Annotation of gforth/regexp.fs, revision 1.24
1.1 pazsan 1: \ Regexp compile
2:
1.13 anton 3: \ Copyright (C) 2005,2006,2007,2008 Free Software Foundation, Inc.
1.3 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
1.10 anton 9: \ as published by the Free Software Foundation, either version 3
1.3 anton 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
1.10 anton 18: \ along with this program. If not, see http://www.gnu.org/licenses/.
1.3 anton 19:
1.1 pazsan 20: \ The idea of the parser is the following:
21: \ As long as there's a match, continue
22: \ On a mismatch, LEAVE.
23: \ Insert appropriate control structures on alternative branches
24: \ Keep the old pointer (backtracking) on the stack
25: \ I try to keep the syntax as close to a real regexp system as possible
26: \ All regexp stuff is compiled into one function as forward branching
27: \ state machine
28:
1.2 pazsan 29: \ special control structure
1.1 pazsan 30:
1.8 pazsan 31: : FORK ( compilation -- orig ; run-time f -- ) \ gforth
32: \G AHEAD-like control structure: calls the code after JOIN.
1.1 pazsan 33: POSTPONE call >mark ; immediate restrict
1.8 pazsan 34: : JOIN ( orig -- ) \ gforth
35: \G THEN-like control structure for FORK
36: postpone THEN ; immediate restrict
1.1 pazsan 37:
38: \ Charclasses
39:
40: : +bit ( addr n -- ) + 1 swap c! ;
41: : -bit ( addr n -- ) + 0 swap c! ;
42: : @+ ( addr -- n addr' ) dup @ swap cell+ ;
43:
44: 0 Value cur-class
1.8 pazsan 45: : charclass ( -- ) \ regexp-cg
46: \G Create a charclass
47: Create here dup to cur-class $100 dup allot erase ;
48: : +char ( char -- ) \ regexp-cg
49: \G add a char to the current charclass
50: cur-class swap +bit ;
51: : -char ( char -- ) \ regexp-cg
52: \G remove a char from the current charclass
53: cur-class swap -bit ;
54: : ..char ( start end -- ) \ regexp-cg
55: \G add a range of chars to the current charclass
56: 1+ swap ?DO I +char LOOP ;
1.1 pazsan 57: : or! ( n addr -- ) dup @ rot or swap ! ;
58: : and! ( n addr -- ) dup @ rot and swap ! ;
1.8 pazsan 59: : +class ( class -- ) \ regexp-cg
60: \G union of charclass @var{class} and the current charclass
61: $100 0 ?DO @+ swap
62: cur-class I + or! cell +LOOP drop ;
63: : -class ( class -- ) \ regexp-cg
64: \G subtract the charclass @var{class} from the current charclass
65: $100 0 ?DO @+ swap invert
66: cur-class I + and! cell +LOOP drop ;
1.1 pazsan 67:
68: : char? ( addr class -- addr' flag )
69: >r count r> + c@ ;
70:
71: \ Charclass tests
72:
1.8 pazsan 73: : c? ( addr class -- ) \ regexp-pattern
74: \G check @var{addr} for membership in charclass @var{class}
75: ]] char? 0= ?LEAVE [[ ; immediate
76: : -c? ( addr class -- ) \ regexp-pattern
77: \G check @var{addr} for not membership in charclass @var{class}
78: ]] char? ?LEAVE [[ ; immediate
1.1 pazsan 79:
80: charclass digit '0 '9 ..char
81: charclass blanks 0 bl ..char
82: \ bl +char #tab +char #cr +char #lf +char ctrl L +char
83: charclass letter 'a 'z ..char 'A 'Z ..char
84: charclass any 0 $FF ..char #lf -char
85:
1.8 pazsan 86: : \d ( addr -- addr' ) \ regexp-pattern
87: \G check for digit
88: ]] digit c? [[ ; immediate
89: : \s ( addr -- addr' ) \ regexp-pattern
90: \G check for blanks
91: ]] blanks c? [[ ; immediate
92: : .? ( addr -- addr' ) \ regexp-pattern
93: \G check for any single charachter
94: ]] any c? [[ ; immediate
95: : -\d ( addr -- addr' ) \ regexp-pattern
96: \G check for not digit
97: ]] digit -c? [[ ; immediate
98: : -\s ( addr -- addr' ) \ regexp-pattern
99: \G check for not blank
100: ]] blanks -c? [[ ; immediate
101: : ` ( "char" -- ) \ regexp-pattern
102: \G check for particular char
1.1 pazsan 103: ]] count [[ char ]] Literal <> ?LEAVE [[ ; immediate
1.17 pazsan 104: : -` ( "char" -- ) \ regexp-pattern
105: \G check for particular char
106: ]] count [[ char ]] Literal = ?LEAVE [[ ; immediate
1.1 pazsan 107:
108: \ loop stack
109:
110: Variable loops $40 3 * cells allot
111: : 3@ ( addr -- a b c ) dup >r 2 cells + @ r> 2@ ;
112: : 3! ( a b c addr -- ) dup >r 2! r> 2 cells + ! ;
113: : loops> ( -- addr ) -3 loops +! loops @+ swap cells + 3@ ;
114: : >loops ( addr -- ) loops @+ swap cells + 3! 3 loops +! ;
115: : BEGIN, ( -- ) ]] BEGIN [[ >loops ;
1.23 pazsan 116: : DONE, ( -- ) loops @ IF loops> ]] DONE [[ ELSE ." no done left!" cr THEN ;
1.1 pazsan 117:
118: \ variables
119:
120: Variable vars &18 cells allot
121: Variable varstack 9 cells allot
122: Variable varsmax
123: : >var ( -- addr ) vars @+ swap 2* cells +
124: vars @ varstack @+ swap cells + !
125: 1 vars +! 1 varstack +! ;
126: : var> ( -- addr ) -1 varstack +!
127: varstack @+ swap cells + @
128: 1+ 2* cells vars + ;
129:
130: \ start end
131:
132: 0 Value end$
133: 0 Value start$
134: : !end ( addr u -- addr ) over + to end$ dup to start$ ;
1.12 pazsan 135: : end-rex? ( addr -- addr flag ) dup end$ u< ;
136: : start-rex? ( addr -- addr flag ) dup start$ u> ;
1.1 pazsan 137: : ?end ( addr -- addr ) ]] dup end$ u> ?LEAVE [[ ; immediate
1.20 pazsan 138: : rest$ ( addr -- addr addr u ) dup end$ over - ;
1.1 pazsan 139:
140: \ start and end
141:
1.8 pazsan 142: : \^ ( addr -- addr ) \ regexp-pattern
143: \G check for string start
1.12 pazsan 144: ]] start-rex? ?LEAVE [[ ; immediate
1.8 pazsan 145: : \$ ( addr -- addr ) \ regexp-pattern
146: \G check for string end
1.12 pazsan 147: ]] end-rex? ?LEAVE [[ ; immediate
1.1 pazsan 148:
1.17 pazsan 149: \ A word for string comparison
150:
1.22 pazsan 151: : (str=?) ( addr1 addr u -- addr2 )
1.20 pazsan 152: dup >r 2>r rest$ r@ umin 2r> compare IF rdrop true ELSE r> + false THEN ;
1.22 pazsan 153: : str=? ( addr1 addr u -- addr2 ) ]] (str=?) ?LEAVE [[ ; immediate
154: : ,=" ( addr u -- ) tuck dup ]] rest$ Literal umin SLiteral compare ?LEAVE Literal + [[ ;
1.17 pazsan 155: : =" ( <string>" -- ) \ regexp-pattern
156: \G check for string
157: '" parse ,=" ; immediate
158:
1.1 pazsan 159: \ regexp block
160:
161: \ FORK/JOIN are like AHEAD THEN, but producing a call on AHEAD
162: \ instead of a jump.
163:
1.8 pazsan 164: : (( ( addr u -- ) \ regexp-pattern
165: \G start regexp block
166: vars off varsmax off loops off
1.1 pazsan 167: ]] FORK AHEAD BUT JOIN !end [[ BEGIN, ; immediate
1.8 pazsan 168: : )) ( -- addr f ) \ regexp-pattern
169: \G end regexp block
1.19 pazsan 170: ]] ?end drop true ;S [[
171: DONE, ]] drop false ;S THEN [[ ; immediate
1.1 pazsan 172:
173: \ greedy loops
174:
175: \ Idea: scan as many characters as possible, try the rest of the pattern
176: \ and then back off one pattern at a time
177:
178: : drops ( n -- ) 1+ cells sp@ + sp! ;
179:
1.8 pazsan 180: : {** ( addr -- addr addr ) \ regexp-pattern
181: \G greedy zero-or-more pattern
1.1 pazsan 182: 0 ]] Literal >r BEGIN dup [[ BEGIN, ; immediate
1.8 pazsan 183: ' {** Alias {++ ( addr -- addr addr ) \ regexp-pattern
184: \G greedy one-or-more pattern
185: immediate
186: : n*} ( sys n -- ) \ regexp-pattern
187: \G At least @var{n} pattern
1.12 pazsan 188: >r ]] r> 1+ >r end-rex? 0= UNTIL dup [[ DONE, ]] drop [[
1.1 pazsan 189: r@ ]] r> 1+ Literal U+DO FORK BUT [[
1.19 pazsan 190: ]] IF I' I - [[ r@ 1- ]] Literal + drops true UNLOOP ;S THEN LOOP [[
1.1 pazsan 191: r@ IF r@ ]] Literal drops [[ THEN
1.22 pazsan 192: rdrop ]] dup LEAVE JOIN [[ ; immediate
1.8 pazsan 193: : **} ( sys -- ) \ regexp-pattern
194: \G end of greedy zero-or-more pattern
195: 0 postpone n*} ; immediate
196: : ++} ( sys -- ) \ regexp-pattern
197: \G end of greedy zero-or-more pattern
198: 1 postpone n*} ; immediate
1.1 pazsan 199:
200: \ non-greedy loops
201:
202: \ Idea: Try to match rest of the regexp, and if that fails, try match
203: \ first expr and then try again rest of regexp.
204:
1.8 pazsan 205: : {+ ( addr -- addr addr ) \ regexp-pattern
206: \G non-greedy one-or-more pattern
1.1 pazsan 207: ]] BEGIN [[ BEGIN, ; immediate
1.8 pazsan 208: : {* ( addr -- addr addr ) \ regexp-pattern
209: \G non-greedy zero-or-more pattern
1.19 pazsan 210: ]] {+ dup FORK BUT IF drop true ;S THEN [[ ; immediate
1.8 pazsan 211: : *} ( addr addr' -- addr' ) \ regexp-pattern
212: \G end of non-greedy zero-or-more pattern
1.1 pazsan 213: ]] dup end$ u> UNTIL [[
1.19 pazsan 214: DONE, ]] drop false ;S JOIN [[ ; immediate
1.8 pazsan 215: : +} ( addr addr' -- addr' ) \ regexp-pattern
216: \G end of non-greedy one-or-more pattern
1.19 pazsan 217: ]] dup FORK BUT IF drop true ;S [[
1.23 pazsan 218: DONE, ]] drop dup LEAVE [[ BEGIN, ]] THEN *} [[ ; immediate
1.1 pazsan 219:
1.8 pazsan 220: : // ( -- ) \ regexp-pattern
221: \G search for string
222: ]] {* 1+ *} [[ ; immediate
1.1 pazsan 223:
224: \ alternatives
225:
226: \ idea: try to match one alternative and then the rest of regexp.
227: \ if that fails, jump back to second alternative
228:
1.24 ! pazsan 229: : THENs ( sys -- ) BEGIN dup WHILE ]] THEN [[ REPEAT drop ;
1.1 pazsan 230:
1.8 pazsan 231: : {{ ( addr -- addr addr ) \ regexp-pattern
232: \G Start of alternatives
1.24 ! pazsan 233: 0 ]] dup dup FORK IF 2drop true ;S BUT JOIN [[ vars @ ; immediate
1.8 pazsan 234: : || ( addr addr -- addr addr ) \ regexp-pattern
235: \G separator between alternatives
1.24 ! pazsan 236: vars @ varsmax @ max varsmax ! vars !
! 237: ]] AHEAD BUT THEN drop [[
! 238: ]] dup dup FORK IF 2drop true ;S BUT JOIN [[ vars @ ; immediate
1.21 pazsan 239: : }} ( addr addr -- addr ) \ regexp-pattern
1.8 pazsan 240: \G end of alternatives
1.24 ! pazsan 241: vars @ varsmax @ max vars ! drop
! 242: ]] AHEAD BUT THEN 2drop false ;S [[ THENs ; immediate
1.1 pazsan 243:
244: \ match variables
245:
1.8 pazsan 246: : \( ( addr -- addr ) \ regexp-pattern
247: \G start of matching variable; variables are referred as \\1--9
248: ]] dup [[
1.1 pazsan 249: >var ]] ALiteral ! [[ ; immediate
1.8 pazsan 250: : \) ( addr -- addr ) \ regexp-pattern
251: \G end of matching variable
252: ]] dup [[
1.1 pazsan 253: var> ]] ALiteral ! [[ ; immediate
1.8 pazsan 254: : \0 ( -- addr u ) \ regexp-pattern
255: \G the whole string
256: start$ end$ over - ;
1.1 pazsan 257: : \: ( i -- )
258: Create 2* 1+ cells vars + ,
259: DOES> ( -- addr u ) @ 2@ tuck - ;
260: : \:s ( n -- ) 0 ?DO I \: LOOP ;
261: 9 \:s \1 \2 \3 \4 \5 \6 \7 \8 \9
1.6 pazsan 262:
263: \ replacements, needs string.fs
264:
265: require string.fs
266:
267: 0 Value >>ptr
268: 0 Value <<ptr
269: Variable >>string
1.17 pazsan 270: : s>> ( addr -- addr ) \ regexp-replace
1.8 pazsan 271: \G Start replace pattern region
272: dup to >>ptr ;
273: : << ( run-addr addr u -- run-addr ) \ regexp-replace
274: \G Replace string from start of replace pattern region with
275: \G @var{addr} @var{u}
1.6 pazsan 276: <<ptr >>ptr over - >>string $+!
277: >>string $+! dup to <<ptr ;
1.8 pazsan 278: : <<" ( "string<">" -- ) \ regexp-replace
279: \G Replace string from start of replace pattern region with
280: \G @var{string}
281: '" parse postpone SLiteral postpone << ; immediate
1.6 pazsan 282: : >>string@ ( -- addr u )
1.16 pazsan 283: >>string $@ ;
284: : >>string0 ( addr u -- addr u ) s" " >>string $!
285: 0 to >>ptr over to <<ptr ;
1.6 pazsan 286: : >>next ( -- addr u ) <<ptr end$ over - ;
1.14 pazsan 287: : >>rest ( -- ) >>next >>string $+! ;
288: : s// ( addr u -- ptr )
1.8 pazsan 289: \G start search/replace loop
1.17 pazsan 290: ]] >>string0 (( // s>> [[ ; immediate
291: : >> ( addr -- addr )
292: ]] <<ptr >>ptr u> ?LEAVE ?end [[ ; immediate
293: : //s ( ptr -- )
294: \G search end
295: ]] )) drop >>rest >>string@ [[ ; immediate
1.15 pazsan 296: : //o ( ptr addr u -- addr' u' )
1.14 pazsan 297: \G end search/replace single loop
1.17 pazsan 298: ]] << //s [[ ; immediate
1.14 pazsan 299: : //g ( ptr addr u -- addr' u' )
300: \G end search/replace all loop
1.17 pazsan 301: ]] << LEAVE //s [[ ; immediate
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>