From c3c43fd39c715827de5cc74846ba977c5b1d2181 Mon Sep 17 00:00:00 2001 From: Kaz Kylheku Date: Fri, 20 Apr 2012 00:47:46 -0700 Subject: * parser.y (regtoken): New nonterminal symbol. (regterm): REGTOKEN production factored out to regtoken. (regclass): Reverted prior commmit's changes. (regclassterm): Reverted prior commit, removing REGTOKEN production for character classes, and introduced a regtoken production. So now the keyword symbols are part of the character class abstract syntax. (regtoken): New production rule. * regex.c (regex_space_chars): Converted to internal linkage. (char_set_compile): Handle token keywords in character class abstract syntax. * regex.h (regex_space_chars): External declaration removed. --- ChangeLog | 17 +++++++++++++++++ parser.y | 56 +++++++++++++++++++++----------------------------------- regex.c | 31 ++++++++++++++++++++++++++++++- regex.h | 1 - 4 files changed, 68 insertions(+), 37 deletions(-) diff --git a/ChangeLog b/ChangeLog index 75b111d3..4497be3b 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,3 +1,20 @@ +2012-04-20 Kaz Kylheku + + * parser.y (regtoken): New nonterminal symbol. + (regterm): REGTOKEN production factored out to regtoken. + (regclass): Reverted prior commmit's changes. + (regclassterm): Reverted prior commit, removing REGTOKEN + production for character classes, and introduced a regtoken + production. So now the keyword symbols are part of the + character class abstract syntax. + (regtoken): New production rule. + + * regex.c (regex_space_chars): Converted to internal linkage. + (char_set_compile): Handle token keywords in character class + abstract syntax. + + * regex.h (regex_space_chars): External declaration removed. + 2012-04-19 Kaz Kylheku First cut at implementing \s, \d, \w, \S, \D and \W regex tokens. diff --git a/parser.y b/parser.y index c57bc901..46a0052b 100644 --- a/parser.y +++ b/parser.y @@ -90,7 +90,7 @@ static val parsed_spec; %type repeat_clause repeat_parts_opt o_line %type o_elems_opt o_elems o_elem o_var rep_elem rep_parts_opt %type regex lisp_regex regexpr regbranch -%type regterm regclass regclassterm regrange +%type regterm regtoken regclass regclassterm regrange %type strlit chrlit quasilit quasi_items quasi_item litchars %type regchar %type '(' '[' @@ -796,48 +796,20 @@ regterm : regterm '*' { $$ = list(zeroplus_s, $1, nao); } | ']' { $$ = chr(']'); } | '-' { $$ = chr('-'); } | REGCHAR { $$ = chr($1); } - | REGTOKEN { switch ($1) - { case 's': - $$ = space_k; break; - case 'S': - $$ = cspace_k; break; - case 'd': - $$ = digit_k; break; - case 'D': - $$ = cdigit_k; break; - case 'w': - $$ = word_char_k; break; - case 'W': - $$ = cword_char_k; break; }} + | regtoken { $$ = $1; } | '(' regexpr ')' { $$ = $2; } | '(' error { $$ = nil; yybadtoken(yychar, lit("regex subexpression")); } ; -regclass : regclassterm { $$ = $1; } - | regclassterm regclass { $$ = nappend2($1, $2); } +regclass : regclassterm { $$ = cons($1, nil); } + | regclassterm regclass { $$ = cons($1, $2); } ; -regclassterm : regrange { $$ = cons($1, nil); } - | regchar { $$ = cons(chr($1), nil); } - | REGTOKEN { switch ($1) - { case 's': - $$ = regex_space_chars; - break; - case 'd': - $$ = cons(cons(chr('0'), chr('9')), nil); - break; - case 'w': - $$ = list(cons(chr('A'), chr('Z')), - cons(chr('a'), chr('z')), - chr('_'), nao); - break; - default: - yyerrorf(lit("complemented token " - "\\~a not allowed " - "in regex character class"), - chr($1), nao); } } +regclassterm : regrange { $$ = $1; } + | regchar { $$ = chr($1); } + | regtoken { $$ = $1; } ; regrange : regchar '-' regchar { $$ = cons(chr($1), chr($3)); } @@ -856,6 +828,20 @@ regchar : '?' { $$ = '?'; } | REGCHAR { $$ = $1; } ; +regtoken : REGTOKEN { switch ($1) + { case 's': + $$ = space_k; break; + case 'S': + $$ = cspace_k; break; + case 'd': + $$ = digit_k; break; + case 'D': + $$ = cdigit_k; break; + case 'w': + $$ = word_char_k; break; + case 'W': + $$ = cword_char_k; break; }} + newl : '\n' | error '\n' { yyerror("newline expected after directive"); yyerrok; } diff --git a/regex.c b/regex.c index 9670b901..77989a98 100644 --- a/regex.c +++ b/regex.c @@ -209,6 +209,8 @@ union regex_machine { int opt_derivative_regex = 0; +static val regex_space_chars; + static int L0_full(cset_L0_t *L0) { int i; @@ -634,6 +636,24 @@ static char_set_t *char_set_compile(val args, val comp) min = c_chr(item); if (c_chr(item) > max) max = c_chr(item); + } else if (item == space_k) { + if (max < 0x3000) + max = 0x3000; + if (min > 0x9) + min = 0x9; + } else if (item == digit_k) { + if (max < '9') + max = 9; + if (min > '0') + min = 0; + } else if (item == word_char_k) { + if (min > 'A') + min = 'A'; + if (max < 'z') + max = 'z'; + } else if (item == cspace_k || item == cdigit_k || item == cword_char_k) { + uw_throwf(error_s, lit("bad object in character class syntax: ~s"), + item, nao); } else { assert(0 && "bad regex set"); } @@ -667,6 +687,16 @@ static char_set_t *char_set_compile(val args, val comp) char_set_add_range(set, c_chr(from), c_chr(to)); } else if (typeof(item) == chr_s) { char_set_add(set, c_chr(item)); + } else if (item == space_k) { + val iter; + for (iter = regex_space_chars; iter; iter = cdr(iter)) + char_set_add(set, c_chr(car(iter))); + } else if (item == digit_k) { + char_set_add_range(set, '0', '9'); + } else if (item == word_char_k) { + char_set_add_range(set, 'A', 'Z'); + char_set_add_range(set, 'a', 'z'); + char_set_add(set, '_'); } else { assert(0 && "bad regex set"); } @@ -1852,7 +1882,6 @@ val regsub(val regex, val repl, val str) val space_k, digit_k, word_char_k; val cspace_k, cdigit_k, cword_char_k; -val regex_space_chars; void regex_init(void) { diff --git a/regex.h b/regex.h index 4d137a0f..962a2846 100644 --- a/regex.h +++ b/regex.h @@ -26,7 +26,6 @@ extern val space_k, digit_k, word_char_k; extern val cspace_k, cdigit_k, cword_char_k; -extern val regex_space_chars; val regex_compile(val regex_sexp); val regexp(val); -- cgit v1.2.3