📄 regex.cc
字号:
/* Extended regular expression matching and search library. Copyright (C) 1985, 1989-90 Free Software Foundation, Inc. This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */// This is a translation into C++ of regex.c, the GNU regexp package./* To test, compile with -Dtest. This Dtestable feature turns this into a self-contained program which reads a pattern, describes how it compiles, then reads a string and searches for it. On the other hand, if you compile with both -Dtest and -Dcanned you can run some tests we've already thought of. *//* AIX requires the alloca decl to be the first thing in the file. */#ifdef __GNUC__#define alloca alloca#else#ifdef sparc#include <alloca.h>#else#ifdef _AIX#pragma alloca#elsechar *alloca ();#endif#endif#endif#ifdef emacs/* The `emacs' switch turns on certain special matching commands that make sense only in emacs. */#include "config.h"#include "lisp.h"#include "buffer.h"#include "syntax.h"#else /* not emacs */#include <_G_config.h>#include <string.h>#include <stdlib.h>/* Define the syntax stuff, so we can do the \<, \>, etc. *//* This must be nonzero for the wordchar and notwordchar pattern commands in re_match_2. */#ifndef Sword #define Sword 1#endif#define SYNTAX(c) re_syntax_table[c]#ifdef SYNTAX_TABLEchar *re_syntax_table;#else /* not SYNTAX_TABLE */static char re_syntax_table[256];static voidinit_syntax_once (){ register int c; static int done = 0; if (done) return; memset (re_syntax_table, 0, sizeof re_syntax_table); for (c = 'a'; c <= 'z'; c++) re_syntax_table[c] = Sword; for (c = 'A'; c <= 'Z'; c++) re_syntax_table[c] = Sword; for (c = '0'; c <= '9'; c++) re_syntax_table[c] = Sword; done = 1;}#endif /* SYNTAX_TABLE */#endif /* emacs *//* We write fatal error messages on standard error. */#include <stdio.h>/* isalpha(3) etc. are used for the character classes. */#include <ctype.h>/* Sequents are missing isgraph. */#ifndef isgraph#define isgraph(c) (isprint((c)) && !isspace((c)))#endif/* Get the interface, including the syntax bits. */#include "regex.h"/* These are the command codes that appear in compiled regular expressions, one per byte. Some command codes are followed by argument bytes. A command code can specify any interpretation whatsoever for its arguments. Zero-bytes may appear in the compiled regular expression. The value of `exactn' is needed in search.c (search_buffer) in emacs. So regex.h defines a symbol `RE_EXACTN_VALUE' to be 1; the value of `exactn' we use here must also be 1. */enum regexpcode { unused=0, exactn=1, /* Followed by one byte giving n, then by n literal bytes. */ begline, /* Fail unless at beginning of line. */ endline, /* Fail unless at end of line. */ jump, /* Followed by two bytes giving relative address to jump to. */ on_failure_jump, /* Followed by two bytes giving relative address of place to resume at in case of failure. */ finalize_jump, /* Throw away latest failure point and then jump to address. */ maybe_finalize_jump, /* Like jump but finalize if safe to do so. This is used to jump back to the beginning of a repeat. If the command that follows this jump is clearly incompatible with the one at the beginning of the repeat, such that we can be sure that there is no use backtracking out of repetitions already completed, then we finalize. */ dummy_failure_jump, /* Jump, and push a dummy failure point. This failure point will be thrown away if an attempt is made to use it for a failure. A + construct makes this before the first repeat. Also use it as an intermediary kind of jump when compiling an or construct. */ succeed_n, /* Used like on_failure_jump except has to succeed n times; then gets turned into an on_failure_jump. The relative address following it is useless until then. The address is followed by two bytes containing n. */ jump_n, /* Similar to jump, but jump n times only; also the relative address following is in turn followed by yet two more bytes containing n. */ set_number_at, /* Set the following relative location to the subsequent number. */ anychar, /* Matches any (more or less) one character. */ charset, /* Matches any one char belonging to specified set. First following byte is number of bitmap bytes. Then come bytes for a bitmap saying which chars are in. Bits in each byte are ordered low-bit-first. A character is in the set if its bit is 1. A character too large to have a bit in the map is automatically not in the set. */ charset_not, /* Same parameters as charset, but match any character that is not one of those specified. */ start_memory, /* Start remembering the text that is matched, for storing in a memory register. Followed by one byte containing the register number. Register numbers must be in the range 0 through RE_NREGS. */ stop_memory, /* Stop remembering the text that is matched and store it in a memory register. Followed by one byte containing the register number. Register numbers must be in the range 0 through RE_NREGS. */ duplicate, /* Match a duplicate of something remembered. Followed by one byte containing the index of the memory register. */#ifdef emacs before_dot, /* Succeeds if before point. */ at_dot, /* Succeeds if at point. */ after_dot, /* Succeeds if after point. */#endif begbuf, /* Succeeds if at beginning of buffer. */ endbuf, /* Succeeds if at end of buffer. */ wordchar, /* Matches any word-constituent character. */ notwordchar, /* Matches any char that is not a word-constituent. */ wordbeg, /* Succeeds if at word beginning. */ wordend, /* Succeeds if at word end. */ wordbound, /* Succeeds if at a word boundary. */ notwordbound,/* Succeeds if not at a word boundary. */#ifdef emacs syntaxspec, /* Matches any character whose syntax is specified. followed by a byte which contains a syntax code, e.g., Sword. */ notsyntaxspec /* Matches any character whose syntax differs from that specified. */#endif }; /* Number of failure points to allocate space for initially, when matching. If this number is exceeded, more space is allocated, so it is not a hard limit. */#ifndef NFAILURES#define NFAILURES 80#endif#ifndef SIGN_EXTEND_CHAR#ifdef __STDC__#define SIGN_EXTEND_CHAR(c) ((signed char)(c))#else#define SIGN_EXTEND_CHAR(c) (((c)^128) - 128) /* As in Harbison and Steele. */#endif#endif /* not SIGN_EXTEND_CHAR *//* Store NUMBER in two contiguous bytes starting at DESTINATION. */#define STORE_NUMBER(destination, number) \ { (destination)[0] = (number) & 0377; \ (destination)[1] = (number) >> 8; } /* Same as STORE_NUMBER, except increment the destination pointer to the byte after where the number is stored. Watch out that values for DESTINATION such as p + 1 won't work, whereas p will. */#define STORE_NUMBER_AND_INCR(destination, number) \ { STORE_NUMBER(destination, number); \ (destination) += 2; }/* Put into DESTINATION a number stored in two contingous bytes starting at SOURCE. */#define EXTRACT_NUMBER(destination, source) \ { (destination) = *(source) & 0377; \ (destination) += SIGN_EXTEND_CHAR (*(char *)((source) + 1)) << 8; }/* Same as EXTRACT_NUMBER, except increment the pointer for source to point to second byte of SOURCE. Note that SOURCE has to be a value such as p, not, e.g., p + 1. */#define EXTRACT_NUMBER_AND_INCR(destination, source) \ { EXTRACT_NUMBER (destination, source); \ (source) += 2; }/* Specify the precise syntax of regexps for compilation. This provides for compatibility for various utilities which historically have different, incompatible syntaxes. The argument SYNTAX is a bit-mask comprised of the various bits defined in regex.h. */intre_set_syntax (int syntax){ int ret; ret = obscure_syntax; obscure_syntax = syntax; return ret;}/* Set by re_set_syntax to the current regexp syntax to recognize. */int obscure_syntax = 0;/* Macros for re_compile_pattern, which is found below these definitions. */#define CHAR_CLASS_MAX_LENGTH 6/* Fetch the next character in the uncompiled pattern, translating it if necessary. */#define PATFETCH(c) \ {if (p == pend) goto end_of_pattern; \ c = * (const unsigned char *) p++; \ if (translate) c = translate[c]; }/* Fetch the next character in the uncompiled pattern, with no translation. */#define PATFETCH_RAW(c) \ {if (p == pend) goto end_of_pattern; \ c = * (const unsigned char *) p++; }#define PATUNFETCH p--/* If the buffer isn't allocated when it comes in, use this. */#define INIT_BUF_SIZE 28/* Make sure we have at least N more bytes of space in buffer. */#define GET_BUFFER_SPACE(n) \ { \ while (b - bufp->buffer + (n) >= bufp->allocated) \ EXTEND_BUFFER; \ }/* Make sure we have one more byte of buffer space and then add CH to it. */#define BUFPUSH(ch) \ { \ GET_BUFFER_SPACE (1); \ *b++ = (char) (ch); \ } /* Extend the buffer by twice its current size via reallociation and reset the pointers that pointed into the old allocation to point to the correct places in the new allocation. If extending the buffer results in it being larger than 1 << 16, then flag memory exhausted. */#define EXTEND_BUFFER \ { char *old_buffer = bufp->buffer; \ if (bufp->allocated == (1L<<16)) goto too_big; \ bufp->allocated *= 2; \ if (bufp->allocated > (1L<<16)) bufp->allocated = (1L<<16); \ bufp->buffer = (char *) realloc (bufp->buffer, bufp->allocated); \ if (bufp->buffer == 0) \ goto memory_exhausted; \ b = (b - old_buffer) + bufp->buffer; \ if (fixup_jump) \ fixup_jump = (fixup_jump - old_buffer) + bufp->buffer; \ if (laststart) \ laststart = (laststart - old_buffer) + bufp->buffer; \ begalt = (begalt - old_buffer) + bufp->buffer; \ if (pending_exact) \ pending_exact = (pending_exact - old_buffer) + bufp->buffer; \ }/* Set the bit for character C in a character set list. */#define SET_LIST_BIT(c) (b[(c) / BYTEWIDTH] |= 1 << ((c) % BYTEWIDTH))/* Get the next unsigned number in the uncompiled pattern. */#define GET_UNSIGNED_NUMBER(num) \ { if (p != pend) \ { \ PATFETCH (c); \ while (isdigit (c)) \ { \ if (num < 0) \ num = 0; \ num = num * 10 + c - '0'; \ if (p == pend) \ break; \ PATFETCH (c); \ } \ } \ }/* Subroutines for re_compile_pattern. */static void store_jump (char *from, char opcode, char *to);static void insert_jump (char op, char *from, char *to, char *current_end);static void store_jump_n (char *from, char opcode, char *to, unsigned n);static void insert_jump_n (char, char *, char *, char *, unsigned);static void insert_op_2 (char, char *, char *_end, int, int);/* re_compile_pattern takes a regular-expression string and converts it into a buffer full of byte commands for matching. PATTERN is the address of the pattern string SIZE is the length of it. BUFP is a struct re_pattern_buffer * which points to the info on where to store the byte commands. This structure contains a char * which points to the actual space, which should have been obtained with malloc. re_compile_pattern may use realloc to grow the buffer space. The number of bytes of commands can be found out by looking in the `struct re_pattern_buffer' that bufp pointed to, after re_compile_pattern returns. */char *re_compile_pattern (const char *pattern, int size, struct re_pattern_buffer *bufp){ register char *b = bufp->buffer; register const char *p = pattern; const char *pend = pattern + size; register unsigned c, c1; const char *p1; unsigned char *translate = (unsigned char *) bufp->translate; /* Address of the count-byte of the most recently inserted `exactn' command. This makes it possible to tell whether a new exact-match character can be added to that command or requires a new `exactn' command. */ char *pending_exact = 0; /* Address of the place where a forward-jump should go to the end of the containing expression. Each alternative of an `or', except the last, ends with a forward-jump of this sort. */ char *fixup_jump = 0; /* Address of start of the most recently finished expression. This tells postfix * where to find the start of its operand. */ char *laststart = 0; /* In processing a repeat, 1 means zero matches is allowed. */ char zero_times_ok; /* In processing a repeat, 1 means many matches is allowed. */ char many_times_ok; /* Address of beginning of regexp, or inside of last \(. */ char *begalt = b; /* In processing an interval, at least this many matches must be made. */ int lower_bound; /* In processing an interval, at most this many matches can be made. */ int upper_bound;
⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -