Commit | Line | Data |
---|---|---|
03363afd YO |
1 | # regcomp.sym |
2 | # | |
3 | # File has two sections, divided by a line of dashes '-'. | |
4 | # | |
5 | # Empty rows after #-comment are removed from input are ignored | |
6 | # | |
7 | # First section is for regops, second sectionis for regmatch-states | |
8 | # | |
3dab1dad YO |
9 | # Note that the order in this file is important. |
10 | # | |
03363afd YO |
11 | # Format for first section: |
12 | # NAME \t TYPE, arg-description [num-args] [longjump-len] \t DESCRIPTION | |
13 | # | |
3dab1dad | 14 | # |
c476f425 | 15 | # run perl regen.pl after editing this file |
3dab1dad | 16 | |
03363afd YO |
17 | |
18 | ||
1de06328 YO |
19 | #* Exit points (0,1) |
20 | ||
d09b2d29 IZ |
21 | END END, no End of program. |
22 | SUCCEED END, no Return from a subroutine, basically. | |
23 | ||
1de06328 YO |
24 | #* Anchors: (2..13) |
25 | ||
d09b2d29 IZ |
26 | BOL BOL, no Match "" at beginning of line. |
27 | MBOL BOL, no Same, assuming multiline. | |
28 | SBOL BOL, no Same, assuming singleline. | |
b85d18e9 | 29 | EOS EOL, no Match "" at end of string. |
d09b2d29 IZ |
30 | EOL EOL, no Match "" at end of line. |
31 | MEOL EOL, no Same, assuming multiline. | |
32 | SEOL EOL, no Same, assuming singleline. | |
33 | BOUND BOUND, no Match "" at any word boundary | |
34 | BOUNDL BOUND, no Match "" at any word boundary | |
35 | NBOUND NBOUND, no Match "" at any word non-boundary | |
36 | NBOUNDL NBOUND, no Match "" at any word non-boundary | |
37 | GPOS GPOS, no Matches where last m//g left off. | |
38 | ||
1de06328 YO |
39 | #* [Special] alternatives: (14..30) |
40 | ||
22c35a8c | 41 | REG_ANY REG_ANY, no Match any one character (except newline). |
22c35a8c | 42 | SANY REG_ANY, no Match any one character. |
f33976b4 | 43 | CANY REG_ANY, no Match any one byte. |
d09b2d29 IZ |
44 | ANYOF ANYOF, sv Match character in (or not in) this class. |
45 | ALNUM ALNUM, no Match any alphanumeric character | |
46 | ALNUML ALNUM, no Match any alphanumeric char in locale | |
47 | NALNUM NALNUM, no Match any non-alphanumeric character | |
48 | NALNUML NALNUM, no Match any non-alphanumeric char in locale | |
49 | SPACE SPACE, no Match any whitespace character | |
50 | SPACEL SPACE, no Match any whitespace char in locale | |
51 | NSPACE NSPACE, no Match any non-whitespace character | |
52 | NSPACEL NSPACE, no Match any non-whitespace char in locale | |
53 | DIGIT DIGIT, no Match any numeric character | |
b8c5462f | 54 | DIGITL DIGIT, no Match any numeric character in locale |
d09b2d29 | 55 | NDIGIT NDIGIT, no Match any non-numeric character |
b8c5462f | 56 | NDIGITL NDIGIT, no Match any non-numeric character in locale |
a0ed51b3 | 57 | CLUMP CLUMP, no Match any combining character sequence |
d09b2d29 | 58 | |
1de06328 YO |
59 | #* Alternation (31) |
60 | ||
61 | # BRANCH The set of branches constituting a single choice are hooked | |
d09b2d29 IZ |
62 | # together with their "next" pointers, since precedence prevents |
63 | # anything being concatenated to any individual branch. The | |
64 | # "next" pointer of the last BRANCH in a choice points to the | |
65 | # thing following the whole choice. This is also where the | |
66 | # final "next" pointer of each individual branch points; each | |
67 | # branch starts with the operand node of a BRANCH node. | |
68 | # | |
69 | BRANCH BRANCH, node Match this alternative, or the next... | |
70 | ||
1de06328 YO |
71 | #*Back pointer (32) |
72 | ||
d09b2d29 IZ |
73 | # BACK Normal "next" pointers all implicitly point forward; BACK |
74 | # exists to make loop structures possible. | |
75 | # not used | |
76 | BACK BACK, no Match "", "next" ptr points backward. | |
77 | ||
1de06328 YO |
78 | #*Literals (33..35) |
79 | ||
d09b2d29 IZ |
80 | EXACT EXACT, sv Match this string (preceded by length). |
81 | EXACTF EXACT, sv Match this string, folded (prec. by length). | |
82 | EXACTFL EXACT, sv Match this string, folded in locale (w/len). | |
83 | ||
1de06328 YO |
84 | #*Do nothing types (36..37) |
85 | ||
d09b2d29 IZ |
86 | NOTHING NOTHING,no Match empty string. |
87 | # A variant of above which delimits a group, thus stops optimizations | |
88 | TAIL NOTHING,no Match empty string. Can jump here from outside. | |
89 | ||
1de06328 YO |
90 | #*Loops (38..44) |
91 | ||
d09b2d29 IZ |
92 | # STAR,PLUS '?', and complex '*' and '+', are implemented as circular |
93 | # BRANCH structures using BACK. Simple cases (one character | |
94 | # per match) are implemented with STAR and PLUS for speed | |
95 | # and to minimize recursive plunges. | |
96 | # | |
97 | STAR STAR, node Match this (simple) thing 0 or more times. | |
98 | PLUS PLUS, node Match this (simple) thing 1 or more times. | |
99 | ||
100 | CURLY CURLY, sv 2 Match this simple thing {n,m} times. | |
101 | CURLYN CURLY, no 2 Match next-after-this simple thing | |
102 | # {n,m} times, set parenths. | |
103 | CURLYM CURLY, no 2 Match this medium-complex thing {n,m} times. | |
104 | CURLYX CURLY, sv 2 Match this complex thing {n,m} times. | |
105 | ||
106 | # This terminator creates a loop structure for CURLYX | |
107 | WHILEM WHILEM, no Do curly processing and see if rest matches. | |
108 | ||
1de06328 YO |
109 | #*Buffer related (45..49) |
110 | ||
d09b2d29 IZ |
111 | # OPEN,CLOSE,GROUPP ...are numbered at compile time. |
112 | OPEN OPEN, num 1 Mark this point in input as start of #n. | |
113 | CLOSE CLOSE, num 1 Analogous to OPEN. | |
114 | ||
115 | REF REF, num 1 Match some already matched string | |
116 | REFF REF, num 1 Match already matched string, folded | |
117 | REFFL REF, num 1 Match already matched string, folded in loc. | |
118 | ||
1de06328 YO |
119 | #*Grouping assertions (50..54) |
120 | ||
d09b2d29 IZ |
121 | IFMATCH BRANCHJ,off 1 2 Succeeds if the following matches. |
122 | UNLESSM BRANCHJ,off 1 2 Fails if the following matches. | |
123 | SUSPEND BRANCHJ,off 1 1 "Independent" sub-RE. | |
124 | IFTHEN BRANCHJ,off 1 1 Switch, should be preceeded by switcher . | |
125 | GROUPP GROUPP, num 1 Whether the group matched. | |
126 | ||
1de06328 YO |
127 | #*Support for long RE (55..56) |
128 | ||
d09b2d29 IZ |
129 | LONGJMP LONGJMP,off 1 1 Jump far away. |
130 | BRANCHJ BRANCHJ,off 1 1 BRANCH with long offset. | |
131 | ||
1de06328 YO |
132 | #*The heavy worker (57..58) |
133 | ||
d09b2d29 IZ |
134 | EVAL EVAL, evl 1 Execute some Perl code. |
135 | ||
1de06328 YO |
136 | #*Modifiers (59..60) |
137 | ||
d09b2d29 IZ |
138 | MINMOD MINMOD, no Next operator is not greedy. |
139 | LOGICAL LOGICAL,no Next opcode should set the flag only. | |
140 | ||
1de06328 | 141 | # This is not used yet (61) |
d09b2d29 IZ |
142 | RENUM BRANCHJ,off 1 1 Group with independently numbered parens. |
143 | ||
1de06328 YO |
144 | #*Trie Related (62..64) |
145 | ||
146 | # Behave the same as A|LIST|OF|WORDS would. The '..C' variants have | |
147 | # inline charclass data (ascii only), the 'C' store it in the structure. | |
148 | # NOTE: the relative order of the TRIE-like regops is signifigant | |
ce5e9471 | 149 | |
3dab1dad | 150 | TRIE TRIE, trie 1 Match many EXACT(FL?)? at once. flags==type |
786e8c11 | 151 | TRIEC TRIE, trie charclass Same as TRIE, but with embedded charclass data |
3dab1dad | 152 | |
1de06328 YO |
153 | # For start classes, contains an added fail table. |
154 | AHOCORASICK TRIE, trie 1 Aho Corasick stclass. flags==type | |
155 | AHOCORASICKC TRIE, trie charclass Same as AHOCORASICK, but with embedded charclass data | |
156 | ||
6bda09f9 YO |
157 | #*Recursion (65) |
158 | RECURSE RECURSE, num/ofs 2L recurse to paren arg1 at (signed) ofs arg2 | |
159 | SRECURSE RECURSE, no recurse to start of pattern | |
03363afd | 160 | |
1de06328 YO |
161 | # NEW STUFF ABOVE THIS LINE -- Please update counts below. |
162 | ||
03363afd YO |
163 | ################################################################################ |
164 | ||
165 | #*SPECIAL REGOPS (65, 66) | |
1de06328 YO |
166 | |
167 | # This is not really a node, but an optimized away piece of a "long" node. | |
168 | # To simplify debugging output, we mark it as if it were a node | |
169 | OPTIMIZED NOTHING,off Placeholder for dump. | |
170 | ||
3dab1dad YO |
171 | # Special opcode with the property that no opcode in a compiled program |
172 | # will ever be of this type. Thus it can be used as a flag value that | |
173 | # no other opcode has been seen. END is used similarly, in that an END | |
174 | # node cant be optimized. So END implies "unoptimizable" and PSEUDO mean | |
175 | # "not seen anything to optimize yet". | |
176 | PSEUDO PSEUDO,off Pseudo opcode for internal use. | |
1de06328 | 177 | |
03363afd YO |
178 | ------------------------------------------------------------------------------- |
179 | # Format for second section: | |
180 | # REGOP \t typelist [ \t typelist] [# Comment] | |
181 | # typelist= namelist | |
182 | # = namelist:FAIL | |
183 | # = name:count | |
184 | ||
185 | # Anything below is a state | |
186 | # | |
187 | # | |
188 | TRIE next:FAIL | |
189 | EVAL AB:FAIL | |
c476f425 DM |
190 | CURLYX end:FAIL |
191 | WHILEM A_pre,A_min,A_max,B_min,B_max:FAIL | |
03363afd YO |
192 | BRANCH next:FAIL |
193 | CURLYM A,B:FAIL | |
194 | IFMATCH A:FAIL | |
195 | CURLY B_min_known,B_min,B_max:FAIL | |
196 |