap-test-fail :passed pl-ap-test-pass :total pl-ap-test-count :failures pl-ap-test-failures}))

lib/prolog/tests/programs/family.pl

@@ -1,24 +0,0 @@
-%% family — facts + transitive ancestor + derived relations.
-%% Five-generation tree: tom -> bob -> {ann, pat} -> jim, plus tom's
-%% other child liz.
-
-parent(tom, bob).
-parent(tom, liz).
-parent(bob, ann).
-parent(bob, pat).
-parent(pat, jim).
-
-male(tom).
-male(bob).
-male(jim).
-male(pat).
-female(liz).
-female(ann).
-
-father(F, C) :- parent(F, C), male(F).
-mother(M, C) :- parent(M, C), female(M).
-
-ancestor(X, Y) :- parent(X, Y).
-ancestor(X, Y) :- parent(X, Z), ancestor(Z, Y).
-
-sibling(X, Y) :- parent(P, X), parent(P, Y), \=(X, Y).

lib/prolog/tests/programs/family.sx

@@ -1,116 +0,0 @@
-;; lib/prolog/tests/programs/family.sx — facts + ancestor + sibling relations.
-
-(define pl-fa-test-count 0)
-(define pl-fa-test-pass 0)
-(define pl-fa-test-fail 0)
-(define pl-fa-test-failures (list))
-
-(define
-  pl-fa-test!
-  (fn
-    (name got expected)
-    (begin
-      (set! pl-fa-test-count (+ pl-fa-test-count 1))
-      (if
-        (= got expected)
-        (set! pl-fa-test-pass (+ pl-fa-test-pass 1))
-        (begin
-          (set! pl-fa-test-fail (+ pl-fa-test-fail 1))
-          (append!
-            pl-fa-test-failures
-            (str name "\n    expected: " expected "\n    got:      " got)))))))
-
-(define
-  pl-fa-goal
-  (fn
-    (src env)
-    (pl-instantiate (nth (first (pl-parse (str "g :- " src "."))) 2) env)))
-
-(define
-  pl-fa-prog-src
-  "parent(tom, bob). parent(tom, liz). parent(bob, ann). parent(bob, pat). parent(pat, jim). male(tom). male(bob). male(jim). male(pat). female(liz). female(ann). father(F, C) :- parent(F, C), male(F). mother(M, C) :- parent(M, C), female(M). ancestor(X, Y) :- parent(X, Y). ancestor(X, Y) :- parent(X, Z), ancestor(Z, Y). sibling(X, Y) :- parent(P, X), parent(P, Y), \\=(X, Y).")
-
-(define pl-fa-db (pl-mk-db))
-(pl-db-load! pl-fa-db (pl-parse pl-fa-prog-src))
-
-(pl-fa-test!
-  "parent(tom, bob) is a fact"
-  (pl-solve-once!
-    pl-fa-db
-    (pl-fa-goal "parent(tom, bob)" {})
-    (pl-mk-trail))
-  true)
-
-(pl-fa-test!
-  "parent(tom, ann) — not a direct parent"
-  (pl-solve-once!
-    pl-fa-db
-    (pl-fa-goal "parent(tom, ann)" {})
-    (pl-mk-trail))
-  false)
-
-(pl-fa-test!
-  "5 parent/2 facts in total"
-  (pl-solve-count!
-    pl-fa-db
-    (pl-fa-goal "parent(X, Y)" {})
-    (pl-mk-trail))
-  5)
-
-(pl-fa-test!
-  "ancestor(tom, jim) — three-step transitive"
-  (pl-solve-once!
-    pl-fa-db
-    (pl-fa-goal "ancestor(tom, jim)" {})
-    (pl-mk-trail))
-  true)
-
-(pl-fa-test!
-  "tom has 5 ancestors-of: bob, liz, ann, pat, jim"
-  (pl-solve-count!
-    pl-fa-db
-    (pl-fa-goal "ancestor(tom, X)" {})
-    (pl-mk-trail))
-  5)
-
-(pl-fa-test!
-  "father(bob, ann) succeeds"
-  (pl-solve-once!
-    pl-fa-db
-    (pl-fa-goal "father(bob, ann)" {})
-    (pl-mk-trail))
-  true)
-
-(pl-fa-test!
-  "father(liz, ann) fails (liz is female)"
-  (pl-solve-once!
-    pl-fa-db
-    (pl-fa-goal "father(liz, ann)" {})
-    (pl-mk-trail))
-  false)
-
-(pl-fa-test!
-  "mother(liz, X) fails (liz has no children)"
-  (pl-solve-once!
-    pl-fa-db
-    (pl-fa-goal "mother(liz, X)" {})
-    (pl-mk-trail))
-  false)
-
-(pl-fa-test!
-  "sibling(ann, pat) succeeds"
-  (pl-solve-once!
-    pl-fa-db
-    (pl-fa-goal "sibling(ann, pat)" {})
-    (pl-mk-trail))
-  true)
-
-(pl-fa-test!
-  "sibling(ann, ann) fails by \\="
-  (pl-solve-once!
-    pl-fa-db
-    (pl-fa-goal "sibling(ann, ann)" {})
-    (pl-mk-trail))
-  false)
-
-(define pl-family-tests-run! (fn () {:failed pl-fa-test-fail :passed pl-fa-test-pass :total pl-fa-test-count :failures pl-fa-test-failures}))

lib/prolog/tests/programs/member.pl

@@ -1,4 +0,0 @@
-%% member/2 — list membership.
-%% Generates all solutions on backtracking when the element is unbound.
-member(X, [X|_]).
-member(X, [_|T]) :- member(X, T).

lib/prolog/tests/programs/member.sx

@@ -1,91 +0,0 @@
-;; lib/prolog/tests/programs/member.sx — member/2 generator.
-
-(define pl-mb-test-count 0)
-(define pl-mb-test-pass 0)
-(define pl-mb-test-fail 0)
-(define pl-mb-test-failures (list))
-
-(define
-  pl-mb-test!
-  (fn
-    (name got expected)
-    (begin
-      (set! pl-mb-test-count (+ pl-mb-test-count 1))
-      (if
-        (= got expected)
-        (set! pl-mb-test-pass (+ pl-mb-test-pass 1))
-        (begin
-          (set! pl-mb-test-fail (+ pl-mb-test-fail 1))
-          (append!
-            pl-mb-test-failures
-            (str name "\n    expected: " expected "\n    got:      " got)))))))
-
-(define
-  pl-mb-goal
-  (fn
-    (src env)
-    (pl-instantiate (nth (first (pl-parse (str "g :- " src "."))) 2) env)))
-
-(define pl-mb-prog-src "member(X, [X|_]). member(X, [_|T]) :- member(X, T).")
-
-(define pl-mb-db (pl-mk-db))
-(pl-db-load! pl-mb-db (pl-parse pl-mb-prog-src))
-
-(pl-mb-test!
-  "member(2, [1, 2, 3]) succeeds"
-  (pl-solve-once!
-    pl-mb-db
-    (pl-mb-goal "member(2, [1, 2, 3])" {})
-    (pl-mk-trail))
-  true)
-
-(pl-mb-test!
-  "member(4, [1, 2, 3]) fails"
-  (pl-solve-once!
-    pl-mb-db
-    (pl-mb-goal "member(4, [1, 2, 3])" {})
-    (pl-mk-trail))
-  false)
-
-(pl-mb-test!
-  "member(X, []) fails"
-  (pl-solve-once!
-    pl-mb-db
-    (pl-mb-goal "member(X, [])" {})
-    (pl-mk-trail))
-  false)
-
-(pl-mb-test!
-  "member(X, [a, b, c]) generates 3 solutions"
-  (pl-solve-count!
-    pl-mb-db
-    (pl-mb-goal "member(X, [a, b, c])" {})
-    (pl-mk-trail))
-  3)
-
-(define pl-mb-env-1 {})
-(define pl-mb-goal-1 (pl-mb-goal "member(X, [11, 22, 33])" pl-mb-env-1))
-(pl-solve-once! pl-mb-db pl-mb-goal-1 (pl-mk-trail))
-
-(pl-mb-test!
-  "member(X, [11, 22, 33]) first solution X = 11"
-  (pl-num-val (pl-walk-deep (dict-get pl-mb-env-1 "X")))
-  11)
-
-(pl-mb-test!
-  "member(2, [1, 2, 3, 2, 1]) matches twice on backtrack"
-  (pl-solve-count!
-    pl-mb-db
-    (pl-mb-goal "member(2, [1, 2, 3, 2, 1])" {})
-    (pl-mk-trail))
-  2)
-
-(pl-mb-test!
-  "member with unbound list cell unifies"
-  (pl-solve-once!
-    pl-mb-db
-    (pl-mb-goal "member(a, [X, b, c])" {})
-    (pl-mk-trail))
-  true)
-
-(define pl-member-tests-run! (fn () {:failed pl-mb-test-fail :passed pl-mb-test-pass :total pl-mb-test-count :failures pl-mb-test-failures}))

lib/prolog/tests/programs/nqueens.pl

@@ -1,27 +0,0 @@
-%% nqueens — permutation-and-test formulation.
-%% Caller passes the row list [1..N]; queens/2 finds N column placements
-%% s.t. no two queens attack on a diagonal. Same-column attacks are
-%% structurally impossible — Qs is a permutation, all distinct.
-%%
-%% No `>/2` `</2` `=</2` built-ins yet, so range/3 is omitted; tests pass
-%; the literal range list. Once the operator table lands and arithmetic
-%% comparison built-ins are in, range/3 can be added.
-queens(L, Qs) :- permute(L, Qs), safe(Qs).
-
-permute([], []).
-permute(L, [H|T]) :- select(H, L, R), permute(R, T).
-
-select(X, [X|T], T).
-select(X, [H|T], [H|R]) :- select(X, T, R).
-
-safe([]).
-safe([Q|Qs]) :- safe(Qs), no_attack(Q, Qs, 1).
-
-no_attack(_, [], _).
-no_attack(Q, [Q1|Qs], D) :-
-  is(D2, +(Q, D)),
-  \=(D2, Q1),
-  is(D3, -(Q, D)),
-  \=(D3, Q1),
-  is(D1, +(D, 1)),
-  no_attack(Q, Qs, D1).

lib/prolog/tests/programs/nqueens.sx

@@ -1,108 +0,0 @@
-;; lib/prolog/tests/programs/nqueens.sx — N-queens via permute + safe.
-
-(define pl-nq-test-count 0)
-(define pl-nq-test-pass 0)
-(define pl-nq-test-fail 0)
-(define pl-nq-test-failures (list))
-
-(define
-  pl-nq-test!
-  (fn
-    (name got expected)
-    (begin
-      (set! pl-nq-test-count (+ pl-nq-test-count 1))
-      (if
-        (= got expected)
-        (set! pl-nq-test-pass (+ pl-nq-test-pass 1))
-        (begin
-          (set! pl-nq-test-fail (+ pl-nq-test-fail 1))
-          (append!
-            pl-nq-test-failures
-            (str name "\n    expected: " expected "\n    got:      " got)))))))
-
-(define
-  pl-nq-term-to-sx
-  (fn
-    (t)
-    (cond
-      ((pl-num? t) (pl-num-val t))
-      ((pl-atom? t) (pl-atom-name t))
-      (true (list :complex)))))
-
-(define
-  pl-nq-list-walked
-  (fn
-    (w)
-    (cond
-      ((and (pl-atom? w) (= (pl-atom-name w) "[]")) (list))
-      ((and (pl-compound? w) (= (pl-fun w) ".") (= (len (pl-args w)) 2))
-        (cons
-          (pl-nq-term-to-sx (first (pl-args w)))
-          (pl-nq-list-walked (nth (pl-args w) 1))))
-      (true (list :not-list)))))
-
-(define pl-nq-list-to-sx (fn (t) (pl-nq-list-walked (pl-walk-deep t))))
-
-(define
-  pl-nq-goal
-  (fn
-    (src env)
-    (pl-instantiate (nth (first (pl-parse (str "g :- " src "."))) 2) env)))
-
-(define
-  pl-nq-prog-src
-  "queens(L, Qs) :- permute(L, Qs), safe(Qs). permute([], []). permute(L, [H|T]) :- select(H, L, R), permute(R, T). select(X, [X|T], T). select(X, [H|T], [H|R]) :- select(X, T, R). safe([]). safe([Q|Qs]) :- safe(Qs), no_attack(Q, Qs, 1). no_attack(_, [], _). no_attack(Q, [Q1|Qs], D) :- is(D2, +(Q, D)), \\=(D2, Q1), is(D3, -(Q, D)), \\=(D3, Q1), is(D1, +(D, 1)), no_attack(Q, Qs, D1).")
-
-(define pl-nq-db (pl-mk-db))
-(pl-db-load! pl-nq-db (pl-parse pl-nq-prog-src))
-
-(pl-nq-test!
-  "queens([1], Qs) → 1 solution"
-  (pl-solve-count!
-    pl-nq-db
-    (pl-nq-goal "queens([1], Qs)" {})
-    (pl-mk-trail))
-  1)
-
-(pl-nq-test!
-  "queens([1, 2], Qs) → 0 solutions"
-  (pl-solve-count!
-    pl-nq-db
-    (pl-nq-goal "queens([1, 2], Qs)" {})
-    (pl-mk-trail))
-  0)
-
-(pl-nq-test!
-  "queens([1, 2, 3], Qs) → 0 solutions"
-  (pl-solve-count!
-    pl-nq-db
-    (pl-nq-goal "queens([1, 2, 3], Qs)" {})
-    (pl-mk-trail))
-  0)
-
-(pl-nq-test!
-  "queens([1, 2, 3, 4], Qs) → 2 solutions"
-  (pl-solve-count!
-    pl-nq-db
-    (pl-nq-goal "queens([1, 2, 3, 4], Qs)" {})
-    (pl-mk-trail))
-  2)
-
-(pl-nq-test!
-  "queens([1, 2, 3, 4, 5], Qs) → 10 solutions"
-  (pl-solve-count!
-    pl-nq-db
-    (pl-nq-goal "queens([1, 2, 3, 4, 5], Qs)" {})
-    (pl-mk-trail))
-  10)
-
-(define pl-nq-env-1 {})
-(define pl-nq-goal-1 (pl-nq-goal "queens([1, 2, 3, 4], Qs)" pl-nq-env-1))
-(pl-solve-once! pl-nq-db pl-nq-goal-1 (pl-mk-trail))
-
-(pl-nq-test!
-  "queens([1..4], Qs) first solution = [2, 4, 1, 3]"
-  (pl-nq-list-to-sx (dict-get pl-nq-env-1 "Qs"))
-  (list 2 4 1 3))
-
-(define pl-nqueens-tests-run! (fn () {:failed pl-nq-test-fail :passed pl-nq-test-pass :total pl-nq-test-count :failures pl-nq-test-failures}))

lib/prolog/tests/programs/reverse.pl

@@ -1,7 +0,0 @@
-%% reverse/2 — naive reverse via append/3.
-%% Quadratic — accumulates the reversed prefix one append per cons.
-reverse([], []).
-reverse([H|T], R) :- reverse(T, RT), append(RT, [H], R).
-
-append([], L, L).
-append([H|T], L, [H|R]) :- append(T, L, R).

lib/prolog/tests/programs/reverse.sx

@@ -1,113 +0,0 @@
-;; lib/prolog/tests/programs/reverse.sx — naive reverse/2 via append/3.
-;;
-;; Mirrors reverse.pl (embedded as a string here).
-
-(define pl-rv-test-count 0)
-(define pl-rv-test-pass 0)
-(define pl-rv-test-fail 0)
-(define pl-rv-test-failures (list))
-
-(define
-  pl-rv-test!
-  (fn
-    (name got expected)
-    (begin
-      (set! pl-rv-test-count (+ pl-rv-test-count 1))
-      (if
-        (= got expected)
-        (set! pl-rv-test-pass (+ pl-rv-test-pass 1))
-        (begin
-          (set! pl-rv-test-fail (+ pl-rv-test-fail 1))
-          (append!
-            pl-rv-test-failures
-            (str name "\n    expected: " expected "\n    got:      " got)))))))
-
-(define
-  pl-rv-term-to-sx
-  (fn
-    (t)
-    (cond
-      ((pl-num? t) (pl-num-val t))
-      ((pl-atom? t) (pl-atom-name t))
-      (true (list :complex)))))
-
-(define
-  pl-rv-list-walked
-  (fn
-    (w)
-    (cond
-      ((and (pl-atom? w) (= (pl-atom-name w) "[]")) (list))
-      ((and (pl-compound? w) (= (pl-fun w) ".") (= (len (pl-args w)) 2))
-        (cons
-          (pl-rv-term-to-sx (first (pl-args w)))
-          (pl-rv-list-walked (nth (pl-args w) 1))))
-      (true (list :not-list)))))
-
-(define pl-rv-list-to-sx (fn (t) (pl-rv-list-walked (pl-walk-deep t))))
-
-(define
-  pl-rv-goal
-  (fn
-    (src env)
-    (pl-instantiate (nth (first (pl-parse (str "g :- " src "."))) 2) env)))
-
-(define
-  pl-rv-prog-src
-  "reverse([], []). reverse([H|T], R) :- reverse(T, RT), append(RT, [H], R). append([], L, L). append([H|T], L, [H|R]) :- append(T, L, R).")
-
-(define pl-rv-db (pl-mk-db))
-(pl-db-load! pl-rv-db (pl-parse pl-rv-prog-src))
-
-(define pl-rv-env-1 {})
-(define pl-rv-goal-1 (pl-rv-goal "reverse([], X)" pl-rv-env-1))
-(pl-solve-once! pl-rv-db pl-rv-goal-1 (pl-mk-trail))
-
-(pl-rv-test!
-  "reverse([], X) → X = []"
-  (pl-rv-list-to-sx (dict-get pl-rv-env-1 "X"))
-  (list))
-
-(define pl-rv-env-2 {})
-(define pl-rv-goal-2 (pl-rv-goal "reverse([1], X)" pl-rv-env-2))
-(pl-solve-once! pl-rv-db pl-rv-goal-2 (pl-mk-trail))
-
-(pl-rv-test!
-  "reverse([1], X) → X = [1]"
-  (pl-rv-list-to-sx (dict-get pl-rv-env-2 "X"))
-  (list 1))
-
-(define pl-rv-env-3 {})
-(define pl-rv-goal-3 (pl-rv-goal "reverse([1, 2, 3], X)" pl-rv-env-3))
-(pl-solve-once! pl-rv-db pl-rv-goal-3 (pl-mk-trail))
-
-(pl-rv-test!
-  "reverse([1, 2, 3], X) → X = [3, 2, 1]"
-  (pl-rv-list-to-sx (dict-get pl-rv-env-3 "X"))
-  (list 3 2 1))
-
-(define pl-rv-env-4 {})
-(define pl-rv-goal-4 (pl-rv-goal "reverse([a, b, c, d], X)" pl-rv-env-4))
-(pl-solve-once! pl-rv-db pl-rv-goal-4 (pl-mk-trail))
-
-(pl-rv-test!
-  "reverse([a, b, c, d], X) → X = [d, c, b, a]"
-  (pl-rv-list-to-sx (dict-get pl-rv-env-4 "X"))
-  (list "d" "c" "b" "a"))
-
-(pl-rv-test!
-  "reverse([1, 2, 3], [3, 2, 1]) succeeds"
-  (pl-solve-once!
-    pl-rv-db
-    (pl-rv-goal "reverse([1, 2, 3], [3, 2, 1])" {})
-    (pl-mk-trail))
-  true)
-
-(pl-rv-test!
-  "reverse([1, 2], [1, 2]) fails"
-  (pl-solve-once!
-    pl-rv-db
-    (pl-rv-goal "reverse([1, 2], [1, 2])" {})
-    (pl-mk-trail))
-  false)
-
-(define pl-reverse-tests-run! (fn () {:failed pl-rv-test-fail :passed pl-rv-test-pass :total pl-rv-test-count :failures pl-rv-test-failures}))

lib/prolog/tests/solve.sx

@@ -1,618 +0,0 @@
-;; lib/prolog/tests/solve.sx — DFS solver unit tests
-
-(define pl-s-test-count 0)
-(define pl-s-test-pass 0)
-(define pl-s-test-fail 0)
-(define pl-s-test-failures (list))
-
-(define
-  pl-s-test!
-  (fn
-    (name got expected)
-    (begin
-      (set! pl-s-test-count (+ pl-s-test-count 1))
-      (if
-        (= got expected)
-        (set! pl-s-test-pass (+ pl-s-test-pass 1))
-        (begin
-          (set! pl-s-test-fail (+ pl-s-test-fail 1))
-          (append!
-            pl-s-test-failures
-            (str name "\n    expected: " expected "\n    got:      " got)))))))
-
-(define
-  pl-s-goal
-  (fn
-    (src env)
-    (pl-instantiate (nth (first (pl-parse (str "g :- " src "."))) 2) env)))
-
-(define pl-s-empty-db (pl-mk-db))
-
-(pl-s-test!
-  "true succeeds"
-  (pl-solve-once! pl-s-empty-db (pl-s-goal "true" {}) (pl-mk-trail))
-  true)
-
-(pl-s-test!
-  "fail fails"
-  (pl-solve-once! pl-s-empty-db (pl-s-goal "fail" {}) (pl-mk-trail))
-  false)
-
-(pl-s-test!
-  "= identical atoms"
-  (pl-solve-once!
-    pl-s-empty-db
-    (pl-s-goal "=(a, a)" {})
-    (pl-mk-trail))
-  true)
-
-(pl-s-test!
-  "= different atoms"
-  (pl-solve-once!
-    pl-s-empty-db
-    (pl-s-goal "=(a, b)" {})
-    (pl-mk-trail))
-  false)
-
-(pl-s-test!
-  "= var to atom"
-  (pl-solve-once!
-    pl-s-empty-db
-    (pl-s-goal "=(X, foo)" {})
-    (pl-mk-trail))
-  true)
-
-(define pl-s-env-bind {})
-(define pl-s-trail-bind (pl-mk-trail))
-(define pl-s-goal-bind (pl-s-goal "=(X, foo)" pl-s-env-bind))
-(pl-solve-once! pl-s-empty-db pl-s-goal-bind pl-s-trail-bind)
-
-(pl-s-test!
-  "X bound to foo after =(X, foo)"
-  (pl-atom-name (pl-walk-deep (dict-get pl-s-env-bind "X")))
-  "foo")
-
-(pl-s-test!
-  "true , true succeeds"
-  (pl-solve-once!
-    pl-s-empty-db
-    (pl-s-goal "true, true" {})
-    (pl-mk-trail))
-  true)
-
-(pl-s-test!
-  "true , fail fails"
-  (pl-solve-once!
-    pl-s-empty-db
-    (pl-s-goal "true, fail" {})
-    (pl-mk-trail))
-  false)
-
-(pl-s-test!
-  "consistent X bindings succeed"
-  (pl-solve-once!
-    pl-s-empty-db
-    (pl-s-goal "=(X, a), =(X, a)" {})
-    (pl-mk-trail))
-  true)
-
-(pl-s-test!
-  "conflicting X bindings fail"
-  (pl-solve-once!
-    pl-s-empty-db
-    (pl-s-goal "=(X, a), =(X, b)" {})
-    (pl-mk-trail))
-  false)
-
-(define pl-s-db1 (pl-mk-db))
-(pl-db-load!
-  pl-s-db1
-  (pl-parse "parent(tom, bob). parent(bob, liz). parent(bob, ann)."))
-
-(pl-s-test!
-  "fact lookup hit"
-  (pl-solve-once!
-    pl-s-db1
-    (pl-s-goal "parent(tom, bob)" {})
-    (pl-mk-trail))
-  true)
-
-(pl-s-test!
-  "fact lookup miss"
-  (pl-solve-once!
-    pl-s-db1
-    (pl-s-goal "parent(tom, liz)" {})
-    (pl-mk-trail))
-  false)
-
-(pl-s-test!
-  "all parent solutions"
-  (pl-solve-count!
-    pl-s-db1
-    (pl-s-goal "parent(X, Y)" {})
-    (pl-mk-trail))
-  3)
-
-(pl-s-test!
-  "fixed first arg solutions"
-  (pl-solve-count!
-    pl-s-db1
-    (pl-s-goal "parent(bob, Y)" {})
-    (pl-mk-trail))
-  2)
-
-(define pl-s-db2 (pl-mk-db))
-(pl-db-load!
-  pl-s-db2
-  (pl-parse
-    "parent(tom, bob). parent(bob, ann). ancestor(X, Y) :- parent(X, Y). ancestor(X, Z) :- parent(X, Y), ancestor(Y, Z)."))
-
-(pl-s-test!
-  "rule direct ancestor"
-  (pl-solve-once!
-    pl-s-db2
-    (pl-s-goal "ancestor(tom, bob)" {})
-    (pl-mk-trail))
-  true)
-
-(pl-s-test!
-  "rule transitive ancestor"
-  (pl-solve-once!
-    pl-s-db2
-    (pl-s-goal "ancestor(tom, ann)" {})
-    (pl-mk-trail))
-  true)
-
-(pl-s-test!
-  "rule no path"
-  (pl-solve-once!
-    pl-s-db2
-    (pl-s-goal "ancestor(ann, tom)" {})
-    (pl-mk-trail))
-  false)
-
-(define pl-s-env-undo {})
-(define pl-s-trail-undo (pl-mk-trail))
-(define pl-s-goal-undo (pl-s-goal "=(X, a), fail" pl-s-env-undo))
-(pl-solve-once! pl-s-empty-db pl-s-goal-undo pl-s-trail-undo)
-
-(pl-s-test!
-  "trail undone after failure leaves X unbound"
-  (pl-var-bound? (dict-get pl-s-env-undo "X"))
-  false)
-
-(define pl-s-db-cut1 (pl-mk-db))
-
-(pl-db-load! pl-s-db-cut1 (pl-parse "g :- !. g :- true."))
-
-(pl-s-test!
-  "bare cut succeeds"
-  (pl-solve-once! pl-s-db-cut1 (pl-s-goal "g" {}) (pl-mk-trail))
-  true)
-
-(pl-s-test!
-  "cut commits to first matching clause"
-  (pl-solve-count! pl-s-db-cut1 (pl-s-goal "g" {}) (pl-mk-trail))
-  1)
-
-(define pl-s-db-cut2 (pl-mk-db))
-
-(pl-db-load! pl-s-db-cut2 (pl-parse "a(1). a(2). g(X) :- a(X), !."))
-
-(pl-s-test!
-  "cut commits to first a solution"
-  (pl-solve-count! pl-s-db-cut2 (pl-s-goal "g(X)" {}) (pl-mk-trail))
-  1)
-
-(define pl-s-db-cut3 (pl-mk-db))
-
-(pl-db-load!
-  pl-s-db-cut3
-  (pl-parse "a(1). a(2). g(X) :- a(X), !, fail. g(99)."))
-
-(pl-s-test!
-  "cut then fail blocks alt clauses"
-  (pl-solve-count! pl-s-db-cut3 (pl-s-goal "g(X)" {}) (pl-mk-trail))
-  0)
-
-(define pl-s-db-cut4 (pl-mk-db))
-
-(pl-db-load!
-  pl-s-db-cut4
-  (pl-parse "a(1). b(10). b(20). g(X, Y) :- a(X), !, b(Y)."))
-
-(pl-s-test!
-  "post-cut goal backtracks freely"
-  (pl-solve-count!
-    pl-s-db-cut4
-    (pl-s-goal "g(X, Y)" {})
-    (pl-mk-trail))
-  2)
-
-(define pl-s-db-cut5 (pl-mk-db))
-
-(pl-db-load!
-  pl-s-db-cut5
-  (pl-parse "r(1). r(2). q :- r(X), !. p :- q. p :- true."))
-
-(pl-s-test!
-  "inner cut does not commit outer predicate"
-  (pl-solve-count! pl-s-db-cut5 (pl-s-goal "p" {}) (pl-mk-trail))
-  2)
-
-(pl-s-test!
-  "\\= different atoms succeeds"
-  (pl-solve-once!
-    pl-s-empty-db
-    (pl-s-goal "\\=(a, b)" {})
-    (pl-mk-trail))
-  true)
-
-(pl-s-test!
-  "\\= same atoms fails"
-  (pl-solve-once!
-    pl-s-empty-db
-    (pl-s-goal "\\=(a, a)" {})
-    (pl-mk-trail))
-  false)
-
-(pl-s-test!
-  "\\= var-vs-atom would unify so fails"
-  (pl-solve-once!
-    pl-s-empty-db
-    (pl-s-goal "\\=(X, a)" {})
-    (pl-mk-trail))
-  false)
-
-(define pl-s-env-ne {})
-
-(define pl-s-trail-ne (pl-mk-trail))
-
-(define pl-s-goal-ne (pl-s-goal "\\=(X, a)" pl-s-env-ne))
-
-(pl-solve-once! pl-s-empty-db pl-s-goal-ne pl-s-trail-ne)
-
-(pl-s-test!
-  "\\= leaves no bindings"
-  (pl-var-bound? (dict-get pl-s-env-ne "X"))
-  false)
-
-(pl-s-test!
-  "; left succeeds"
-  (pl-solve-once!
-    pl-s-empty-db
-    (pl-s-goal ";(true, fail)" {})
-    (pl-mk-trail))
-  true)
-
-(pl-s-test!
-  "; right succeeds when left fails"
-  (pl-solve-once!
-    pl-s-empty-db
-    (pl-s-goal ";(fail, true)" {})
-    (pl-mk-trail))
-  true)
-
-(pl-s-test!
-  "; both fail"
-  (pl-solve-once!
-    pl-s-empty-db
-    (pl-s-goal ";(fail, fail)" {})
-    (pl-mk-trail))
-  false)
-
-(pl-s-test!
-  "; both branches counted"
-  (pl-solve-count!
-    pl-s-empty-db
-    (pl-s-goal ";(true, true)" {})
-    (pl-mk-trail))
-  2)
-
-(define pl-s-db-call (pl-mk-db))
-
-(pl-db-load! pl-s-db-call (pl-parse "p(1). p(2)."))
-
-(pl-s-test!
-  "call(true) succeeds"
-  (pl-solve-once!
-    pl-s-db-call
-    (pl-s-goal "call(true)" {})
-    (pl-mk-trail))
-  true)
-
-(pl-s-test!
-  "call(p(X)) yields all solutions"
-  (pl-solve-count!
-    pl-s-db-call
-    (pl-s-goal "call(p(X))" {})
-    (pl-mk-trail))
-  2)
-
-(pl-s-test!
-  "call of bound goal var resolves"
-  (pl-solve-once!
-    pl-s-db-call
-    (pl-s-goal "=(G, true), call(G)" {})
-    (pl-mk-trail))
-  true)
-
-(define pl-s-db-ite (pl-mk-db))
-
-(pl-db-load! pl-s-db-ite (pl-parse "p(1). p(2). q(yes). q(no)."))
-
-(pl-s-test!
-  "if-then-else: cond true → then runs"
-  (pl-solve-once!
-    pl-s-db-ite
-    (pl-s-goal ";(->(true, =(X, ok)), =(X, fallback))" {})
-    (pl-mk-trail))
-  true)
-
-(define pl-s-env-ite1 {})
-
-(pl-solve-once!
-  pl-s-db-ite
-  (pl-s-goal ";(->(true, =(X, ok)), =(X, fallback))" pl-s-env-ite1)
-  (pl-mk-trail))
-
-(pl-s-test!
-  "if-then-else: cond true binds via then"
-  (pl-atom-name (pl-walk-deep (dict-get pl-s-env-ite1 "X")))
-  "ok")
-
-(pl-s-test!
-  "if-then-else: cond false → else"
-  (pl-solve-once!
-    pl-s-db-ite
-    (pl-s-goal ";(->(fail, =(X, ok)), =(X, fallback))" {})
-    (pl-mk-trail))
-  true)
-
-(define pl-s-env-ite2 {})
-
-(pl-solve-once!
-  pl-s-db-ite
-  (pl-s-goal ";(->(fail, =(X, ok)), =(X, fallback))" pl-s-env-ite2)
-  (pl-mk-trail))
-
-(pl-s-test!
-  "if-then-else: cond false binds via else"
-  (pl-atom-name (pl-walk-deep (dict-get pl-s-env-ite2 "X")))
-  "fallback")
-
-(pl-s-test!
-  "if-then-else: cond commits to first solution (count = 1)"
-  (pl-solve-count!
-    pl-s-db-ite
-    (pl-s-goal ";(->(p(X), =(Y, found)), =(Y, none))" {})
-    (pl-mk-trail))
-  1)
-
-(pl-s-test!
-  "if-then-else: then can backtrack"
-  (pl-solve-count!
-    pl-s-db-ite
-    (pl-s-goal ";(->(true, p(X)), =(X, none))" {})
-    (pl-mk-trail))
-  2)
-
-(pl-s-test!
-  "if-then-else: else can backtrack"
-  (pl-solve-count!
-    pl-s-db-ite
-    (pl-s-goal ";(->(fail, =(X, ignored)), p(X))" {})
-    (pl-mk-trail))
-  2)
-
-(pl-s-test!
-  "standalone -> with true cond succeeds"
-  (pl-solve-once!
-    pl-s-db-ite
-    (pl-s-goal "->(true, =(X, hi))" {})
-    (pl-mk-trail))
-  true)
-
-(pl-s-test!
-  "standalone -> with false cond fails"
-  (pl-solve-once!
-    pl-s-db-ite
-    (pl-s-goal "->(fail, =(X, hi))" {})
-    (pl-mk-trail))
-  false)
-
-(pl-s-test!
-  "write(hello)"
-  (begin
-    (pl-output-clear!)
-    (pl-solve-once!
-      pl-s-empty-db
-      (pl-s-goal "write(hello)" {})
-      (pl-mk-trail))
-    pl-output-buffer)
-  "hello")
-
-(pl-s-test!
-  "nl outputs newline"
-  (begin
-    (pl-output-clear!)
-    (pl-solve-once! pl-s-empty-db (pl-s-goal "nl" {}) (pl-mk-trail))
-    pl-output-buffer)
-  "\n")
-
-(pl-s-test!
-  "write(42) outputs digits"
-  (begin
-    (pl-output-clear!)
-    (pl-solve-once!
-      pl-s-empty-db
-      (pl-s-goal "write(42)" {})
-      (pl-mk-trail))
-    pl-output-buffer)
-  "42")
-
-(pl-s-test!
-  "write(foo(a, b)) formats compound"
-  (begin
-    (pl-output-clear!)
-    (pl-solve-once!
-      pl-s-empty-db
-      (pl-s-goal "write(foo(a, b))" {})
-      (pl-mk-trail))
-    pl-output-buffer)
-  "foo(a, b)")
-
-(pl-s-test!
-  "write conjunction"
-  (begin
-    (pl-output-clear!)
-    (pl-solve-once!
-      pl-s-empty-db
-      (pl-s-goal "write(a), write(b)" {})
-      (pl-mk-trail))
-    pl-output-buffer)
-  "ab")
-
-(pl-s-test!
-  "write of bound var walks binding"
-  (begin
-    (pl-output-clear!)
-    (pl-solve-once!
-      pl-s-empty-db
-      (pl-s-goal "=(X, hello), write(X)" {})
-      (pl-mk-trail))
-    pl-output-buffer)
-  "hello")
-
-(pl-s-test!
-  "write then nl"
-  (begin
-    (pl-output-clear!)
-    (pl-solve-once!
-      pl-s-empty-db
-      (pl-s-goal "write(hi), nl" {})
-      (pl-mk-trail))
-    pl-output-buffer)
-  "hi\n")
-
-(define pl-s-env-arith1 {})
-
-(pl-solve-once!
-  pl-s-empty-db
-  (pl-s-goal "is(X, 42)" pl-s-env-arith1)
-  (pl-mk-trail))
-
-(pl-s-test!
-  "is(X, 42) binds X to 42"
-  (pl-num-val (pl-walk-deep (dict-get pl-s-env-arith1 "X")))
-  42)
-
-(define pl-s-env-arith2 {})
-
-(pl-solve-once!
-  pl-s-empty-db
-  (pl-s-goal "is(X, +(2, 3))" pl-s-env-arith2)
-  (pl-mk-trail))
-
-(pl-s-test!
-  "is(X, +(2, 3)) binds X to 5"
-  (pl-num-val (pl-walk-deep (dict-get pl-s-env-arith2 "X")))
-  5)
-
-(define pl-s-env-arith3 {})
-
-(pl-solve-once!
-  pl-s-empty-db
-  (pl-s-goal "is(X, *(2, 3))" pl-s-env-arith3)
-  (pl-mk-trail))
-
-(pl-s-test!
-  "is(X, *(2, 3)) binds X to 6"
-  (pl-num-val (pl-walk-deep (dict-get pl-s-env-arith3 "X")))
-  6)
-
-(define pl-s-env-arith4 {})
-
-(pl-solve-once!
-  pl-s-empty-db
-  (pl-s-goal "is(X, -(10, 3))" pl-s-env-arith4)
-  (pl-mk-trail))
-
-(pl-s-test!
-  "is(X, -(10, 3)) binds X to 7"
-  (pl-num-val (pl-walk-deep (dict-get pl-s-env-arith4 "X")))
-  7)
-
-(define pl-s-env-arith5 {})
-
-(pl-solve-once!
-  pl-s-empty-db
-  (pl-s-goal "is(X, /(10, 2))" pl-s-env-arith5)
-  (pl-mk-trail))
-
-(pl-s-test!
-  "is(X, /(10, 2)) binds X to 5"
-  (pl-num-val (pl-walk-deep (dict-get pl-s-env-arith5 "X")))
-  5)
-
-(define pl-s-env-arith6 {})
-
-(pl-solve-once!
-  pl-s-empty-db
-  (pl-s-goal "is(X, mod(10, 3))" pl-s-env-arith6)
-  (pl-mk-trail))
-
-(pl-s-test!
-  "is(X, mod(10, 3)) binds X to 1"
-  (pl-num-val (pl-walk-deep (dict-get pl-s-env-arith6 "X")))
-  1)
-
-(define pl-s-env-arith7 {})
-
-(pl-solve-once!
-  pl-s-empty-db
-  (pl-s-goal "is(X, abs(-(0, 5)))" pl-s-env-arith7)
-  (pl-mk-trail))
-
-(pl-s-test!
-  "is(X, abs(-(0, 5))) binds X to 5"
-  (pl-num-val (pl-walk-deep (dict-get pl-s-env-arith7 "X")))
-  5)
-
-(define pl-s-env-arith8 {})
-
-(pl-solve-once!
-  pl-s-empty-db
-  (pl-s-goal "is(X, +(2, *(3, 4)))" pl-s-env-arith8)
-  (pl-mk-trail))
-
-(pl-s-test!
-  "is(X, +(2, *(3, 4))) binds X to 14 (nested)"
-  (pl-num-val (pl-walk-deep (dict-get pl-s-env-arith8 "X")))
-  14)
-
-(pl-s-test!
-  "is(5, +(2, 3)) succeeds (LHS num matches)"
-  (pl-solve-once!
-    pl-s-empty-db
-    (pl-s-goal "is(5, +(2, 3))" {})
-    (pl-mk-trail))
-  true)
-
-(pl-s-test!
-  "is(6, +(2, 3)) fails (LHS num mismatch)"
-  (pl-solve-once!
-    pl-s-empty-db
-    (pl-s-goal "is(6, +(2, 3))" {})
-    (pl-mk-trail))
-  false)
-
-(pl-s-test!
-  "is propagates bound vars on RHS"
-  (pl-solve-once!
-    pl-s-empty-db
-    (pl-s-goal "=(Y, 4), is(X, +(Y, 1)), =(X, 5)" {})
-    (pl-mk-trail))
-  true)
-
-(define pl-solve-tests-run! (fn () {:failed pl-s-test-fail :passed pl-s-test-pass :total pl-s-test-count :failures pl-s-test-failures}))

plans/forth-on-sx.md

@@ -69,36 +69,333 @@ Representation:
 - [x] Tests in `lib/forth/tests/test-phase2.sx` — 26/26 pass
 
 ### Phase 3 — control flow + first Hayes tests green
-- [ ] `IF`, `ELSE`, `THEN` — compile to SX `if`
-- [ ] `BEGIN`, `UNTIL`, `WHILE`, `REPEAT`, `AGAIN` — compile to loops
-- [ ] `DO`, `LOOP`, `+LOOP`, `I`, `J`, `LEAVE` — counted loops (needs a return stack)
-- [ ] Return stack: `>R`, `R>`, `R@`, `2>R`, `2R>`, `2R@`
-- [ ] Vendor John Hayes' test suite to `lib/forth/ans-tests/`
-- [ ] `lib/forth/conformance.sh` + runner; `scoreboard.json` + `scoreboard.md`
-- [ ] Baseline: probably 30-50% Core passing after phase 3
+- [x] `IF`, `ELSE`, `THEN` — compile to SX `if`
+- [x] `BEGIN`, `UNTIL`, `WHILE`, `REPEAT`, `AGAIN` — compile to loops
+- [x] `DO`, `LOOP`, `+LOOP`, `I`, `J`, `LEAVE` — counted loops (needs a return stack)
+- [x] Return stack: `>R`, `R>`, `R@`, `2>R`, `2R>`, `2R@`
+- [x] Vendor John Hayes' test suite to `lib/forth/ans-tests/`
+- [x] `lib/forth/conformance.sh` + runner; `scoreboard.json` + `scoreboard.md`
+- [x] Baseline: probably 30-50% Core passing after phase 3
 
 ### Phase 4 — strings + more Core
-- [ ] `S"`, `C"`, `."`, `TYPE`, `COUNT`, `CMOVE`, `FILL`, `BLANK`
-- [ ] `CHAR`, `[CHAR]`, `KEY`, `ACCEPT`
-- [ ] `BASE` manipulation: `DECIMAL`, `HEX`
-- [ ] `DEPTH`, `SP@`, `SP!`
-- [ ] Drive Hayes Core pass-rate up
+- [x] `S"`, `C"`, `."`, `TYPE`, `COUNT`, `CMOVE`, `FILL`, `BLANK`
+- [x] `CHAR`, `[CHAR]`, `KEY`, `ACCEPT`
+- [x] `BASE` manipulation: `DECIMAL`, `HEX`
+- [x] `DEPTH`, `SP@`, `SP!`
+- [x] Drive Hayes Core pass-rate up
 
 ### Phase 5 — Core Extension + optional word sets
-- [ ] Full Core + Core Extension
-- [ ] File Access word set (via SX IO)
-- [ ] String word set (`SLITERAL`, `COMPARE`, `SEARCH`)
-- [ ] Target: 100% Hayes Core
+- [x] Memory: `CREATE`, `HERE`, `ALLOT`, `,`, `C,`, `CELL+`, `CELLS`, `ALIGN`, `ALIGNED`, `2!`, `2@`
+- [x] Unsigned compare: `U<`, `U>`
+- [x] Mixed/double-cell math: `S>D`, `M*`, `UM*`, `UM/MOD`, `FM/MOD`, `SM/REM`, `*/`, `*/MOD`
+- [x] Double-cell ops: `D+`, `D-`, `D=`, `D<`, `D0=`, `2DUP`, `2DROP`, `2OVER`, `2SWAP` (already), plus `D>S`, `DABS`, `DNEGATE`
+- [x] Number formatting: `<#`, `#`, `#S`, `#>`, `HOLD`, `SIGN`, `.R`, `U.`, `U.R`
+- [x] Parsing/dictionary: `WORD`, `FIND`, `EXECUTE`, `'`, `[']`, `LITERAL`, `POSTPONE`, `>BODY` (DOES> deferred — needs runtime-rebind of last CREATE)
+- [x] Source/state: `EVALUATE`, `STATE`, `[`, `]` (`SOURCE`/`>IN` stubbed; tokenized input means the exact byte/offset semantics aren't useful here)
+- [x] Misc Core: `WITHIN`, `MAX`/`MIN` (already), `ABORT`, `ABORT"`, `EXIT`, `UNLOOP`
+- [x] File Access word set (in-memory — `read-file` is not reachable from the epoch eval env)
+- [x] String word set (`SLITERAL`, `COMPARE`, `SEARCH`)
+- [x] Target: 100% Hayes Core (97% achieved — remaining 5 errors all in `GI5`'s multi-`WHILE`-per-`BEGIN` non-standard pattern, plus one stuck `dict-set!` chunk and 14 numeric-edge fails)
 
 ### Phase 6 — speed
-- [ ] Inline primitive calls during compile (skip dict lookup)
-- [ ] Tail-call optimise colon-def endings
-- [ ] JIT cooperation: mark compiled colon-defs as VM-eligible
+- [x] Inline primitive calls during compile (skip dict lookup)
+- [x] Tail-call optimise colon-def endings
+- [x] JIT cooperation: mark compiled colon-defs as VM-eligible
 
 ## Progress log
 
 _Newest first._
 
+- **Phase 6 close — JIT cooperation hooks (Hayes unchanged at 618/638).**
+  Every word record now carries `:vm-eligible? true` and a
+  `:call-count` counter that `forth-execute-word` bumps on every
+  invocation. The flag is a hint for downstream JIT consumers — our
+  bodies are plain SX lambdas already, so the existing SX VM's
+  on-first-call JIT lifts them into bytecode automatically; the
+  metadata just makes that fact discoverable. Added
+  `forth-hot-words state threshold` returning `(name count)`
+  pairs above a threshold so a future tracing JIT can pick out
+  hot definitions to specialise. Phase 6 boxes all ticked.
+  All 306 internal tests green; Hayes Core stays at 618/638.
+
+- **Phase 6 — TCO at colon-def endings (Hayes unchanged at 618/638).**
+  `forth-run-body` now special-cases the final op when it's a plain
+  function (not a branch dict): we call it in tail position with no
+  pc-increment and no recursive `forth-run-body` call. This means
+  the SX CEK can collapse the continuation frame, so chains like
+  `: A ... B ; : B ... C ; …` and `RECURSE` deep-recursion test
+  cases run without piling up frames at each colon-def boundary.
+  All 306 internal tests still green; verified 5000-deep
+  `COUNTDOWN RECURSE` still terminates fine.
+
+- **Phase 6 — inline primitive calls (Hayes unchanged at 618/638).**
+  `forth-compile-call` now appends the looked-up word's body fn
+  directly to the colon-def body instead of wrapping it in
+  `(fn (s) (forth-execute-word s w))`. `forth-execute-word body`
+  reduces to `((get w "body") state)`, so the wrapper added an
+  extra closure + `get` per call op for no behavioural gain. Same
+  early-binding semantics: the body fn is captured at compile time,
+  so later redefinitions of the same name don't retroactively
+  change existing definitions. All 306 internal tests still green;
+  Hayes Core stays at 618/638. Pure optimisation.
+
+- **Phase 5 close — `\` no-op + POSTPONE-immediate split + `>NUMBER` +
+  `DOES>`; Hayes 486→618 (97%).** Big closing-out iteration.
+  Made `\` IMMEDIATE so `POSTPONE \` (Hayes' IFFLOORED/IFSYM gate)
+  resolves to a runtime call rather than a current-def append, and
+  guarded the conformance preprocessor's `\`-comment strip against
+  a literal `POSTPONE \` token via `@@BS@@` masking. Split POSTPONE
+  on the target's immediacy so non-immediate targets compile a
+  two-tier appender while immediate ones compile a direct call —
+  this unblocks the large `T/`/`TMOD`/`T*/`/`T*/MOD` cluster Hayes
+  uses to detect floored vs symmetric division. `>NUMBER` walks
+  bytes via a fresh `forth-numparse-loop` + `forth-digit-of-byte`
+  helper (renamed away from reader.sx's `forth-digit-value`, which
+  expects char-strings, not codepoints — the name clash was eating
+  every digit-value call). Implemented `DOES>` by:
+  1) tracking the last CREATE on `state.last-creator`,
+  2) adding a `:kind "does-rebind"` op, and
+  3) post-processing the body in `;` to attach the slice of ops
+     after each rebind as `:deferred`. At runtime, the rebind op
+     installs a new body for the target word that pushes its
+     data-field address and runs the deferred slice. Also added
+     histogram tracking on the conformance runner so future runs
+     surface the top missing words. Hayes: 618/638 pass (97%),
+     14 fail, 6 error (5× GI5 multi-WHILE, 1× dict-set! chunk).
+
+- **Phase 5 — String word set `COMPARE`/`SEARCH`/`SLITERAL` (+9).**
+  `COMPARE` walks bytes via the new `forth-compare-bytes-loop`,
+  returning -1/0/1 with standard prefix semantics (shorter string
+  compares less than its extension). `SEARCH` scans the haystack
+  with a helper `forth-search-bytes` and `forth-match-at`, returning
+  the tail after the first match or the original string with flag=0.
+  Empty needle returns at offset 0 with flag=-1 per ANS. `SLITERAL`
+  is IMMEDIATE: pops `(c-addr u)` at compile time, copies the bytes
+  into a fresh allocation, and emits the two pushes so the compiled
+  word yields the interned string at runtime.
+
+- **Phase 5 — File Access word set (in-memory backing; +4).**
+  `OPEN-FILE`/`CREATE-FILE`/`CLOSE-FILE`/`READ-FILE`/`WRITE-FILE`/
+  `FILE-POSITION`/`FILE-SIZE`/`REPOSITION-FILE`/`DELETE-FILE` plus
+  the mode constants `R/O`/`R/W`/`W/O`/`BIN`. File handles live on
+  `state.files` (fileid → {content, pos, path}) with a
+  `state.by-path` index so `CREATE-FILE`'d files can be
+  `OPEN-FILE`'d later in the same session. Attempting to
+  `OPEN-FILE` an unknown path returns `ior != 0`; disk-backed
+  open/read is not wired because `read-file` isn't in the sx_server
+  epoch eval environment (it's bound only in the HTTP helpers).
+  Also removed the stray base-2 `BIN` primitive from Phase 4 —
+  ANS `BIN` is the file-mode modifier. Hayes Core unchanged at
+  486/638 since core.fr doesn't exercise file words.
+
+- **Phase 5 — `WITHIN`/`ABORT`/`ABORT"`/`EXIT`/`UNLOOP` (+7;
+  Hayes 477→486, 76%).** `WITHIN` uses the ANS two's-complement
+  trick: `(n1-n2) U< (n3-n2)`. `ABORT` wipes the data/return/control
+  stacks and raises — the conformance runner catches it at the
+  chunk boundary. `ABORT"` parses its message like `S"`, then at
+  runtime pops a flag and raises only if truthy. `EXIT` adds a new
+  `:kind "exit"` op that the PC-driven body runner treats as a
+  jump-to-end; added a matching cond clause in `forth-step-op`.
+  `UNLOOP` pops two from the return stack — usable paired with
+  `EXIT` to bail from inside `DO`/`LOOP`.
+
+- **Phase 5 — `[`, `]`, `STATE`, `EVALUATE` (+5; Hayes 463→477, 74%).**
+  `[` (IMMEDIATE) clears `state.compiling`, `]` sets it. `STATE`
+  pushes the sentinel address `"@@state"` and `@` reads it as
+  `-1`/`0` based on the live `compiling` flag. `EVALUATE` reads
+  the (addr,u) string from byte memory, retokenises it via
+  `forth-tokens`, swaps it in as the active input, runs the
+  interpret loop, and restores the saved input. `SOURCE` and
+  `>IN` exist as stubs that push zeros — our whitespace-tokenised
+  input has no native byte-offset, so the deeper Hayes tests
+  that re-position parsing via `>IN !` stay marked as errors
+  rather than silently misbehaving.
+
+- **Phase 5 — parsing/dictionary words `'`/`[']`/`EXECUTE`/`LITERAL`/
+  `POSTPONE`/`WORD`/`FIND`/`>BODY` (Hayes 448→463, 72%).** xt is
+  represented as the SX dict reference of the word record, so
+  `'`/`[']` push the looked-up record and `EXECUTE` calls
+  `forth-execute-word` on the popped value. `LITERAL` (IMMEDIATE)
+  pops a value at compile time and emits a push-op. `POSTPONE`
+  (IMMEDIATE) compiles into the *outer* def an op that, when run
+  during a *later* compile, appends a call-w op to whatever def is
+  current — the standard two-tier compile semantic. Added
+  `state.last-defined` tracked by every primitive/colon definition
+  so `IMMEDIATE` can target the most-recent word even after `;`
+  closes the def. CREATE now stashes its data-field address on the
+  word record so `>BODY` can recover it. `WORD`/`FIND` use the byte
+  memory and counted-string layout already in place.
+  `DOES>` is deferred — needs a runtime mechanism to rebind the
+  last-CREATE'd word's action.
+
+- **Phase 5 — pictured numeric output: `<#`/`#`/`#S`/`#>`/`HOLD`/`SIGN` +
+  `U.`/`U.R`/`.R` (+9; Hayes 446→448, 70%).** Added a `state.hold`
+  list of single-character strings — `<#` resets it, `HOLD` and
+  `SIGN` prepend, `#` divides ud by BASE and prepends one digit,
+  `#S` loops `#` until ud is zero (running once even on zero),
+  `#>` drops ud and copies the joined hold buffer into mem,
+  pushing `(addr, len)`. `U.` / `.R` / `U.R` use a separate
+  `forth-num-to-string` for one-shot decimal/hex output and
+  `forth-spaces-str` for right-justify padding.
+
+- **Phase 5 — double-cell ops `D+`/`D-`/`DNEGATE`/`DABS`/`D=`/`D<`/`D0=`/
+  `D0<`/`DMAX`/`DMIN` (+18; Hayes unchanged).** Doubles get rebuilt
+  from `(lo, hi)` cells via `forth-double-from-cells-s`, the op runs
+  in bignum, and we push back via `forth-double-push-s`. Hayes Core
+  doesn't exercise D-words (those live in Gerry Jackson's separate
+  `doublest.fth` Double word-set tests we have not vendored), so the
+  scoreboard stays at 446/638 — but the words now exist for any
+  consumer that needs them.
+
+- **Phase 5 — mixed/double-cell math; Hayes 342→446 (69%).** Added
+  `S>D`, `D>S`, `M*`, `UM*`, `UM/MOD`, `FM/MOD`, `SM/REM`, `*/`, `*/MOD`.
+  Doubles ride on the stack as `(lo, hi)` with `hi` on top.
+  Helpers `forth-double-push-{u,s}` / `forth-double-from-cells-{u,s}`
+  split & rebuild via 32-bit unsigned mod/div, picking the negative
+  path explicitly so we don't form `2^64 + small` (float precision
+  drops at ULP=2^12 once you cross 2^64). `M*`/`UM*` use bignum
+  multiply then split; `*/`/`*/MOD` use bignum intermediate and
+  truncated division. Hayes: 446 pass / 185 error / 7 fail.
+
+- **Phase 5 — memory primitives + unsigned compare; Hayes 268→342 (53%).**
+  Added `CREATE`/`HERE`/`ALLOT`/`,`/`C,`/`CELL+`/`CELLS`/`ALIGN`/`ALIGNED`/
+  `2!`/`2@`/`U<`/`U>`. Generalised `@`/`!`/`+!` to dispatch on address
+  type: string addresses still go through `state.vars` (VARIABLE/VALUE
+  cells) while integer addresses now fall through to `state.mem` —
+  letting CREATE-allocated cells coexist with existing variables.
+  Decomposed the original "Full Core + Core Extension" box into
+  smaller unticked sub-bullets so iterations land per cluster.
+  Hayes: 342 pass / 292 error / 4 fail (53%). 237/237 internal.
+
+- **Phase 4 close — LSHIFT/RSHIFT, 32-bit arith truncation, early
+  binding; Hayes 174→268 (42%).** Added `LSHIFT` / `RSHIFT` as logical
+  shifts on 32-bit unsigned values, converted through
+  `forth-to-unsigned`/`forth-from-unsigned`. All arithmetic
+  primitives (`+` `-` `*` `/` `MOD` `NEGATE` `ABS` `1+` `1-` `2+`
+  `2-` `2*` `2/`) now clip results to 32-bit signed via a new
+  `forth-clip` helper, so loop idioms that rely on `2*` shifting the
+  MSB out (e.g. Hayes' `BITS` counter) actually terminate.
+  Changed colon-def call compilation from late-binding to early
+  binding: `forth-compile-call` now resolves the target word at
+  compile time, which makes `: GDX 123 ; : GDX GDX 234 ;` behave
+  per ANS (inner `GDX` → old def, not infinite recursion). `RECURSE`
+  keeps its late-binding thunk via the new `forth-compile-recurse`
+  helper. Raised `MAX_CHUNKS` default to 638 (full `core.fr`) now
+  that the BITS and COUNT-BITS loops terminate. Hayes: 268 pass /
+  368 error / 2 fail.
+
+- **Phase 4 — `SP@`/`SP!` (+4; Hayes unchanged; `DEPTH` was already present).**
+  `SP@` pushes the current data-stack depth (our closest analogue to a
+  stack pointer — SX lists have no addressable backing). `SP!` pops a
+  target depth and truncates the stack via `drop` on the dstack list.
+  This preserves the save/restore idiom `SP@ … SP!` even though the
+  returned "pointer" is really a count.
+
+- **Phase 4 — `BASE`/`DECIMAL`/`HEX`/`BIN`/`OCTAL` (+9; Hayes unchanged).**
+  Moved `base` from its top-level state slot into `state.vars["base"]`
+  so the regular `@`/`!`/VARIABLE machinery works on it.
+  `BASE` pushes the sentinel address `"base"`; `DECIMAL`/`HEX`/`BIN`/
+  `OCTAL` are thin primitives that write into that slot. Parser
+  reads through `vars` now. Hayes unchanged because the runner had
+  already been stubbing `HEX`/`DECIMAL` — now real words, stubs
+  removed from `hayes-runner.sx`.
+
+- **Phase 4 — `CHAR`/`[CHAR]`/`KEY`/`ACCEPT` (+7 / Hayes 168→174).**
+  `CHAR` parses the next token and pushes the first-char code. `[CHAR]`
+  is IMMEDIATE: in compile mode it embeds the code as a compiled push
+  op, in interpret mode it pushes inline. `KEY`/`ACCEPT` read from an
+  optional `state.keybuf` string — empty buffer makes `KEY` raise
+  `"no input available"` (matches ANS when stdin is closed) and
+  `ACCEPT` returns `0`. Enough for Hayes to get past CHAR-gated
+  clusters; real interactive IO lands later.
+
+- **Phase 4 — strings: `S"`/`C"`/`."`/`TYPE`/`COUNT`/`CMOVE`/`CMOVE>`/`MOVE`/`FILL`/`BLANK`/`C@`/`C!`/`CHAR+`/`CHARS` (+16 / Hayes 165→168).**
+  Added a byte-addressable memory model to state: `mem` (dict keyed by
+  stringified address → integer byte) and `here` (next-free integer
+  addr). Helpers `forth-alloc-bytes!` / `forth-mem-write-string!` /
+  `forth-mem-read-string`. `S"`/`C"`/`."` are IMMEDIATE parsing words
+  that consume tokens until one ends with `"`, then either copy content
+  into memory at compile time (and emit a push of `addr`/`addr len` for
+  the colon-def body) or do it inline in interpret mode. `TYPE` emits
+  `u` bytes from `addr` via `char-from-code`. `COUNT` reads the length
+  byte at a counted-string address and pushes (`addr+1`, `u`). `FILL`,
+  `BLANK` (FILL with space), `CMOVE` (forward), `CMOVE>` (backward),
+  and `MOVE` (auto-directional) mutate the byte dict. 193/193 internal
+  tests, Hayes 168/590 (+3).
+
+- **Phase 3 — Hayes conformance runner + baseline scoreboard (165/590, 28%).**
+  `lib/forth/conformance.sh` preprocesses `ans-tests/core.fr` (strips `\`
+  and `( ... )` comments + `TESTING` lines), splits the source on every
+  `}T` so each Hayes test plus the small declaration blocks between
+  them are one safe-resume chunk, and emits an SX driver that feeds
+  the chunks through `lib/forth/hayes-runner.sx`. The runner registers
+  `T{`/`->`/`}T` as Forth primitives that snapshot the dstack depth on
+  `T{`, record actual on `->`, compare on `}T`, and install stub
+  `HEX`/`DECIMAL`/`TESTING` so metadata doesn't halt the stream. Errors
+  raised inside a chunk are caught by `guard` and the state is reset,
+  so one bad test does not break the rest. Outputs
+  `scoreboard.json` + `scoreboard.md`.
+
+  First-run baseline: 165 pass / 425 error / 0 fail on the first 590
+  chunks. The default cap sits at 590 because `core.fr` chunks beyond
+  that rely on unsigned-integer wrap-around (e.g. `COUNT-BITS` with
+  `BEGIN DUP WHILE … 2* REPEAT`) which never terminates on our
+  bignum-based Forth; raise `MAX_CHUNKS` once those tests unblock.
+  Majority of errors are missing Phase-4 words (`RSHIFT`, `LSHIFT`,
+  `CELLS`, `S"`, `CHAR`, `SOURCE`, etc.) — each one implemented should
+  convert a cluster of errors to passes.
+
+- **Phase 3 — vendor Gerry Jackson's forth2012-test-suite.** Added
+  `lib/forth/ans-tests/{tester.fr, core.fr, coreexttest.fth}` from
+  https://github.com/gerryjackson/forth2012-test-suite (master, fetched
+  2026-04-24). `tester.fr` is Hayes' `T{ ... -> ... }T` harness; `core.fr`
+  is the ~1000-line Core word tests; `coreexttest.fth` is Core Ext
+  (parked for later phases). Files are pristine — the conformance runner
+  (next iteration) will consume them.
+
+- **Phase 3 — `DO`/`LOOP`/`+LOOP`/`I`/`J`/`LEAVE` + return stack words (+16).**
+  Counted loops compile onto the same PC-driven body runner. DO emits an
+  enter-op (pops limit+start from data stack, pushes them to rstack) and
+  pushes a `{:kind "do" :back PC :leaves ()}` marker onto cstack. LOOP/+LOOP
+  emit a dict op (`:kind "loop"`/`"+loop"` with target=back-cell). The step
+  handler pops index & reads limit, increments, and either restores the
+  updated index + jumps back, or drops the frame and advances. LEAVE walks
+  cstack for the innermost DO marker, emits a `:kind "leave"` dict op with
+  a fresh target cell, and registers it on the marker's leaves list. LOOP
+  patches all registered leave-targets to the exit PC and drops the marker.
+  The leave op pops two from rstack (unloop) and branches. `I` peeks rtop;
+  `J` reads rstack index 2 (below inner frame). Added non-immediate
+  return-stack words `>R`, `R>`, `R@`, `2>R`, `2R>`, `2R@`. Nested
+  DO/LOOP with J tested; LEAVE in nested loops exits only the inner.
+  177/177 green.
+
+- **Phase 3 — `BEGIN`/`UNTIL`/`WHILE`/`REPEAT`/`AGAIN` (+9).** Indefinite-loop
+  constructs built on the same PC-driven body runner introduced for `IF`.
+  BEGIN records the current body length on `state.cstack` (a plain numeric
+  back-target). UNTIL/AGAIN pop that back-target and emit a `bif`/`branch`
+  op whose target cell is set to the recorded PC. WHILE emits a forward
+  `bif` with a fresh target cell and pushes it on the cstack *above* the
+  BEGIN marker; REPEAT pops both (while-target first, then back-pc), emits
+  an unconditional branch back to BEGIN, then patches the while-target to
+  the current body length — so WHILE's false flag jumps past the REPEAT.
+  Mixed compile-time layout (numeric back-targets + dict forward targets
+  on the same cstack) is OK because the immediate words pop them in the
+  order they expect. AGAIN works structurally but lacks a test without a
+  usable mid-loop exit; revisit once `EXIT` lands. 161/161 green.
+
+- **Phase 3 start — `IF`/`ELSE`/`THEN` (+18).** `lib/forth/compiler.sx`
+  + `tests/test-phase3.sx`. Colon-def body switched from `for-each` to
+  a PC-driven runner so branch ops can jump: ops now include dict tags
+  `{"kind" "bif"|"branch" "target" cell}` alongside the existing
+  `(fn (s) ...)` shape. IF compiles a `bif` with a fresh target cell
+  pushed to `state.cstack`; ELSE emits an unconditional `branch`,
+  patches the IF's target to the instruction after this branch, and
+  pushes the new target; THEN patches the most recent target to the
+  current body length. Nested IF/ELSE/THEN works via the cstack.
+  Also fixed `EMIT`: `code-char` → `char-from-code` (spec-correct
+  primitive name) so Phase 1/2 tests run green on sx_server.
+  152/152 (Phase 1 + 2 + 3) green.
+
 - **Phase 2 complete — colon defs, compile mode, VARIABLE/CONSTANT/VALUE/TO, @/!/+! (+26).**
   `lib/forth/compiler.sx` plus `tests/test-phase2.sx`.
   Colon-def body is a list of ops (one per source token) wrapped in a single

plans/prolog-on-sx.md

@@ -39,34 +39,34 @@ Representation choices (finalise in phase 1, document here):
 ## Roadmap
 
 ### Phase 1 — tokenizer + term parser (no operator table)
-- [x] Tokenizer: atoms (lowercase/quoted), variables (uppercase/`_`), numbers, strings, punct `( ) , . [ ] | ! :-`, comments (`%`, `/* */`)
-- [x] Parser: clauses `head :- body.` and facts `head.`; terms `atom | Var | number | compound(args) | [list,sugar]`
-- [x] **Skip for phase 1:** operator table. `X is Y + 1` must be written `is(X, '+'(Y, 1))`; `=` written `=(X, Y)`. Operators land in phase 4.
-- [x] Unit tests in `lib/prolog/tests/parse.sx` — 25 pass
+- [ ] Tokenizer: atoms (lowercase/quoted), variables (uppercase/`_`), numbers, strings, punct `( ) , . [ ] | ! :-`, comments (`%`, `/* */`)
+- [ ] Parser: clauses `head :- body.` and facts `head.`; terms `atom | Var | number | compound(args) | [list,sugar]`
+- [ ] **Skip for phase 1:** operator table. `X is Y + 1` must be written `is(X, '+'(Y, 1))`; `=` written `=(X, Y)`. Operators land in phase 4.
+- [ ] Unit tests in `lib/prolog/tests/parse.sx`
 
 ### Phase 2 — unification + trail
-- [x] `make-var`, `walk` (follow binding chain), `prolog-unify!` (terms + trail → bool), `trail-undo-to!`
-- [x] Occurs-check off by default, exposed as flag
-- [x] 30+ unification tests in `lib/prolog/tests/unify.sx`: atoms, vars, compounds, lists, cyclic (no-occurs-check), mutual occurs — 47 pass
+- [ ] `make-var`, `walk` (follow binding chain), `prolog-unify!` (terms + trail → bool), `trail-undo-to!`
+- [ ] Occurs-check off by default, exposed as flag
+- [ ] 30+ unification tests in `lib/prolog/tests/unify.sx`: atoms, vars, compounds, lists, cyclic (no-occurs-check), mutual occurs
 
 ### Phase 3 — clause DB + DFS solver + cut + first classic programs
-- [x] Clause DB: `"functor/arity" → list-of-clauses`, loader inserts — `pl-mk-db` / `pl-db-add!` / `pl-db-load!` / `pl-db-lookup` / `pl-db-lookup-goal`, 14 tests in `tests/clausedb.sx`
-- [x] Solver: DFS with choice points backed by delimited continuations (`lib/callcc.sx`). On goal entry, capture; per matching clause, unify head + recurse body; on failure, undo trail, try next — first cut: trail-based undo + CPS k (no shift/reset yet, per briefing gotcha). Built-ins so far: `true/0`, `fail/0`, `=/2`, `,/2`. Refactor to delimited conts later.
-- [x] Cut (`!`): cut barrier at current choice-point frame; collapse all up to barrier — two-cut-box scheme: each `pl-solve-user!` creates a fresh inner-cut-box (set by `!` in this predicate's body) AND snapshots the outer-cut-box state on entry. After body fails, abandon clause alternatives if (a) inner was set or (b) outer transitioned false→true during this call. Lets post-cut goals backtrack normally while blocking pre-cut alternatives. 6 cut tests cover bare cut, clause-commit, choice-commit, cut+fail, post-cut backtracking, nested-cut isolation.
-- [x] Built-ins: `=/2`, `\\=/2`, `true/0`, `fail/0`, `!/0`, `,/2`, `;/2`, `->/2` inside `;`, `call/1`, `write/1`, `nl/0` — all 11 done. `write/1` and `nl/0` use a global `pl-output-buffer` string + `pl-output-clear!` for testability; `pl-format-term` walks deep then renders atoms/nums/strs/compounds/vars (var → `_<id>`). Note: cut-transparency via `;` not testable yet without operator support — `;(,(a,!), b)` parser-rejects because `,` is body-operator-only; revisit in phase 4.
-- [x] Arithmetic `is/2` with `+ - * / mod abs` — `pl-eval-arith` walks deep, recurses on compounds, dispatches on functor; binary `+ - * / mod`, binary AND unary `-`, unary `abs`. `is/2` evaluates RHS, wraps as `("num" v)`, unifies via `pl-solve-eq!`. 11 tests cover each op + nested + ground LHS match/mismatch + bound-var-on-RHS chain.
+- [ ] Clause DB: `"functor/arity" → list-of-clauses`, loader inserts
+- [ ] Solver: DFS with choice points backed by delimited continuations (`lib/callcc.sx`). On goal entry, capture; per matching clause, unify head + recurse body; on failure, undo trail, try next
+- [ ] Cut (`!`): cut barrier at current choice-point frame; collapse all up to barrier
+- [ ] Built-ins: `=/2`, `\\=/2`, `true/0`, `fail/0`, `!/0`, `,/2`, `;/2`, `->/2` inside `;`, `call/1`, `write/1`, `nl/0`
+- [ ] Arithmetic `is/2` with `+ - * / mod abs`
 - [ ] Classic programs in `lib/prolog/tests/programs/`:
-  - [x] `append.pl` — list append (with backtracking) — `lib/prolog/tests/programs/append.{pl,sx}`. 6 tests cover: build (`append([], L, X)`, `append([1,2], [3,4], X)`), check ground match/mismatch, full split-backtracking (`append(X, Y, [1,2,3])` → 4 solutions), single-deduce (`append(X, [3], [1,2,3])` → X=[1,2]).
-  - [x] `reverse.pl` — naive reverse — `lib/prolog/tests/programs/reverse.{pl,sx}`. Naive reverse via append: `reverse([H|T], R) :- reverse(T, RT), append(RT, [H], R)`. 6 tests cover empty, singleton, 3-list, 4-atom-list, ground match, ground mismatch.
-  - [x] `member.pl` — generate all solutions via backtracking — `lib/prolog/tests/programs/member.{pl,sx}`. Classic 2-clause `member(X, [X|_])` + `member(X, [_|T]) :- member(X, T)`. 7 tests cover bound-element hit/miss, empty list, generator (count = list length), first-solution binding, duplicate matches counted twice, anonymous head-cell unification.
-  - [x] `nqueens.pl` — 8-queens — `lib/prolog/tests/programs/nqueens.{pl,sx}`. Permute-and-test formulation: `queens(L, Qs) :- permute(L, Qs), safe(Qs)` + `select` + `safe` + `no_attack`. Tested at N=1 (1), N=2 (0), N=3 (0), N=4 (2), N=5 (10) plus first-solution check at N=4 = `[2, 4, 1, 3]`. N=8 omitted — interpreter is too slow (40320 perms); add once compiled clauses or constraint-style placement land. `range/3` skipped pending arithmetic-comparison built-ins (`>/2` etc.).
-  - [x] `family.pl` — facts + rules (parent/ancestor) — `lib/prolog/tests/programs/family.{pl,sx}`. 5 parent facts + male/female + derived `father`/`mother`/`ancestor`/`sibling`. 10 tests cover direct facts, fact count, transitive ancestor through 3 generations, descendant counting, gender-restricted father/mother, sibling via shared parent + `\=`.
-- [x] `lib/prolog/conformance.sh` + runner, `scoreboard.json` + `scoreboard.md` — bash script feeds load + eval epoch script to sx_server, parses each suite's `{:failed N :passed N :total N :failures (...)}` line, writes JSON (machine) + MD (human) scoreboards. Exit non-zero on any failure. `SX_SERVER` env var overrides binary path. First scoreboard: 183 / 183.
-- [x] Target: all 5 classic programs passing — append (6) + reverse (6) + member (7) + nqueens (6) + family (10) = 35 program tests, all green. Phase 3 architecturally complete bar the conformance harness/scoreboard.
+  - [ ] `append.pl` — list append (with backtracking)
+  - [ ] `reverse.pl` — naive reverse
+  - [ ] `member.pl` — generate all solutions via backtracking
+  - [ ] `nqueens.pl` — 8-queens
+  - [ ] `family.pl` — facts + rules (parent/ancestor)
+- [ ] `lib/prolog/conformance.sh` + runner, `scoreboard.json` + `scoreboard.md`
+- [ ] Target: all 5 classic programs passing
 
 ### Phase 4 — operator table + more built-ins (next run)
-- [x] Operator table parsing (prefix/infix/postfix, precedence, assoc) — `pl-op-table` (15 entries: `, ; -> = \= is < > =< >= + - * / mod`); precedence-climbing parser via `pp-parse-primary` + `pp-parse-term-prec` + `pp-parse-op-rhs`. Parens override precedence. Args inside compounds parsed at 999 so `,` stays as separator. xfx/xfy/yfx supported; prefix/postfix deferred (so `-5` still tokenises as bare atom + num as before). Comparison built-ins `</2 >/2 =</2 >=/2` added. New `tests/operators.sx` 19 tests cover assoc/precedence/parens + solver via infix.
-- [x] `assert/1`, `asserta/1`, `assertz/1`, `retract/1` — `assert` aliases `assertz`. Helpers `pl-rt-to-ast` (deep-walk + replace runtime vars with `_G<id>` parse markers) + `pl-build-clause` (detect `:-` head). `assertz` uses `pl-db-add!`; `asserta` uses new `pl-db-prepend!`. `retract` walks goal, looks up by functor/arity, tries each clause via unification, removes first match by index (`pl-list-without`). 11 tests in `tests/dynamic.sx`. Rule-asserts deferred — `:-` not in op table yet, so only fact-shaped clauses for now.
+- [ ] Operator table parsing (prefix/infix/postfix, precedence, assoc)
+- [ ] `assert/1`, `asserta/1`, `assertz/1`, `retract/1`
 - [ ] `findall/3`, `bagof/3`, `setof/3`
 - [ ] `copy_term/2`, `functor/3`, `arg/3`, `=../2`
 - [ ] String/atom predicates
@@ -88,22 +88,6 @@ Representation choices (finalise in phase 1, document here):
 
 _Newest first. Agent appends on every commit._
 
-- 2026-04-25 — Dynamic clauses: `assert/1`, `assertz/1`, `asserta/1`, `retract/1`. New helpers `pl-rt-to-ast` (deep-walk runtime term → parse-AST, mapping unbound runtime vars to `_G<id>` markers so `pl-instantiate-fresh` produces fresh vars per call) + `pl-build-clause` + `pl-db-prepend!` + `pl-list-without`. `retract` keeps runtime vars (so the caller's vars get bound), walks head for the functor/arity key, tries each stored clause via `pl-unify!`, removes the first match by index. 11 tests in `tests/dynamic.sx`; conformance script gained dynamic row. Total **213** (+11). Rule-form asserts (`(H :- B)`) deferred until `:-` is in the op table.
-- 2026-04-25 — Phase 4 starts: operator-table parsing. Parser rewrite uses precedence climbing (xfx/xfy/yfx); 15-op table covers control (`, ; ->`), comparison (`= \\= is < > =< >=`), arithmetic (`+ - * / mod`). Parens override. Backwards-compatible: prefix-syntax compounds (`=(X, Y)`, `+(2, 3)`) still parse as before; existing 183 tests untouched. Added comparison built-ins `</2 >/2 =</2 >=/2` to runtime (eval both sides, compare). New `tests/operators.sx` 19 tests; conformance script gained an operators row. Total **202** (+19). Prefix/postfix deferred — `-5` keeps old bare-atom semantics.
-- 2026-04-25 — Conformance harness landed. `lib/prolog/conformance.sh` runs all 9 suites in one sx_server epoch, parses the `{:failed/:passed/:total/:failures}` summary lines, and writes `scoreboard.json` + `scoreboard.md`. `SX_SERVER` env var overrides the binary path; default points at the main-repo build. Phase 3 fully complete: 183 / 183 passing across parse/unify/clausedb/solve/append/reverse/member/nqueens/family.
-- 2026-04-25 — `family.pl` fifth classic program — completes the 5-program target. 5-fact pedigree + male/female + derived father/mother/ancestor/sibling. 10 tests cover fact lookup + count, transitive ancestor through 3 generations, descendant counting (5), gender-restricted derivations, sibling via shared parent guarded by `\=`. Total 183 (+10). All 5 classic programs ticked; Phase 3 needs only conformance harness + scoreboard left.
-- 2026-04-25 — `nqueens.pl` fourth classic program. Permute-and-test variant exercises every Phase-3 feature: lists with `[H|T]` cons sugar, multi-clause backtracking, recursive `permute`/`select`/`safe`/`no_attack`, `is/2` arithmetic on diagonals, `\=/2` for diagonal-conflict check. 6 tests at N ∈ {1,2,3,4,5} with expected counts {1,0,0,2,10} + first-solution `[2,4,1,3]`. N=5 takes ~30s (120 perms × safe-check); N=8 omitted as it would be ~thousands of seconds. Total 173 (+6).
-- 2026-04-25 — `member.pl` third classic program. Standard 2-clause definition; 7 tests cover bound-element hit/miss, empty-list fail, generator-count = list length, first-solution binding (X=11), duplicate elements matched twice on backtrack, anonymous-head unification (`member(a, [X, b, c])` binds X=a). Total 167 (+7).
-- 2026-04-25 — `reverse.pl` second classic program. Naive reverse defined via append. 6 tests (empty/singleton/3-list/4-atom-list/ground match/ground mismatch). Confirms the solver handles non-trivial recursive composition: `reverse([1,2,3], R)` recurses to depth 3 then unwinds via 3 nested `append`s. Total 160 (+6).
-- 2026-04-25 — `append.pl` first classic program. `lib/prolog/tests/programs/append.pl` is the canonical 2-clause source; `append.sx` embeds the source as a string (no file-read primitive in SX yet) and runs 6 tests covering build, check, full split-backtrack (4 solutions), and deduction modes. Helpers `pl-ap-list-to-sx` / `pl-ap-term-to-sx` convert deep-walked Prolog lists (`("compound" "." (h t))` / `("atom" "[]")`) to SX lists for structural assertion. Total 154 (+6).
-- 2026-04-25 — `is/2` arithmetic landed. `pl-eval-arith` recursively evaluates ground RHS expressions (binary `+ - * /`, `mod`; binary+unary `-`; unary `abs`); `is/2` wraps the value as `("num" v)` and unifies via `pl-solve-eq!`, so it works in all three modes — bind unbound LHS, check ground LHS for equality, propagate from earlier var bindings on RHS. 11 tests, total 148 (+11). Without operator support, expressions must be written prefix: `is(X, +(2, *(3, 4)))`.
-- 2026-04-25 — `write/1` + `nl/0` landed using global string buffer (`pl-output-buffer` + `pl-output-clear!` + `pl-output-write!`). `pl-format-term` walks deep + dispatches on atom/num/str/compound/var; `pl-format-args` recursively comma-joins. 7 new tests cover atom/num/compound formatting, conjunction order, var-walk, and `nl`. Built-ins box (`=/2`, `\\=/2`, `true/0`, `fail/0`, `!/0`, `,/2`, `;/2`, `->/2`, `call/1`, `write/1`, `nl/0`) now ticked. Total 137 (+7).
-- 2026-04-25 — `->/2` if-then-else landed (both `;(->(C,T), E)` and standalone `->(C, T)` ≡ `(C -> T ; fail)`). `pl-solve-or!` now special-cases `->` in left arg → `pl-solve-if-then-else!`. Cond runs in a fresh local cut-box (ISO opacity for cut inside cond). Then-branch can backtrack, else-branch can backtrack, but cond commits to first solution. 9 new tests covering both forms, both branches, binding visibility, cond-commit, then-backtrack, else-backtrack. Total 130 (+9).
-- 2026-04-25 — Built-ins `\=/2`, `;/2`, `call/1` landed. `pl-solve-not-eq!` (try unify, always undo, succeed iff unify failed). `pl-solve-or!` (try left, on failure check cut and only try right if not cut). `call/1` opens a fresh inner cut-box (ISO opacity: cut inside `call(G)` commits G, not caller). 11 new tests in `tests/solve.sx` cover atoms+vars for `\=`, both branches + count for `;`, and `call/1` against atoms / compounds / bound goal vars. Total 121 (+11). Box not yet ticked — `->/2`, `write/1`, `nl/0` still pending.
-- 2026-04-25 — Cut (`!/0`) landed. `pl-cut?` predicate; solver functions all take a `cut-box`; `pl-solve-user!` creates a fresh inner-cut-box and snapshots `outer-was-cut`; `pl-try-clauses!` abandons alternatives when inner.cut OR (outer.cut transitioned false→true during this call). 6 new cut tests in `tests/solve.sx` covering bare cut, clause-commit, choice-commit, cut+fail blocks alt clauses, post-cut goal backtracks freely, inner cut isolation. Total 110 (+6).
-- 2026-04-25 — Phase 3 DFS solver landed (CPS, trail-based backtracking; delimited conts deferred). `pl-solve!` + `pl-solve-eq!` + `pl-solve-user!` + `pl-try-clauses!` + `pl-solve-once!` + `pl-solve-count!` in runtime.sx. Built-ins: `true/0`, `fail/0`, `=/2`, `,/2`. New `tests/solve.sx` 18/18 green covers atomic goals, =, conjunction, fact lookup, multi-solution count, recursive ancestor rule, trail-undo verification. Bug fix: `pl-instantiate` had no `("clause" h b)` case → vars in rule head/body were never instantiated, so rule resolution silently failed against runtime-var goals. Added clause case to recurse with shared var-env. Total 104 (+18).
-- 2026-04-24 — Phase 3 clause DB landed: `pl-mk-db` + `pl-head-key` / `pl-clause-key` / `pl-goal-key` + `pl-db-add!` / `pl-db-load!` / `pl-db-lookup` / `pl-db-lookup-goal` in runtime.sx. New `tests/clausedb.sx` 14/14 green. Total 86 (+14). Loader preserves declaration order (append!).
-- 2026-04-24 — Verified phase 1+2 already implemented on loops/prolog: `pl-parse-tests-run!` 25/25, `pl-unify-tests-run!` 47/47 (72 total). Ticked phase 1+2 boxes.
 - _(awaiting phase 1)_
 
 ## Blockers