1# regcomp.sym 2# 3# File has two sections, divided by a line of dashes '-'. 4# 5# Lines beginning with # are ignored, except for those that start with #* 6# which are included in pod/perldebguts.pod. # within a line may be part 7# of a description. 8# 9# First section is for regops, second section is for regmatch-states 10# 11# Note that the order in this file is important. 12# 13# Format for first section: 14# NAME \s+ TYPE, arg-description [num-args] [flags] [longjump] ; DESCRIPTION 15# flag <S> means is REGNODE_SIMPLE; flag <V> means is REGNODE_VARIES; <.> is 16# a placeholder 17# longjump is 1 if the (first) argument holds the next offset. 18# 19# 20# run perl regen.pl after editing this file 21 22 23 24#* Exit points 25 26END END, no ; End of program. 27SUCCEED END, no ; Return from a subroutine, basically. 28 29#* Line Start Anchors: 30#Note flags field for SBOL indicates if it is a /^/ or a /\A/ 31SBOL BOL, no ; Match "" at beginning of line: /^/, /\A/ 32MBOL BOL, no ; Same, assuming multiline: /^/m 33 34#* Line End Anchors: 35SEOL EOL, no ; Match "" at end of line: /$/ 36MEOL EOL, no ; Same, assuming multiline: /$/m 37EOS EOL, no ; Match "" at end of string: /\z/ 38 39#* Match Start Anchors: 40GPOS GPOS, no ; Matches where last m//g left off. 41 42#* Word Boundary Opcodes: 43# The regops that have varieties that vary depending on the character set regex 44# modifiers have to ordered thusly: /d, /l, /u, /a, /aa. This is because code 45# in regcomp.c uses the enum value of the modifier as an offset from the /d 46# version. The complements must come after the non-complements. 47# BOUND, POSIX and their complements are affected, as well as EXACTF. 48BOUND BOUND, no ; Like BOUNDA for non-utf8, otherwise match "" between any Unicode \w\W or \W\w 49BOUNDL BOUND, no ; Like BOUND/BOUNDU, but \w and \W are defined by current locale 50BOUNDU BOUND, no ; Match "" at any boundary of a given type using /u rules. 51BOUNDA BOUND, no ; Match "" at any boundary between \w\W or \W\w, where \w is [_a-zA-Z0-9] 52# All NBOUND nodes are required by code in regexec.c to be greater than all BOUND ones 53NBOUND NBOUND, no ; Like NBOUNDA for non-utf8, otherwise match "" between any Unicode \w\w or \W\W 54NBOUNDL NBOUND, no ; Like NBOUND/NBOUNDU, but \w and \W are defined by current locale 55NBOUNDU NBOUND, no ; Match "" at any non-boundary of a given type using using /u rules. 56NBOUNDA NBOUND, no ; Match "" betweeen any \w\w or \W\W, where \w is [_a-zA-Z0-9] 57 58#* [Special] alternatives: 59REG_ANY REG_ANY, no 0 S ; Match any one character (except newline). 60SANY REG_ANY, no 0 S ; Match any one character. 61ANYOF ANYOF, sv charclass S ; Match character in (or not in) this class, single char match only 62ANYOFD ANYOF, sv charclass S ; Like ANYOF, but /d is in effect 63ANYOFL ANYOF, sv charclass S ; Like ANYOF, but /l is in effect 64ANYOFPOSIXL ANYOF, sv charclass_posixl S ; Like ANYOFL, but matches [[:posix:]] classes 65ANYOFH ANYOF, sv 1 S ; Like ANYOF, but only has "High" matches, none in the bitmap; non-zero flags "f" means "f" is the first UTF-8 byte shared in common by all code points matched 66ANYOFM ANYOFM byte 1 S ; Like ANYOF, but matches an invariant byte as determined by the mask and arg 67NANYOFM ANYOFM byte 1 S ; complement of ANYOFM 68 69#* POSIX Character Classes: 70# Order of the below is important. See ordering comment above. 71POSIXD POSIXD, none 0 S ; Some [[:class:]] under /d; the FLAGS field gives which one 72POSIXL POSIXD, none 0 S ; Some [[:class:]] under /l; the FLAGS field gives which one 73POSIXU POSIXD, none 0 S ; Some [[:class:]] under /u; the FLAGS field gives which one 74POSIXA POSIXD, none 0 S ; Some [[:class:]] under /a; the FLAGS field gives which one 75NPOSIXD NPOSIXD, none 0 S ; complement of POSIXD, [[:^class:]] 76NPOSIXL NPOSIXD, none 0 S ; complement of POSIXL, [[:^class:]] 77NPOSIXU NPOSIXD, none 0 S ; complement of POSIXU, [[:^class:]] 78NPOSIXA NPOSIXD, none 0 S ; complement of POSIXA, [[:^class:]] 79# End of order is important 80 81CLUMP CLUMP, no 0 V ; Match any extended grapheme cluster sequence 82 83#* Alternation 84 85#* BRANCH The set of branches constituting a single choice are 86#* hooked together with their "next" pointers, since 87#* precedence prevents anything being concatenated to 88#* any individual branch. The "next" pointer of the last 89#* BRANCH in a choice points to the thing following the 90#* whole choice. This is also where the final "next" 91#* pointer of each individual branch points; each branch 92#* starts with the operand node of a BRANCH node. 93#* 94BRANCH BRANCH, node 0 V ; Match this alternative, or the next... 95 96#*Literals 97# NOTE: the relative ordering of these types is important do not change it 98 99EXACT EXACT, str ; Match this string (preceded by length). 100EXACTL EXACT, str ; Like EXACT, but /l is in effect (used so locale-related warnings can be checked for). 101EXACTF EXACT, str ; Match this string using /id rules (w/len); (string not UTF-8, not guaranteed to be folded). 102EXACTFL EXACT, str ; Match this string using /il rules (w/len); (string not guaranteed to be folded). 103EXACTFU EXACT, str ; Match this string using /iu rules (w/len); (string folded iff in UTF-8; non-UTF8 folded length <= unfolded). 104EXACTFAA EXACT, str ; Match this string using /iaa rules (w/len) (string folded iff in UTF-8; non-UTF8 folded length <= unfolded). 105 106# End of important relative ordering. 107 108EXACTFUP EXACT, str ; Match this string using /iu rules (w/len); (string not UTF-8, not guaranteed to be folded; and its Problematic). 109# In order for a non-UTF-8 EXACTFAA to think the pattern is pre-folded when 110# matching a UTF-8 target string, there would have to be something like an 111# EXACTFAA_MICRO which would not be considered pre-folded for UTF-8 targets, 112# since the fold of the MICRO SIGN would not be done, and would be 113# representable in the UTF-8 target string. 114 115EXACTFLU8 EXACT, str ; Like EXACTFU, but use /il, UTF-8, folded, and everything in it is above 255. 116EXACTFAA_NO_TRIE EXACT, str ; Match this string using /iaa rules (w/len) (string not UTF-8, not guaranteed to be folded, not currently trie-able). 117 118 119EXACT_ONLY8 EXACT, str ; Like EXACT, but only UTF-8 encoded targets can match 120EXACTFU_ONLY8 EXACT, str ; Like EXACTFU, but only UTF-8 encoded targets can match 121# One could add EXACTFAA8 and and something that has the same effect for /l, 122# but these would be extremely uncommon 123 124EXACTFU_S_EDGE EXACT, str ; /di rules, but nothing in it precludes /ui, except begins and/or ends with [Ss]; (string not UTF-8; compile-time only). 125 126#*Do nothing types 127 128NOTHING NOTHING, no ; Match empty string. 129#*A variant of above which delimits a group, thus stops optimizations 130TAIL NOTHING, no ; Match empty string. Can jump here from outside. 131 132#*Loops 133 134#* STAR,PLUS '?', and complex '*' and '+', are implemented as 135#* circular BRANCH structures. Simple cases 136#* (one character per match) are implemented with STAR 137#* and PLUS for speed and to minimize recursive plunges. 138#* 139STAR STAR, node 0 V ; Match this (simple) thing 0 or more times. 140PLUS PLUS, node 0 V ; Match this (simple) thing 1 or more times. 141 142CURLY CURLY, sv 2 V ; Match this simple thing {n,m} times. 143CURLYN CURLY, no 2 V ; Capture next-after-this simple thing 144CURLYM CURLY, no 2 V ; Capture this medium-complex thing {n,m} times. 145CURLYX CURLY, sv 2 V ; Match this complex thing {n,m} times. 146 147#*This terminator creates a loop structure for CURLYX 148WHILEM WHILEM, no 0 V ; Do curly processing and see if rest matches. 149 150#*Buffer related 151 152#*OPEN,CLOSE,GROUPP ...are numbered at compile time. 153OPEN OPEN, num 1 ; Mark this point in input as start of #n. 154CLOSE CLOSE, num 1 ; Close corresponding OPEN of #n. 155SROPEN SROPEN, none ; Same as OPEN, but for script run 156SRCLOSE SRCLOSE, none ; Close preceding SROPEN 157 158REF REF, num 1 V ; Match some already matched string 159REFF REF, num 1 V ; Match already matched string, using /di rules. 160REFFL REF, num 1 V ; Match already matched string, using /li rules. 161# N?REFF[AU] could have been implemented using the FLAGS field of the 162# regnode, but by having a separate node type, we can use the existing switch 163# statement to avoid some tests 164REFFU REF, num 1 V ; Match already matched string, usng /ui. 165REFFA REF, num 1 V ; Match already matched string, using /aai rules. 166 167#*Named references. Code in regcomp.c assumes that these all are after 168#*the numbered references 169NREF REF, no-sv 1 V ; Match some already matched string 170NREFF REF, no-sv 1 V ; Match already matched string, using /di rules. 171NREFFL REF, no-sv 1 V ; Match already matched string, using /li rules. 172NREFFU REF, num 1 V ; Match already matched string, using /ui rules. 173NREFFA REF, num 1 V ; Match already matched string, using /aai rules. 174 175#*Support for long RE 176LONGJMP LONGJMP, off 1 . 1 ; Jump far away. 177BRANCHJ BRANCHJ, off 1 V 1 ; BRANCH with long offset. 178 179#*Special Case Regops 180IFMATCH BRANCHJ, off 1 . 1 ; Succeeds if the following matches; non-zero flags "f", next_off "o" means lookbehind assertion starting "f..(f-o)" characters before current 181UNLESSM BRANCHJ, off 1 . 1 ; Fails if the following matches; non-zero flags "f", next_off "o" means lookbehind assertion starting "f..(f-o)" characters before current 182SUSPEND BRANCHJ, off 1 V 1 ; "Independent" sub-RE. 183IFTHEN BRANCHJ, off 1 V 1 ; Switch, should be preceded by switcher. 184GROUPP GROUPP, num 1 ; Whether the group matched. 185 186 187#*The heavy worker 188 189EVAL EVAL, evl/flags 2L ; Execute some Perl code. 190 191#*Modifiers 192 193MINMOD MINMOD, no ; Next operator is not greedy. 194LOGICAL LOGICAL, no ; Next opcode should set the flag only. 195 196#*This is not used yet 197RENUM BRANCHJ, off 1 . 1 ; Group with independently numbered parens. 198 199#*Trie Related 200 201#* Behave the same as A|LIST|OF|WORDS would. The '..C' variants 202#* have inline charclass data (ascii only), the 'C' store it in the 203#* structure. 204# NOTE: the relative order of the TRIE-like regops is significant 205 206TRIE TRIE, trie 1 ; Match many EXACT(F[ALU]?)? at once. flags==type 207TRIEC TRIE,trie charclass ; Same as TRIE, but with embedded charclass data 208 209# For start classes, contains an added fail table. 210AHOCORASICK TRIE, trie 1 ; Aho Corasick stclass. flags==type 211AHOCORASICKC TRIE,trie charclass ; Same as AHOCORASICK, but with embedded charclass data 212 213#*Regex Subroutines 214GOSUB GOSUB, num/ofs 2L ; recurse to paren arg1 at (signed) ofs arg2 215 216#*Special conditionals 217NGROUPP NGROUPP, no-sv 1 ; Whether the group matched. 218INSUBP INSUBP, num 1 ; Whether we are in a specific recurse. 219DEFINEP DEFINEP, none 1 ; Never execute directly. 220 221#*Backtracking Verbs 222ENDLIKE ENDLIKE, none ; Used only for the type field of verbs 223OPFAIL ENDLIKE, no-sv 1 ; Same as (?!), but with verb arg 224ACCEPT ENDLIKE, no-sv/num 2L ; Accepts the current matched string, with verbar 225 226#*Verbs With Arguments 227VERB VERB, no-sv 1 ; Used only for the type field of verbs 228PRUNE VERB, no-sv 1 ; Pattern fails at this startpoint if no-backtracking through this 229MARKPOINT VERB, no-sv 1 ; Push the current location for rollback by cut. 230SKIP VERB, no-sv 1 ; On failure skip forward (to the mark) before retrying 231COMMIT VERB, no-sv 1 ; Pattern fails outright if backtracking through this 232CUTGROUP VERB, no-sv 1 ; On failure go to the next alternation in the group 233 234#*Control what to keep in $&. 235KEEPS KEEPS, no ; $& begins here. 236 237#*New charclass like patterns 238LNBREAK LNBREAK, none ; generic newline pattern 239 240# NEW STUFF SOMEWHERE ABOVE THIS LINE 241 242################################################################################ 243 244#*SPECIAL REGOPS 245 246#* This is not really a node, but an optimized away piece of a "long" 247#* node. To simplify debugging output, we mark it as if it were a node 248OPTIMIZED NOTHING, off ; Placeholder for dump. 249 250#* Special opcode with the property that no opcode in a compiled program 251#* will ever be of this type. Thus it can be used as a flag value that 252#* no other opcode has been seen. END is used similarly, in that an END 253#* node cant be optimized. So END implies "unoptimizable" and PSEUDO 254#* mean "not seen anything to optimize yet". 255PSEUDO PSEUDO, off ; Pseudo opcode for internal use. 256 257------------------------------------------------------------------------------- 258# Format for second section: 259# REGOP \t typelist [ \t typelist] 260# typelist= namelist 261# = namelist:FAIL 262# = name:count 263 264# Anything below is a state 265# 266# 267TRIE next:FAIL 268EVAL B,postponed_AB:FAIL 269CURLYX end:FAIL 270WHILEM A_pre,A_min,A_max,B_min,B_max:FAIL 271BRANCH next:FAIL 272CURLYM A,B:FAIL 273IFMATCH A:FAIL 274CURLY B_min,B_max:FAIL 275COMMIT next:FAIL 276MARKPOINT next:FAIL 277SKIP next:FAIL 278CUTGROUP next:FAIL 279KEEPS next:FAIL 280