universe@10: /*
universe@34:  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS HEADER. 
universe@48:  * Copyright 2017 Mike Becker. All rights reserved.
universe@34:  * 
universe@34:  * Redistribution and use in source and binary forms, with or without
universe@34:  * modification, are permitted provided that the following conditions are met:
universe@34:  * 
universe@34:  * 1. Redistributions of source code must retain the above copyright
universe@34:  * notice, this list of conditions and the following disclaimer.
universe@34:  * 
universe@34:  * 2. Redistributions in binary form must reproduce the above copyright
universe@34:  * notice, this list of conditions and the following disclaimer in the
universe@34:  * documentation and/or other materials provided with the distribution.
universe@34:  * 
universe@34:  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
universe@34:  * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
universe@34:  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
universe@34:  * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
universe@34:  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
universe@34:  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
universe@34:  * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
universe@34:  * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
universe@34:  * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
universe@34:  * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 
universe@34:  *
universe@20:  * scanner.c
universe@10:  *
universe@10:  *  Created on: 23.05.2011
universe@20:  *      Author: Mike
universe@10:  */
universe@1: 
universe@1: 
universe@10: #include "scanner.h"
universe@10: #include "suffix_fnc.h"
universe@21: #include "bfile_heuristics.h"
universe@27: #include "regex_parser.h"
universe@23: #include <sys/stat.h>
universe@3: 
universe@41: typedef struct filelist filelist_t;
universe@23: 
universe@41: struct filelist {
universe@41:   char *displayname;
universe@41:   int displayname_len;
universe@41:   char *filename;
universe@41:   int st_mode;
universe@41:   filelist_t *next;
universe@41: };
universe@41: 
universe@41: filelist_t *buildFileList(scanner_t scanner, settings_t* settings,
universe@41:     filelist_t* list) {
universe@41:   
universe@23:   DIR *dirf;
universe@3:   struct dirent *entry;
universe@23:   struct stat statbuf;
universe@41:   
universe@23:   if ((dirf = opendir(scanner.dir)) == NULL) {
universe@34:     printf("%s", scanner.dir);
universe@23:     perror("  Directory access failed");
universe@23:     return 0;
universe@23:   }
universe@23: 
universe@23:   while ((entry = readdir(dirf)) != NULL) {
universe@3:     if (strcmp(entry->d_name, ".") != 0 && strcmp(entry->d_name, "..") != 0) {
universe@41:       
universe@41:       /* Create new filelist entry */
universe@41:       filelist_t *newentry = (filelist_t*) malloc(sizeof(filelist_t));
universe@42:       newentry->next = NULL;
universe@41:       
universe@42:       newentry->displayname_len = strlen(entry->d_name);
universe@42:       newentry->displayname = (char*) malloc(newentry->displayname_len+1);
universe@42:       memcpy(newentry->displayname, entry->d_name, newentry->displayname_len);
universe@42:       newentry->displayname[newentry->displayname_len] = 0;
universe@41:       
universe@42:       newentry->st_mode = 0;
universe@41:       
universe@40:       /* Construct absolute pathname string */
universe@41:       size_t dirnamelen = strlen(scanner.dir);
universe@42:       char *filename = (char*) malloc(2+dirnamelen+newentry->displayname_len);
universe@41:       memcpy(filename, scanner.dir, dirnamelen);
universe@41:       filename[dirnamelen] = settings->fileSeparator;
universe@42:       memcpy(filename+dirnamelen+1, entry->d_name, newentry->displayname_len);
universe@42:       filename[1+dirnamelen+newentry->displayname_len] = 0;
universe@42:       newentry->filename = filename;
universe@14: 
universe@22:       /* Check for subdirectory */
universe@23:       if (stat(filename, &statbuf) == 0) {
universe@42:         newentry->st_mode = statbuf.st_mode;
universe@23:       } else {
universe@23:         perror("  Error in stat call");
universe@3:         continue;
universe@3:       }
universe@42:       
universe@42:       if (list) {
universe@42:         // create fake root to have a pointer on the true root
universe@42:         filelist_t root;
universe@42:         root.next = list;
universe@42:         filelist_t *parent = &root;
universe@42:         while (parent->next &&
universe@42:             (strcasecmp(parent->next->displayname, newentry->displayname) < 0 ||
universe@42:               (!S_ISDIR(newentry->st_mode) && S_ISDIR(parent->next->st_mode))
universe@42:             ) &&
universe@42:             (!S_ISDIR(newentry->st_mode) || S_ISDIR(parent->next->st_mode))
universe@42:             ) {
universe@42:           parent = parent->next;
universe@42:         }
universe@42:         newentry->next = parent->next;
universe@42:         parent->next = newentry;
universe@42:         list = root.next;
universe@42:       } else {
universe@42:         list = newentry;
universe@42:       }
universe@41:     }
universe@41:   }
universe@41:   
universe@41:   closedir(dirf);
universe@41:   
universe@41:   return list;
universe@41: }
universe@3: 
universe@44: int scanDirectory(scanner_t scanner, settings_t* settings,
universe@44:     string_list_t* output) {
universe@41: 
universe@41:   int lines, a;
universe@41:   int lineSum = 0;
universe@41:   bool bfile;
universe@44:   char *outbuf;
universe@41: 
universe@41:   filelist_t *filelist = buildFileList(scanner, settings, NULL);
universe@41: 
universe@41:   while (filelist != NULL) {
universe@41: 
universe@41:     /* Scan subdirectories */
universe@42:     if (!S_ISREG(filelist->st_mode)) {
universe@44:       if (settings->recursive && S_ISDIR(filelist->st_mode)) {
universe@44:         string_list_t *recoutput = new_string_list_t();
universe@44:         lines = scanDirectory(
universe@44:             (scanner_t) {filelist->filename, scanner.spaces+1},
universe@44:             settings, recoutput);
universe@44:         lineSum += lines;
universe@44:         if (!settings->matchesOnly || recoutput->count > 0) {
universe@44:           outbuf = (char*) malloc(81);
universe@44:           snprintf(outbuf, 81, "%*s/%*s%13d lines\n",
universe@44:               filelist->displayname_len+scanner.spaces, filelist->displayname,
universe@44:               60-filelist->displayname_len-scanner.spaces-1, "", lines);
universe@44:           add_string(output, outbuf);
universe@44:           for (int i = 0 ; i < recoutput->count ; i++) {
universe@44:             add_string(output, recoutput->items[i]);
universe@44:           }
universe@44:         }
universe@44:         destroy_string_list_t(recoutput);
universe@44:       } else {
universe@44:         outbuf = (char*) malloc(81);
universe@44:         snprintf(outbuf, 81, "%*s\n", filelist->displayname_len+scanner.spaces,
universe@41:           filelist->displayname);
universe@44:         add_string(output, outbuf);
universe@41:       }
universe@41:     } else {
universe@30:       if ((settings->includeSuffixes->count == 0
universe@41:         || testSuffix(filelist->displayname, settings->includeSuffixes))
universe@41:         && !testSuffix(filelist->displayname, settings->excludeSuffixes)) {
universe@41: 
universe@25:         /* Count lines */
universe@25:         lines = 0;
universe@25:         bfile = false;
universe@25:         bfile_reset(settings->bfileHeuristics);
universe@27:         char line_buffer[REGEX_MAX_LINELENGTH];
universe@25:         int line_buffer_offset = 0;
universe@25: 
universe@41:         FILE *file = fopen(filelist->filename, "r");
universe@3:         if (file == NULL) {
universe@44:           outbuf = (char*) malloc(81);
universe@44:           snprintf(outbuf, 81, "%*s", filelist->displayname_len+scanner.spaces,
universe@41:               filelist->displayname);
universe@44:           add_string(output, outbuf);
universe@3:           perror("  File acces failed");
universe@41:         } else {
universe@41:           do {
universe@41:             a = fgetc(file);
universe@3: 
universe@41:             bfile = bfile_check(settings->bfileHeuristics, a);
universe@3: 
universe@41:             if (a == 10 || a == EOF) {
universe@41:               line_buffer[line_buffer_offset] = 0;
universe@41:               if (regex_parser_do(settings->regex, line_buffer) == 0) {
universe@41:                 /* Only subtract lines when matching has finished */
universe@41:                 if (!regex_parser_matching(settings->regex)) {
universe@41:                   lines -= settings->regex->matched_lines;
universe@41:                 }
universe@41:               }
universe@21: 
universe@41:               line_buffer_offset = 0;
universe@41:               lines++;
universe@41:             } else {
universe@41:               if (line_buffer_offset < REGEX_MAX_LINELENGTH) {
universe@41:                 line_buffer[line_buffer_offset] = a;
universe@41:                 line_buffer_offset++;
universe@41:               } else {
universe@41:                 line_buffer[line_buffer_offset-1] = 0;
universe@41:                 settings->confusing_lnlen = true;
universe@28:               }
universe@28:             }
universe@41:           } while (!bfile && a != EOF);
universe@41:           fclose(file);
universe@25: 
universe@41:           /* Print and sum line count */
universe@41:           if (bfile) {
universe@41:             if (!settings->matchesOnly) {
universe@44:               outbuf = (char*) malloc(81);
universe@44:               snprintf(outbuf, 81,
universe@44:                   "%*s%*s%19s\n", filelist->displayname_len+scanner.spaces,
universe@41:                   filelist->displayname,
universe@41:                   60-filelist->displayname_len-scanner.spaces, "", "binary");
universe@44:               add_string(output, outbuf);
universe@41:             }
universe@25:           } else {
universe@41:             lineSum += lines;
universe@44:             outbuf = (char*) malloc(81);
universe@44:             snprintf(outbuf, 81, "%*s%*s%13d lines\n",
universe@41:                 filelist->displayname_len+scanner.spaces, filelist->displayname,
universe@41:                 60-filelist->displayname_len-scanner.spaces, "", lines);
universe@44:             add_string(output, outbuf);
universe@3:           }
universe@21:         }
universe@16:       } else {
universe@3:         if (!settings->matchesOnly) {
universe@22:           /* Print hint */
universe@44:           outbuf = (char*) malloc(81);
universe@44:           snprintf(outbuf, 81, "%*s%*s%19s\n",
universe@41:               filelist->displayname_len+scanner.spaces, filelist->displayname,
universe@41:               60-filelist->displayname_len-scanner.spaces, "", "no match");
universe@44:           add_string(output, outbuf);
universe@3:         }
universe@3:       }
universe@3:     }
universe@41:     
universe@41:     free(filelist->filename);
universe@41:     free(filelist->displayname);
universe@41:     filelist_t *freethis = filelist;
universe@41:     filelist = filelist->next;
universe@41:     free(freethis);
universe@3:   }
universe@23: 
universe@3:   return lineSum;
universe@3: }