From ee78102672afdede839489fa0b1932b64335eaaf Mon Sep 17 00:00:00 2001 From: Bobby Bingham Date: Tue, 25 Jul 2017 20:49:17 -0500 Subject: move regex printing code to print-regex.c --- Makefile | 2 +- print-regex.c | 124 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ regex.c | 124 ---------------------------------------------------------- 3 files changed, 125 insertions(+), 125 deletions(-) create mode 100644 print-regex.c delete mode 100644 regex.c diff --git a/Makefile b/Makefile index 91c36ad..79099ba 100644 --- a/Makefile +++ b/Makefile @@ -25,7 +25,7 @@ parse-%.c parse-%.h: parse-%.y $(CC) $(CFLAGS) -c $< -o $@ pattern2regex.o: lex-pattern.h parse-pattern.h -pattern2regex: pattern2regex.o lex-pattern.o parse-pattern.o ast.o regex.o +pattern2regex: pattern2regex.o lex-pattern.o parse-pattern.o ast.o print-regex.o $(TARGETS): $(CC) $(CFLAGS) $^ -o $@ diff --git a/print-regex.c b/print-regex.c new file mode 100644 index 0000000..02a58fa --- /dev/null +++ b/print-regex.c @@ -0,0 +1,124 @@ +#include +#include + +#include "ast.h" + +static const int precedence[ATOM_MAX] = { + [ATOM_ALPHABETIC] = 1, + [ATOM_NUMERIC] = 1, + [ATOM_UPPERCASE] = 1, + [ATOM_LOWERCASE] = 1, + [ATOM_CONTROL] = 1, + [ATOM_PUNCTUATION] = 1, + [ATOM_EVERYTHING] = 1, + [ATOM_REPETITION] = 2, + [ATOM_SEQUENCE] = 3, + [ATOM_LITERAL] = 3, + [ATOM_ALTERNATION] = 4, +}; + +static void print_repeat(const struct repeat *r) +{ + if (r->max < 0) { + switch (r->min) { + case 0: printf("*"); break; + case 1: printf("+"); break; + default: printf("{%ld,}", r->min); break; + } + } else if (r->min == 0 && r->max == 1) { + printf("?"); + } else if (r->min == r->max) { + printf("{%ld}", r->min); + } else { + printf("{%ld,%ld}", r->min, r->max); + } +} + +static void print_literal(const char *s) +{ + for (; *s; s++) { + switch (*s) { + case '.': + case '[': + case '{': + case '}': + case '(': + case ')': + case '\\': + case '*': + case '+': + case '?': + case '|': + case '^': + case '$': + putchar('\\'); + + default: + putchar(*s); + } + } +} + +static int get_precedence(const struct atom *a) +{ + if (a->type == ATOM_LITERAL) { + switch (strlen(a->u.literal)) { + case 0: return -1; + case 1: return precedence[ATOM_EVERYTHING]; + default: return precedence[ATOM_SEQUENCE]; + } + } + + return precedence[a->type]; +} + +static void print_atom(const struct atom *a); + +static void print_child_atom(const struct atom *child, const struct atom *parent) +{ + int group = get_precedence(child) > get_precedence(parent); + if (group) printf("("); + print_atom(child); + if (group) printf(")"); +} + +static void print_atom(const struct atom *a) +{ + switch (a->type) { + case ATOM_ALTERNATION: + print_child_atom(a->u.children[0], a); + printf("|"); + print_child_atom(a->u.children[1], a); + break; + + case ATOM_SEQUENCE: + print_child_atom(a->u.children[0], a); + print_child_atom(a->u.children[1], a); + break; + + case ATOM_REPETITION: + print_child_atom(a->u.repeat.child, a); + print_repeat(&a->u.repeat.counts); + break; + + case ATOM_LITERAL: + print_literal(a->u.literal); + break; + + case ATOM_ALPHABETIC: printf("[:alpha:]"); break; + case ATOM_NUMERIC: printf("[:digit:]"); break; + case ATOM_UPPERCASE: printf("[:upper:]"); break; + case ATOM_LOWERCASE: printf("[:lower:]"); break; + case ATOM_CONTROL: printf("[:cntrl:]"); break; + case ATOM_PUNCTUATION: printf("[:punct:]"); break; + case ATOM_EVERYTHING: printf("."); break; + } +} + +void print_regex(const struct atom *a) +{ + printf("^"); + print_atom(a); + printf("$\n"); +} + diff --git a/regex.c b/regex.c deleted file mode 100644 index 02a58fa..0000000 --- a/regex.c +++ /dev/null @@ -1,124 +0,0 @@ -#include -#include - -#include "ast.h" - -static const int precedence[ATOM_MAX] = { - [ATOM_ALPHABETIC] = 1, - [ATOM_NUMERIC] = 1, - [ATOM_UPPERCASE] = 1, - [ATOM_LOWERCASE] = 1, - [ATOM_CONTROL] = 1, - [ATOM_PUNCTUATION] = 1, - [ATOM_EVERYTHING] = 1, - [ATOM_REPETITION] = 2, - [ATOM_SEQUENCE] = 3, - [ATOM_LITERAL] = 3, - [ATOM_ALTERNATION] = 4, -}; - -static void print_repeat(const struct repeat *r) -{ - if (r->max < 0) { - switch (r->min) { - case 0: printf("*"); break; - case 1: printf("+"); break; - default: printf("{%ld,}", r->min); break; - } - } else if (r->min == 0 && r->max == 1) { - printf("?"); - } else if (r->min == r->max) { - printf("{%ld}", r->min); - } else { - printf("{%ld,%ld}", r->min, r->max); - } -} - -static void print_literal(const char *s) -{ - for (; *s; s++) { - switch (*s) { - case '.': - case '[': - case '{': - case '}': - case '(': - case ')': - case '\\': - case '*': - case '+': - case '?': - case '|': - case '^': - case '$': - putchar('\\'); - - default: - putchar(*s); - } - } -} - -static int get_precedence(const struct atom *a) -{ - if (a->type == ATOM_LITERAL) { - switch (strlen(a->u.literal)) { - case 0: return -1; - case 1: return precedence[ATOM_EVERYTHING]; - default: return precedence[ATOM_SEQUENCE]; - } - } - - return precedence[a->type]; -} - -static void print_atom(const struct atom *a); - -static void print_child_atom(const struct atom *child, const struct atom *parent) -{ - int group = get_precedence(child) > get_precedence(parent); - if (group) printf("("); - print_atom(child); - if (group) printf(")"); -} - -static void print_atom(const struct atom *a) -{ - switch (a->type) { - case ATOM_ALTERNATION: - print_child_atom(a->u.children[0], a); - printf("|"); - print_child_atom(a->u.children[1], a); - break; - - case ATOM_SEQUENCE: - print_child_atom(a->u.children[0], a); - print_child_atom(a->u.children[1], a); - break; - - case ATOM_REPETITION: - print_child_atom(a->u.repeat.child, a); - print_repeat(&a->u.repeat.counts); - break; - - case ATOM_LITERAL: - print_literal(a->u.literal); - break; - - case ATOM_ALPHABETIC: printf("[:alpha:]"); break; - case ATOM_NUMERIC: printf("[:digit:]"); break; - case ATOM_UPPERCASE: printf("[:upper:]"); break; - case ATOM_LOWERCASE: printf("[:lower:]"); break; - case ATOM_CONTROL: printf("[:cntrl:]"); break; - case ATOM_PUNCTUATION: printf("[:punct:]"); break; - case ATOM_EVERYTHING: printf("."); break; - } -} - -void print_regex(const struct atom *a) -{ - printf("^"); - print_atom(a); - printf("$\n"); -} - -- cgit v1.2.3