Annotation of gforth/hash.fs, revision 1.33

1.1       pazsan      1: \ Hashed dictionaries                                  15jul94py
                      2: 
1.33    ! anton       3: \ Copyright (C) 1995,1998,2000,2003 Free Software Foundation, Inc.
1.10      anton       4: 
                      5: \ This file is part of Gforth.
                      6: 
                      7: \ Gforth is free software; you can redistribute it and/or
                      8: \ modify it under the terms of the GNU General Public License
                      9: \ as published by the Free Software Foundation; either version 2
                     10: \ of the License, or (at your option) any later version.
                     11: 
                     12: \ This program is distributed in the hope that it will be useful,
                     13: \ but WITHOUT ANY WARRANTY; without even the implied warranty of
                     14: \ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
                     15: \ GNU General Public License for more details.
                     16: 
                     17: \ You should have received a copy of the GNU General Public License
                     18: \ along with this program; if not, write to the Free Software
1.26      anton      19: \ Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
1.10      anton      20: 
1.28      jwilke     21: [IFUNDEF] erase
                     22: : erase ( addr len -- ) 0 fill ;
                     23: [THEN]
                     24: 
1.19      jwilke     25: [IFUNDEF] allocate
1.18      jwilke     26: : reserve-mem here swap allot ;
                     27: \ move to a kernel/memory.fs
                     28: [ELSE]
                     29: : reserve-mem allocate throw ;
                     30: [THEN]
                     31: 
                     32: [IFUNDEF] hashbits
1.19      jwilke     33: 11 Value hashbits
1.18      jwilke     34: [THEN]
1.2       anton      35: 1 hashbits lshift Value Hashlen
1.1       pazsan     36: 
1.18      jwilke     37: \ compute hash key                                     15jul94py
                     38: 
1.28      jwilke     39: has? ec [IF] [IFUNDEF] hash
                     40: : hash ( addr len -- key )
                     41:   over c@ swap 1- IF swap char+ c@ + ELSE nip THEN
                     42:   [ Hashlen 1- ] literal and ;
                     43: [THEN] [THEN]
                     44: 
1.18      jwilke     45: [IFUNDEF] hash
                     46: : hash ( addr len -- key )
                     47:     hashbits (hashkey1) ;
                     48: [THEN]
                     49: 
1.1       pazsan     50: Variable insRule        insRule on
1.4       pazsan     51: Variable revealed
1.1       pazsan     52: 
1.4       pazsan     53: \ Memory handling                                      10oct94py
1.1       pazsan     54: 
1.28      jwilke     55: AVariable HashPointer
1.29      pazsan     56: Variable HashIndex     \ Number of wordlists
                     57: Variable HashPop       \ Number of words
1.28      jwilke     58: 0 AValue HashTable
1.1       pazsan     59: 
1.18      jwilke     60: \ forward declarations
1.28      jwilke     61: 0 AValue hashsearch-map
1.23      anton      62: Defer hash-alloc ( addr -- addr )
1.18      jwilke     63: 
1.4       pazsan     64: \ DelFix and NewFix are from bigFORTH                  15jul94py
1.1       pazsan     65: 
                     66: : DelFix ( addr root -- ) dup @ 2 pick ! ! ;
                     67: : NewFix  ( root len # -- addr )
1.18      jwilke     68:   BEGIN  2 pick @ ?dup  0= WHILE  2dup * reserve-mem
1.1       pazsan     69:          over 0 ?DO  dup 4 pick DelFix 2 pick +  LOOP  drop
                     70:   REPEAT  >r drop r@ @ rot ! r@ swap erase r> ;
                     71: 
1.12      anton      72: : bucket ( addr len wordlist -- bucket-addr )
                     73:     \ @var{bucket-addr} is the address of a cell that points to the first
                     74:     \ element in the list of the bucket for the string @var{addr len}
                     75:     wordlist-extend @ -rot hash xor ( bucket# )
1.13      anton      76:     cells HashTable + ;
1.2       anton      77: 
                     78: : hash-find ( addr len wordlist -- nfa / false )
1.27      anton      79:     >r 2dup r> bucket @ (hashlfind) ;
1.1       pazsan     80: 
                     81: \ hash vocabularies                                    16jul94py
                     82: 
                     83: : lastlink! ( addr link -- )
                     84:   BEGIN  dup @ dup  WHILE  nip  REPEAT  drop ! ;
                     85: 
1.14      anton      86: : (reveal ( nfa wid -- )
1.12      anton      87:     over name>string rot bucket >r
                     88:     HashPointer 2 Cells $400 NewFix
                     89:     tuck cell+ ! r> insRule @
                     90:     IF
                     91:        dup @ 2 pick ! !
                     92:     ELSE
                     93:        lastlink!
                     94:     THEN
1.29      pazsan     95:     revealed on 1 HashPop +! 0 hash-alloc drop ;
1.12      anton      96: 
1.14      anton      97: : hash-reveal ( nfa wid -- )
                     98:     2dup (reveal) (reveal ;
1.1       pazsan     99: 
1.18      jwilke    100: : inithash ( wid -- )
                    101:     wordlist-extend
1.29      pazsan    102:     insRule @ >r  insRule off  1 hash-alloc over ! 3 cells -
1.21      pazsan    103:     dup wordlist-id
1.18      jwilke    104:     BEGIN  @ dup  WHILE  2dup swap (reveal  REPEAT
                    105:     2drop  r> insRule ! ;
                    106: 
1.4       pazsan    107: : addall  ( -- )
1.29      pazsan    108:     HashPop off voclink
1.18      jwilke    109:     BEGIN  @ dup WHILE
                    110:           dup 0 wordlist-link -
1.24      pazsan    111:           dup wordlist-map @ reveal-method @ ['] hash-reveal = 
1.18      jwilke    112:           IF  inithash ELSE drop THEN
                    113:     REPEAT  drop ;
1.4       pazsan    114: 
                    115: : clearhash  ( -- )
1.13      anton     116:     HashTable Hashlen cells bounds
1.4       pazsan    117:     DO  I @
1.15      pazsan    118:        BEGIN  dup  WHILE
1.23      anton     119:            dup @ swap HashPointer DelFix
                    120:        REPEAT
                    121:        I !
                    122:        cell +LOOP
                    123:     HashIndex off 
1.18      jwilke    124:     voclink
1.23      anton     125:     BEGIN ( wordlist-link-addr )
                    126:        @ dup
                    127:     WHILE ( wordlist-link )
                    128:        dup 0 wordlist-link - ( wordlist-link wid ) 
                    129:        dup wordlist-map @ hashsearch-map = 
                    130:        IF ( wordlist-link wid )
                    131:            0 swap wordlist-extend !
                    132:        ELSE
                    133:            drop
                    134:        THEN
                    135:     REPEAT
                    136:     drop ;
1.18      jwilke    137: 
                    138: : rehashall  ( wid -- ) 
                    139:   drop revealed @ 
                    140:   IF   clearhash addall revealed off 
                    141:   THEN ;
1.4       pazsan    142: 
1.18      jwilke    143: : (rehash)   ( wid -- )
                    144:   dup wordlist-extend @ 0=
                    145:   IF   inithash
                    146:   ELSE rehashall THEN ;
                    147: 
1.29      pazsan    148: : hashdouble ( -- )
                    149:     HashTable >r clearhash
                    150:     1 hashbits 1+ dup  to hashbits  lshift  to hashlen
                    151:     r> free >r  0 to HashTable
                    152:     addall r> throw ;
                    153: 
1.28      jwilke    154: const Create (hashsearch-map)
                    155: ' hash-find A, ' hash-reveal A, ' (rehash) A, ' (rehash) A,
                    156: (hashsearch-map) to hashsearch-map
1.4       pazsan    157: 
                    158: \ hash allocate and vocabulary initialization          10oct94py
                    159: 
1.29      pazsan    160: :noname ( n+ -- n )
1.18      jwilke    161:   HashTable 0= 
                    162:   IF  Hashlen cells reserve-mem TO HashTable
                    163:       HashTable Hashlen cells erase THEN
1.29      pazsan    164:   HashIndex @ swap HashIndex +!
1.4       pazsan    165:   HashIndex @ Hashlen >=
1.19      jwilke    166:   [ [IFUNDEF] allocate ]
1.18      jwilke    167:   ABORT" no more space in hashtable"
                    168:   [ [ELSE] ]
1.30      anton     169:   HashPop @ hashlen 2* >= or
1.29      pazsan    170:   IF  hashdouble  THEN 
1.18      jwilke    171:   [ [THEN] ] ; is hash-alloc
1.1       pazsan    172: 
                    173: \ Hash-Find                                            01jan93py
1.19      jwilke    174: has? cross 0= 
1.18      jwilke    175: [IF]
1.16      pazsan    176: : make-hash
1.21      pazsan    177:   hashsearch-map forth-wordlist wordlist-map !
1.18      jwilke    178:   addall ;
                    179:   make-hash \ Baumsuche ist installiert.
                    180: [ELSE]
1.21      pazsan    181:   hashsearch-map forth-wordlist wordlist-map !
1.18      jwilke    182: [THEN]
1.16      pazsan    183: 
1.18      jwilke    184: \ for ec version display that vocabulary goes hashed
1.1       pazsan    185: 
1.18      jwilke    186: : hash-cold  ( -- )
1.19      jwilke    187: [ has? ec [IF] ] ." Hashing..." [ [THEN] ]
1.13      anton     188:   HashPointer off  0 TO HashTable  HashIndex off
1.18      jwilke    189:   addall
                    190: \  voclink
                    191: \  BEGIN  @ dup WHILE
                    192: \         dup 0 wordlist-link - initvoc
                    193: \  REPEAT  drop 
1.19      jwilke    194: [ has? ec [IF] ] ." Done" cr [ [THEN] ] ;
1.18      jwilke    195: 
                    196: ' hash-cold INIT8 chained
1.5       pazsan    197: 
1.1       pazsan    198: : .words  ( -- )
1.13      anton     199:   base @ >r hex HashTable  Hashlen 0
1.4       pazsan    200:   DO  cr  i 2 .r ." : " dup i cells +
1.1       pazsan    201:       BEGIN  @ dup  WHILE
1.20      pazsan    202:              dup cell+ @ name>string type space  REPEAT  drop
1.1       pazsan    203:   LOOP  drop r> base ! ;
                    204: 
1.2       anton     205: \ \ this stuff is for evaluating the hash function
                    206: \ : square dup * ;
                    207: 
                    208: \ : countwl  ( -- sum sumsq )
1.4       pazsan    209: \     \ gives the number of words in the current wordlist
                    210: \     \ and the sum of squares for the sublist lengths
1.2       anton     211: \     0 0
1.13      anton     212: \     hashtable Hashlen cells bounds DO
1.4       pazsan    213: \        0 i BEGIN
                    214: \            @ dup WHILE
                    215: \            swap 1+ swap
                    216: \        REPEAT
                    217: \        drop
                    218: \        swap over square +
                    219: \        >r + r>
                    220: \        1 cells
                    221: \    +LOOP ;
1.2       anton     222: 
                    223: \ : chisq ( -- n )
1.4       pazsan    224: \     \ n should have about the same size as Hashlen
                    225: \     countwl Hashlen 2 pick */ swap - ;

FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>