Tue, 21 May 2013 13:19:37 +0200
replaced direct scanning of directories with a two-step approach (first: create filename list, second: scan)
universe@10 | 1 | /* |
universe@34 | 2 | * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS HEADER. |
universe@36 | 3 | * Copyright 2013 Mike Becker. All rights reserved. |
universe@34 | 4 | * |
universe@34 | 5 | * Redistribution and use in source and binary forms, with or without |
universe@34 | 6 | * modification, are permitted provided that the following conditions are met: |
universe@34 | 7 | * |
universe@34 | 8 | * 1. Redistributions of source code must retain the above copyright |
universe@34 | 9 | * notice, this list of conditions and the following disclaimer. |
universe@34 | 10 | * |
universe@34 | 11 | * 2. Redistributions in binary form must reproduce the above copyright |
universe@34 | 12 | * notice, this list of conditions and the following disclaimer in the |
universe@34 | 13 | * documentation and/or other materials provided with the distribution. |
universe@34 | 14 | * |
universe@34 | 15 | * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" |
universe@34 | 16 | * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
universe@34 | 17 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE |
universe@34 | 18 | * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE |
universe@34 | 19 | * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL |
universe@34 | 20 | * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR |
universe@34 | 21 | * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER |
universe@34 | 22 | * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, |
universe@34 | 23 | * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE |
universe@34 | 24 | * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
universe@34 | 25 | * |
universe@20 | 26 | * scanner.c |
universe@10 | 27 | * |
universe@10 | 28 | * Created on: 23.05.2011 |
universe@20 | 29 | * Author: Mike |
universe@10 | 30 | */ |
universe@1 | 31 | |
universe@1 | 32 | |
universe@10 | 33 | #include "scanner.h" |
universe@10 | 34 | #include "suffix_fnc.h" |
universe@21 | 35 | #include "bfile_heuristics.h" |
universe@27 | 36 | #include "regex_parser.h" |
universe@23 | 37 | #include <sys/stat.h> |
universe@3 | 38 | |
universe@41 | 39 | typedef struct filelist filelist_t; |
universe@23 | 40 | |
universe@41 | 41 | struct filelist { |
universe@41 | 42 | char *displayname; |
universe@41 | 43 | int displayname_len; |
universe@41 | 44 | char *filename; |
universe@41 | 45 | int st_mode; |
universe@41 | 46 | filelist_t *next; |
universe@41 | 47 | }; |
universe@41 | 48 | |
universe@41 | 49 | filelist_t *buildFileList(scanner_t scanner, settings_t* settings, |
universe@41 | 50 | filelist_t* list) { |
universe@41 | 51 | |
universe@23 | 52 | DIR *dirf; |
universe@3 | 53 | struct dirent *entry; |
universe@23 | 54 | struct stat statbuf; |
universe@41 | 55 | filelist_t *listentry = list; |
universe@41 | 56 | |
universe@23 | 57 | if ((dirf = opendir(scanner.dir)) == NULL) { |
universe@34 | 58 | printf("%s", scanner.dir); |
universe@23 | 59 | perror(" Directory access failed"); |
universe@23 | 60 | return 0; |
universe@23 | 61 | } |
universe@23 | 62 | |
universe@23 | 63 | while ((entry = readdir(dirf)) != NULL) { |
universe@3 | 64 | if (strcmp(entry->d_name, ".") != 0 && strcmp(entry->d_name, "..") != 0) { |
universe@41 | 65 | |
universe@41 | 66 | /* Create new filelist entry */ |
universe@41 | 67 | filelist_t *newentry = (filelist_t*) malloc(sizeof(filelist_t)); |
universe@41 | 68 | // TODO: don't just append - create a sorted list! |
universe@41 | 69 | if (listentry) { |
universe@41 | 70 | listentry->next = newentry; |
universe@41 | 71 | } |
universe@41 | 72 | listentry = newentry; |
universe@41 | 73 | if (!list) { |
universe@41 | 74 | list = listentry; |
universe@41 | 75 | } |
universe@41 | 76 | |
universe@41 | 77 | listentry->next = NULL; |
universe@41 | 78 | |
universe@41 | 79 | listentry->displayname_len = strlen(entry->d_name); |
universe@41 | 80 | listentry->displayname = (char*) malloc(listentry->displayname_len+1); |
universe@41 | 81 | memcpy(listentry->displayname, entry->d_name, listentry->displayname_len); |
universe@41 | 82 | listentry->displayname[listentry->displayname_len] = 0; |
universe@41 | 83 | |
universe@41 | 84 | listentry->st_mode = 0; |
universe@41 | 85 | |
universe@40 | 86 | /* Construct absolute pathname string */ |
universe@41 | 87 | size_t dirnamelen = strlen(scanner.dir); |
universe@41 | 88 | char *filename = (char*) malloc(2+dirnamelen+listentry->displayname_len); |
universe@41 | 89 | memcpy(filename, scanner.dir, dirnamelen); |
universe@41 | 90 | filename[dirnamelen] = settings->fileSeparator; |
universe@41 | 91 | memcpy(filename+dirnamelen+1, entry->d_name, listentry->displayname_len); |
universe@41 | 92 | filename[1+dirnamelen+listentry->displayname_len] = 0; |
universe@41 | 93 | listentry->filename = filename; |
universe@14 | 94 | |
universe@22 | 95 | /* Check for subdirectory */ |
universe@23 | 96 | if (stat(filename, &statbuf) == 0) { |
universe@41 | 97 | listentry->st_mode = statbuf.st_mode; |
universe@23 | 98 | } else { |
universe@23 | 99 | perror(" Error in stat call"); |
universe@3 | 100 | continue; |
universe@3 | 101 | } |
universe@41 | 102 | } |
universe@41 | 103 | } |
universe@41 | 104 | |
universe@41 | 105 | closedir(dirf); |
universe@41 | 106 | |
universe@41 | 107 | return list; |
universe@41 | 108 | } |
universe@3 | 109 | |
universe@41 | 110 | int scanDirectory(scanner_t scanner, settings_t* settings) { |
universe@41 | 111 | |
universe@41 | 112 | int lines, a; |
universe@41 | 113 | int lineSum = 0; |
universe@41 | 114 | bool bfile; |
universe@41 | 115 | |
universe@41 | 116 | filelist_t *filelist = buildFileList(scanner, settings, NULL); |
universe@41 | 117 | |
universe@41 | 118 | while (filelist != NULL) { |
universe@41 | 119 | |
universe@41 | 120 | /* Scan subdirectories */ |
universe@41 | 121 | if (!(filelist->st_mode & S_IFREG)) { |
universe@41 | 122 | printf("%*s\n", filelist->displayname_len+scanner.spaces, |
universe@41 | 123 | filelist->displayname); |
universe@41 | 124 | if (settings->recursive && (filelist->st_mode & S_IFDIR)) { |
universe@41 | 125 | scanDirectory((scanner_t) {filelist->filename, scanner.spaces+1}, |
universe@41 | 126 | settings); |
universe@41 | 127 | } |
universe@41 | 128 | } else { |
universe@30 | 129 | if ((settings->includeSuffixes->count == 0 |
universe@41 | 130 | || testSuffix(filelist->displayname, settings->includeSuffixes)) |
universe@41 | 131 | && !testSuffix(filelist->displayname, settings->excludeSuffixes)) { |
universe@41 | 132 | |
universe@25 | 133 | /* Count lines */ |
universe@25 | 134 | lines = 0; |
universe@25 | 135 | bfile = false; |
universe@25 | 136 | bfile_reset(settings->bfileHeuristics); |
universe@27 | 137 | char line_buffer[REGEX_MAX_LINELENGTH]; |
universe@25 | 138 | int line_buffer_offset = 0; |
universe@25 | 139 | |
universe@41 | 140 | FILE *file = fopen(filelist->filename, "r"); |
universe@3 | 141 | if (file == NULL) { |
universe@41 | 142 | printf("%*s", filelist->displayname_len+scanner.spaces, |
universe@41 | 143 | filelist->displayname); |
universe@3 | 144 | perror(" File acces failed"); |
universe@41 | 145 | } else { |
universe@41 | 146 | do { |
universe@41 | 147 | a = fgetc(file); |
universe@3 | 148 | |
universe@41 | 149 | bfile = bfile_check(settings->bfileHeuristics, a); |
universe@3 | 150 | |
universe@41 | 151 | if (a == 10 || a == EOF) { |
universe@41 | 152 | line_buffer[line_buffer_offset] = 0; |
universe@41 | 153 | if (regex_parser_do(settings->regex, line_buffer) == 0) { |
universe@41 | 154 | /* Only subtract lines when matching has finished */ |
universe@41 | 155 | if (!regex_parser_matching(settings->regex)) { |
universe@41 | 156 | lines -= settings->regex->matched_lines; |
universe@41 | 157 | } |
universe@41 | 158 | } |
universe@21 | 159 | |
universe@41 | 160 | line_buffer_offset = 0; |
universe@41 | 161 | lines++; |
universe@41 | 162 | } else { |
universe@41 | 163 | if (line_buffer_offset < REGEX_MAX_LINELENGTH) { |
universe@41 | 164 | line_buffer[line_buffer_offset] = a; |
universe@41 | 165 | line_buffer_offset++; |
universe@41 | 166 | } else { |
universe@41 | 167 | line_buffer[line_buffer_offset-1] = 0; |
universe@41 | 168 | settings->confusing_lnlen = true; |
universe@28 | 169 | } |
universe@28 | 170 | } |
universe@41 | 171 | } while (!bfile && a != EOF); |
universe@41 | 172 | fclose(file); |
universe@25 | 173 | |
universe@41 | 174 | /* Print and sum line count */ |
universe@41 | 175 | if (bfile) { |
universe@41 | 176 | if (!settings->matchesOnly) { |
universe@41 | 177 | printf("%*s%*s%19s\n", filelist->displayname_len+scanner.spaces, |
universe@41 | 178 | filelist->displayname, |
universe@41 | 179 | 60-filelist->displayname_len-scanner.spaces, "", "binary"); |
universe@41 | 180 | } |
universe@25 | 181 | } else { |
universe@41 | 182 | lineSum += lines; |
universe@41 | 183 | printf("%*s%*s%13d lines\n", |
universe@41 | 184 | filelist->displayname_len+scanner.spaces, filelist->displayname, |
universe@41 | 185 | 60-filelist->displayname_len-scanner.spaces, "", lines); |
universe@3 | 186 | } |
universe@21 | 187 | } |
universe@16 | 188 | } else { |
universe@3 | 189 | if (!settings->matchesOnly) { |
universe@22 | 190 | /* Print hint */ |
universe@40 | 191 | printf("%*s%*s%19s\n", |
universe@41 | 192 | filelist->displayname_len+scanner.spaces, filelist->displayname, |
universe@41 | 193 | 60-filelist->displayname_len-scanner.spaces, "", "no match"); |
universe@3 | 194 | } |
universe@3 | 195 | } |
universe@3 | 196 | } |
universe@41 | 197 | |
universe@41 | 198 | free(filelist->filename); |
universe@41 | 199 | free(filelist->displayname); |
universe@41 | 200 | filelist_t *freethis = filelist; |
universe@41 | 201 | filelist = filelist->next; |
universe@41 | 202 | free(freethis); |
universe@3 | 203 | } |
universe@23 | 204 | |
universe@3 | 205 | return lineSum; |
universe@3 | 206 | } |