xfarrow/dircomp
1
0
Fork 0
mirror of https://github.com/xfarrow/dircomp.git synced 2025-06-05 21:49:19 +02:00
Code Issues Packages Projects Releases Wiki Activity

Significantly improved performance. Hash function removed.

Browse Source 
Closes #2
This commit is contained in:
Alessandro Ferro
2023-04-15 22:19:04 +02:00
parent 6e5ff7cfc5
commit 98ae550a9e
3 changed files with 51 additions and 44 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
build.sh
View File

@@ -1,2 +1,2 @@
#! /bin/bash #! /bin/bash
gcc -o3 dircomp.c -lssl -lcrypto -o dircomp gcc -o3 dircomp.c -o dircomp

89
dircomp.c
View File

@@ -81,8 +81,9 @@ bool analyze_directories(char* directory_to_analyze_1, char* directory_to_analyz
printf("\nAnalyzing directories %s %s\n", directory_to_analyze_1, directory_to_analyze_2); printf("\nAnalyzing directories %s %s\n", directory_to_analyze_1, directory_to_analyze_2);
bool is_directory_equal = true; bool is_directory_equal = true;
int stat_result; int stat_result, file_equality_result;
char fullpath_file_helper[512]; char fullpath_file_helper[512];
char fullpath_file_helper2[512];
char* subdirectory1; char* subdirectory1;
char* subdirectory2; char* subdirectory2;
struct dirent *element; struct dirent *element;
@@ -109,27 +110,32 @@ bool analyze_directories(char* directory_to_analyze_1, char* directory_to_analyz
, directory_to_analyze_2); , directory_to_analyze_2);
} }
// Check if files have the same signature // Check if the files are the same
else else
{ {
strcpy(fullpath_file_helper, directory_to_analyze_1); strcpy(fullpath_file_helper, directory_to_analyze_1);
strcat(fullpath_file_helper, "/"); strcat(fullpath_file_helper, "/");
strcat(fullpath_file_helper, element->d_name); strcat(fullpath_file_helper, element->d_name);
unsigned char *hash_file_1 = get_sha1_file(fullpath_file_helper); strcpy(fullpath_file_helper2, directory_to_analyze_2);
strcpy(fullpath_file_helper, directory_to_analyze_2); strcat(fullpath_file_helper2, "/");
strcat(fullpath_file_helper, "/"); strcat(fullpath_file_helper2, element->d_name);
strcat(fullpath_file_helper, element->d_name); file_equality_result = are_files_equal(fullpath_file_helper, fullpath_file_helper2);
unsigned char *hash_file_2 = get_sha1_file(fullpath_file_helper); if (file_equality_result != 1)
if (strcmp(hash_file_1, hash_file_2) != 0)
{ {
is_directory_equal = false; is_directory_equal = false;
if (arguments->v == true) if(file_equality_result == 0 && arguments->v == true)
printf("File %s in %s has a different signature in %s\n", element->d_name {
, directory_to_analyze_1 printf("File %s in %s is different in %s\n" , element->d_name
, directory_to_analyze_2); , directory_to_analyze_1
, directory_to_analyze_2);
}
else if(file_equality_result == -1)
{
printf("Error while comparing file %s in the directories %s, %s", element->d_name
, directory_to_analyze_1
, directory_to_analyze_2);
}
} }
free(hash_file_1);
free(hash_file_2);
} }
} }
@@ -148,7 +154,7 @@ bool analyze_directories(char* directory_to_analyze_1, char* directory_to_analyz
struct stat *dummy_structure = malloc(sizeof(struct stat)); struct stat *dummy_structure = malloc(sizeof(struct stat));
stat_result = stat(fullpath_file_helper, dummy_structure); stat_result = stat(fullpath_file_helper, dummy_structure);
free(dummy_structure); free(dummy_structure);
if (stat_result == -1) if (stat_result == -1) // directory does not exist
{ {
is_directory_equal = false; is_directory_equal = false;
if (arguments->v == true) if (arguments->v == true)
@@ -225,34 +231,37 @@ bool analyze_directories(char* directory_to_analyze_1, char* directory_to_analyz
return is_directory_equal; return is_directory_equal;
} }
unsigned char *get_sha1_file(char *filename) int are_files_equal(char* filename1, char* filename2){
{ struct stat stat1, stat2;
FILE *f = fopen(filename, "rb");
if (f == NULL)
{
fprintf(stderr, "Couldn't open %s\n", filename);
exit(-1);
}
// For a matter of efficiency, we do not read if ( stat(filename1, &stat1) != 0 || stat(filename2, &stat2) != 0)
// the whole file at once. It'd be heavy on RAM. return -1; // error opening files
// Instead, we read BYTES_TO_READ_AT_ONCE at time
#define BYTES_TO_READ_AT_ONCE 512000 // 500KiB if(stat1.st_size != stat2.st_size)
unsigned int bytes; // how many bytes we have actually read from fread return 0; // files are not the same as they have a different dimension
#if BYTES_TO_READ_AT_ONCE > UINT_MAX
#error Trying to read more bytes than what is possible to handle. Recompile using unsigned long or reduce BYTES_TO_READ_AT_ONCE FILE *file1 = fopen(filename1, "rb");
#endif FILE *file2 = fopen(filename2, "rb");
SHA_CTX context; if (file1 == NULL || file2 == NULL)
unsigned char *hash = malloc(SHA_DIGEST_LENGTH * sizeof(unsigned char)); // result will be here
unsigned char databuffer[BYTES_TO_READ_AT_ONCE];
SHA1_Init(&context);
while ((bytes = fread(databuffer, 1, BYTES_TO_READ_AT_ONCE, f)) != 0)
{ {
SHA1_Update(&context, databuffer, bytes); return -1; // error opening files
} }
SHA1_Final(hash, &context); #define BYTES_TO_READ_AT_ONCE 512000
fclose(f); unsigned int bytes;
return hash; #if BYTES_TO_READ_AT_ONCE > UINT_MAX
#error Trying to read more bytes than what is possible to handle. Recompile using unsigned long or reduce BYTES_TO_READ_AT_ONCE
#endif
unsigned char databuffer1[BYTES_TO_READ_AT_ONCE];
unsigned char databuffer2[BYTES_TO_READ_AT_ONCE];
while ((bytes = fread(databuffer1, 1, BYTES_TO_READ_AT_ONCE, file1)) != 0)
{
fread(databuffer2, 1, BYTES_TO_READ_AT_ONCE, file2);
if(strcmp(databuffer1, databuffer2) != 0)
return 0;
}
fclose(file1);
fclose(file2);
return 1;
} }
void print_help(void) void print_help(void)

4
dircomp.h
View File

@@ -14,7 +14,6 @@ FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
#include <unistd.h> #include <unistd.h>
#include <stdbool.h> #include <stdbool.h>
#include <dirent.h> #include <dirent.h>
#include <openssl/sha.h>
#include <string.h> #include <string.h>
#include <limits.h> #include <limits.h>
#include <sys/types.h> #include <sys/types.h>
@@ -33,8 +32,7 @@ struct arguments get_arguments(int, char**);
// Reference: https://www.gnu.org/software/libc/manual/html_node/Directory-Entries.html // Reference: https://www.gnu.org/software/libc/manual/html_node/Directory-Entries.html
bool analyze_directories(char*, char*, struct arguments*); bool analyze_directories(char*, char*, struct arguments*);
// Reference: https://www.openssl.org/docs/man1.1.1/man3/SHA512_Init.html int are_files_equal(char*, char*);
unsigned char* get_sha1_file(char *);
void print_help(void); void print_help(void);