1REGEXP(3) 386BSD Programmer's Manual REGEXP(3) 2 3NNAAMMEE 4 rreeggccoommpp, rreeggeexxeecc, rreeggssuubb, rreeggeerrrroorr - regular expression handlers 5 6SSYYNNOOPPSSIISS 7 ##iinncclluuddee <<rreeggeexxpp..hh>> 8 9 _r_e_g_e_x_p * 10 rreeggccoommpp(_c_o_n_s_t _c_h_a_r *_e_x_p) 11 12 _i_n_t 13 rreeggeexxeecc(_c_o_n_s_t _r_e_g_e_x_p *_p_r_o_g, _c_o_n_s_t _c_h_a_r *_s_t_r_i_n_g) 14 15 _v_o_i_d 16 rreeggssuubb(_c_o_n_s_t _r_e_g_e_x_p *_p_r_o_g, _c_o_n_s_t _c_h_a_r *_s_o_u_r_c_e, _c_h_a_r *_d_e_s_t) 17 18DDEESSCCRRIIPPTTIIOONN 19 The rreeggccoommpp(), rreeggeexxeecc(), rreeggssuubb(), and rreeggeerrrroorr() functions implement 20 egrep(1)-style regular expressions and supporting facilities. 21 22 The rreeggccoommpp() function compiles a regular expression into a structure of 23 type regexp, and returns a pointer to it. The space has been allocated 24 using malloc(3) and may be released by free. 25 26 The rreeggeexxeecc() function matches a NUL-terminated _s_t_r_i_n_g against the 27 compiled regular expression in _p_r_o_g. It returns 1 for success and 0 for 28 failure, and adjusts the contents of _p_r_o_g's _s_t_a_r_t_p and _e_n_d_p (see below) 29 accordingly. 30 31 The members of a regexp structure include at least the following (not 32 necessarily in order): 33 34 char *startp[NSUBEXP]; 35 char *endp[NSUBEXP]; 36 37 where NSUBEXP is defined (as 10) in the header file. Once a successful 38 rreeggeexxeecc() has been done using the rreeggeexxpp(), each _s_t_a_r_t_p- _e_n_d_p pair 39 describes one substring within the _s_t_r_i_n_g, with the _s_t_a_r_t_p pointing to 40 the first character of the substring and the _e_n_d_p pointing to the first 41 character following the substring. The 0th substring is the substring of 42 _s_t_r_i_n_g that matched the whole regular expression. The others are those 43 substrings that matched parenthesized expressions within the regular 44 expression, with parenthesized expressions numbered in left-to-right 45 order of their opening parentheses. 46 47 The rreeggssuubb() function copies _s_o_u_r_c_e to _d_e_s_t, making substitutions 48 according to the most recent rreeggeexxeecc() performed using _p_r_o_g. Each 49 instance of `&' in _s_o_u_r_c_e is replaced by the substring indicated by 50 _s_t_a_r_t_p[] and _e_n_d_p[]. Each instance of `\_n', where _n is a digit, is 51 replaced by the substring indicated by _s_t_a_r_t_p[_n] and _e_n_d_p[_n]. To get a 52 literal `&' or `\_n' into _d_e_s_t, prefix it with `\'; to get a literal `\' 53 preceding `&' or `\_n', prefix it with another `\'. 54 55 The rreeggeerrrroorr() function is called whenever an error is detected in 56 rreeggccoommpp(), rreeggeexxeecc(), or rreeggssuubb(). The default rreeggeerrrroorr() writes the 57 string _m_s_g, with a suitable indicator of origin, on the standard error 58 output and invokes exit(2). The rreeggeerrrroorr() function can be replaced by 59 the user if other actions are desirable. 60 61RREEGGUULLAARR EEXXPPRREESSSSIIOONN SSYYNNTTAAXX 62 A regular expression is zero or more _b_r_a_n_c_h_e_s, separated by `|'. It 63 matches anything that matches one of the branches. 64 65 66 A branch is zero or more _p_i_e_c_e_s, concatenated. It matches a match for 67 the first, followed by a match for the second, etc. 68 69 A piece is an _a_t_o_m possibly followed by `*', `+', or `?'. An atom 70 followed by `*' matches a sequence of 0 or more matches of the atom. An 71 atom followed by `+' matches a sequence of 1 or more matches of the atom. 72 An atom followed by `?' matches a match of the atom, or the null string. 73 74 An atom is a regular expression in parentheses (matching a match for the 75 regular expression), a _r_a_n_g_e (see below), `.' (matching any single 76 character), `^' (matching the null string at the beginning of the input 77 string), `$' (matching the null string at the end of the input string), a 78 `\' followed by a single character (matching that character), or a single 79 character with no other significance (matching that character). 80 81 A _r_a_n_g_e is a sequence of characters enclosed in `[]'. It normally 82 matches any single character from the sequence. If the sequence begins 83 with `^', it matches any single character _n_o_t from the rest of the 84 sequence. If two characters in the sequence are separated by `-', this 85 is shorthand for the full list of ASCII characters between them (e.g. 86 `[0-9]' matches any decimal digit). To include a literal `]' in the 87 sequence, make it the first character (following a possible `^'). To 88 include a literal `-', make it the first or last character. 89 90AAMMBBIIGGUUIITTYY 91 If a regular expression could match two different parts of the input 92 string, it will match the one which begins earliest. If both begin in 93 the same place but match different lengths, or match the same length in 94 different ways, life gets messier, as follows. 95 96 In general, the possibilities in a list of branches are considered in 97 left-to-right order, the possibilities for `*', `+', and `?' are 98 considered longest-first, nested constructs are considered from the 99 outermost in, and concatenated constructs are considered leftmost-first. 100 The match that will be chosen is the one that uses the earliest 101 possibility in the first choice that has to be made. If there is more 102 than one choice, the next will be made in the same manner (earliest 103 possibility) subject to the decision on the first choice. And so forth. 104 105 For example, `(ab|a)b*c' could match `abc' in one of two ways. The first 106 choice is between `ab' and `a'; since `ab' is earlier, and does lead to a 107 successful overall match, it is chosen. Since the `b' is already spoken 108 for, the `b*' must match its last possibility-the empty string-since it 109 must respect the earlier choice. 110 111 In the particular case where no `|'s are present and there is only one 112 `*', `+', or `?', the net effect is that the longest possible match will 113 be chosen. So `ab*', presented with `xabbbby', will match `abbbb'. Note 114 that if `ab*', is tried against `xabyabbbz', it will match `ab' just 115 after `x', due to the begins-earliest rule. (In effect, the decision on 116 where to start the match is the first choice to be made, hence subsequent 117 choices must respect it even if this leads them to less-preferred 118 alternatives.) 119 120RREETTUURRNN VVAALLUUEESS 121 The rreeggccoommpp() function returns NULL for a failure (rreeggeerrrroorr() 122 permitting), where failures are syntax errors, exceeding implementation 123 limits, or applying `+' or `*' to a possibly-null operand. 124 125SSEEEE AALLSSOO 126 ed(1), ex(1), expr(1), egrep(1), fgrep(1), grep(1), regex(3) 127 128HHIISSTTOORRYY 129 Both code and manual page for rreeggccoommpp(), rreeggeexxeecc(), rreeggssuubb(), and 130 rreeggeerrrroorr() were written at the University of Toronto and appeared in 131 4.3BSD-Tahoe. They are intended to be compatible with the Bell V8 132 regexp(3), but are not derived from Bell code. 133 134BBUUGGSS 135 Empty branches and empty regular expressions are not portable to V8. 136 137 The restriction against applying `*' or `+' to a possibly-null operand is 138 an artifact of the simplistic implementation. 139 140 Does not support egrep's newline-separated branches; neither does the V8 141 regexp(3), though. 142 143 Due to emphasis on compactness and simplicity, it's not strikingly fast. 144 It does give special attention to handling simple cases quickly. 145 146BSD Experimental April 19, 1991 3 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199