1 /* $NetBSD: file.c,v 1.11 2019/04/05 13:34:41 christos Exp $ */ 2 /* $FreeBSD: head/usr.bin/grep/file.c 211496 2010-08-19 09:28:59Z des $ */ 3 /* $OpenBSD: file.c,v 1.11 2010/07/02 20:48:48 nicm Exp $ */ 4 5 /*- 6 * Copyright (c) 1999 James Howard and Dag-Erling Coïdan Smørgrav 7 * Copyright (C) 2008-2010 Gabor Kovesdan <gabor@FreeBSD.org> 8 * Copyright (C) 2010 Dimitry Andric <dimitry@andric.com> 9 * All rights reserved. 10 * 11 * Redistribution and use in source and binary forms, with or without 12 * modification, are permitted provided that the following conditions 13 * are met: 14 * 1. Redistributions of source code must retain the above copyright 15 * notice, this list of conditions and the following disclaimer. 16 * 2. Redistributions in binary form must reproduce the above copyright 17 * notice, this list of conditions and the following disclaimer in the 18 * documentation and/or other materials provided with the distribution. 19 * 20 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 21 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 22 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 23 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 24 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 25 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 26 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 27 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 28 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 29 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 30 * SUCH DAMAGE. 31 */ 32 33 #if HAVE_NBTOOL_CONFIG_H 34 #include "nbtool_config.h" 35 #endif 36 37 #include <sys/cdefs.h> 38 __RCSID("$NetBSD: file.c,v 1.11 2019/04/05 13:34:41 christos Exp $"); 39 40 #include <sys/param.h> 41 #include <sys/types.h> 42 #include <sys/stat.h> 43 44 #include <err.h> 45 #include <errno.h> 46 #include <fcntl.h> 47 #include <stddef.h> 48 #include <stdlib.h> 49 #include <string.h> 50 #include <unistd.h> 51 #include <wchar.h> 52 #include <wctype.h> 53 54 #include "grep.h" 55 56 #define MAXBUFSIZ (32 * 1024) 57 #define LNBUFBUMP 80 58 59 #ifndef WITHOUT_GZIP 60 static gzFile gzbufdesc; 61 #endif 62 #ifndef WITHOUT_BZ2 63 static BZFILE* bzbufdesc; 64 #endif 65 66 static unsigned char buffer[MAXBUFSIZ + 1]; 67 static unsigned char *bufpos; 68 static size_t bufrem; 69 70 static unsigned char *lnbuf; 71 static size_t lnbuflen; 72 73 static inline int 74 grep_refill(struct file *f) 75 { 76 ssize_t nr = -1; 77 int bzerr; 78 79 bufpos = buffer; 80 bufrem = 0; 81 82 #ifndef WITHOUT_GZIP 83 if (filebehave == FILE_GZIP) { 84 nr = gzread(gzbufdesc, buffer, MAXBUFSIZ); 85 if (nr == -1) 86 return -1; 87 } 88 #endif 89 #ifndef WITHOUT_BZ2 90 if (filebehave == FILE_BZIP && bzbufdesc != NULL) { 91 nr = BZ2_bzRead(&bzerr, bzbufdesc, buffer, MAXBUFSIZ); 92 switch (bzerr) { 93 case BZ_OK: 94 case BZ_STREAM_END: 95 /* No problem, nr will be okay */ 96 break; 97 case BZ_DATA_ERROR_MAGIC: 98 /* 99 * As opposed to gzread(), which simply returns the 100 * plain file data, if it is not in the correct 101 * compressed format, BZ2_bzRead() instead aborts. 102 * 103 * So, just restart at the beginning of the file again, 104 * and use plain reads from now on. 105 */ 106 BZ2_bzReadClose(&bzerr, bzbufdesc); 107 bzbufdesc = NULL; 108 if (lseek(f->fd, 0, SEEK_SET) == -1) 109 return (-1); 110 nr = read(f->fd, buffer, MAXBUFSIZ); 111 break; 112 default: 113 /* Make sure we exit with an error */ 114 nr = -1; 115 } 116 if (nr == -1) 117 return -1; 118 } 119 #endif 120 if (nr == -1) { 121 nr = read(f->fd, buffer, MAXBUFSIZ); 122 } 123 124 if (nr < 0) 125 return (-1); 126 127 bufrem = nr; 128 return (0); 129 } 130 131 static inline void 132 grep_lnbufgrow(size_t newlen) 133 { 134 135 if (lnbuflen < newlen) { 136 lnbuf = grep_realloc(lnbuf, newlen); 137 lnbuflen = newlen; 138 } 139 } 140 141 static void 142 grep_copyline(size_t off, size_t len) 143 { 144 memcpy(lnbuf + off, bufpos, len); 145 lnbuf[off + len] = '\0'; 146 } 147 148 char * 149 grep_fgetln(struct file *f, size_t *lenp) 150 { 151 unsigned char *p; 152 size_t len; 153 size_t off; 154 ptrdiff_t diff; 155 156 /* Fill the buffer, if necessary */ 157 if (bufrem == 0 && grep_refill(f) != 0) 158 goto error; 159 160 if (bufrem == 0) { 161 /* Return zero length to indicate EOF */ 162 *lenp = 0; 163 return ((char *)bufpos); 164 } 165 166 /* Look for a newline in the remaining part of the buffer */ 167 if ((p = memchr(bufpos, line_sep, bufrem)) != NULL) { 168 ++p; /* advance over newline */ 169 len = p - bufpos; 170 grep_lnbufgrow(len + 1); 171 grep_copyline(0, len); 172 *lenp = len; 173 bufrem -= len; 174 bufpos = p; 175 return (char *)lnbuf; 176 } 177 178 /* We have to copy the current buffered data to the line buffer */ 179 for (len = bufrem, off = 0; ; len += bufrem) { 180 /* Make sure there is room for more data */ 181 grep_lnbufgrow(len + LNBUFBUMP); 182 grep_copyline(off, len - off); 183 off = len; 184 if (grep_refill(f) != 0) 185 goto error; 186 if (bufrem == 0) 187 /* EOF: return partial line */ 188 break; 189 if ((p = memchr(bufpos, line_sep, bufrem)) == NULL) 190 continue; 191 /* got it: finish up the line (like code above) */ 192 ++p; 193 diff = p - bufpos; 194 len += diff; 195 grep_lnbufgrow(len + 1); 196 grep_copyline(off, diff); 197 bufrem -= diff; 198 bufpos = p; 199 break; 200 } 201 *lenp = len; 202 return ((char *)lnbuf); 203 204 error: 205 *lenp = 0; 206 return (NULL); 207 } 208 209 static inline struct file * 210 grep_file_init(struct file *f) 211 { 212 213 #ifndef WITHOUT_GZIP 214 if (filebehave == FILE_GZIP && 215 (gzbufdesc = gzdopen(f->fd, "r")) == NULL) 216 goto error; 217 #endif 218 219 #ifndef WITHOUT_BZ2 220 if (filebehave == FILE_BZIP && 221 (bzbufdesc = BZ2_bzdopen(f->fd, "r")) == NULL) 222 goto error; 223 #endif 224 225 /* Fill read buffer, also catches errors early */ 226 if (grep_refill(f) != 0) 227 goto error; 228 229 /* Check for binary stuff, if necessary */ 230 if (!nulldataflag && binbehave != BINFILE_TEXT && 231 memchr(bufpos, '\0', bufrem) != NULL) 232 f->binary = true; 233 234 return (f); 235 error: 236 close(f->fd); 237 free(f); 238 return (NULL); 239 } 240 241 /* 242 * Opens a file for processing. 243 */ 244 struct file * 245 grep_open(const char *path) 246 { 247 struct file *f; 248 249 f = grep_malloc(sizeof *f); 250 memset(f, 0, sizeof *f); 251 if (path == NULL) { 252 /* Processing stdin implies --line-buffered. */ 253 lbflag = true; 254 f->fd = STDIN_FILENO; 255 } else if ((f->fd = open(path, O_RDONLY)) == -1) { 256 free(f); 257 return (NULL); 258 } 259 260 return (grep_file_init(f)); 261 } 262 263 /* 264 * Closes a file. 265 */ 266 void 267 grep_close(struct file *f) 268 { 269 270 close(f->fd); 271 272 /* Reset read buffer and line buffer */ 273 bufpos = buffer; 274 bufrem = 0; 275 276 free(lnbuf); 277 lnbuf = NULL; 278 lnbuflen = 0; 279 } 280