Optimization

.gitignore

@@ -1,3 +1,5 @@
 /.vscode/
 /demo_sha1
 /test_sha1
+bin/
+build/

CMakeLists.txt

@@ -0,0 +1,67 @@
+cmake_minimum_required(VERSION 3.18)
+project("SHA1_tests")
+
+set(CMAKE_CXX_STANDARD 17)
+set(CMAKE_CXX_FLAGS "-Wall")
+set(CMAKE_CXX_STANDARD_REQUIRED ON)
+set(CMAKE_CXX_EXTENSIONS OFF)
+
+set(EXECUTABLE_OUTPUT_PATH ${CMAKE_SOURCE_DIR}/bin)
+
+get_filename_component(PARENT_DIR ../ ABSOLUTE)
+
+
+if(${CMAKE_SYSTEM_NAME} STREQUAL "Windows")
+	add_definitions(-D_WIN32)
+	add_compile_options("/O2")
+else()
+	add_definitions(-D_LINUX)
+   if(${CMAKE_SYSTEM_NAME} STREQUAL "Linux")
+		message(WARNING "Assuming system in Linux")
+	endif()
+	add_compile_options("-O3")
+	add_compile_options("-g")
+endif()
+
+#---------------------------------------------------------------------
+
+add_library(SHA1_lib INTERFACE)
+
+target_include_directories(
+	SHA1_lib
+	INTERFACE
+	"${CMAKE_SOURCE_DIR}"
+)
+
+#---------------------------------------------------------------------
+
+set(TEST test_sha1)
+
+add_executable(
+	"${TEST}" 
+	test_sha1.cpp
+	test_sha1_file.cpp
+)
+
+target_link_libraries(
+	"${TEST}"
+	PRIVATE
+	SHA1_lib
+)
+
+#---------------------------------------------------------------------
+
+set(TEST demo_sha1)
+
+add_executable(
+	"${TEST}" 
+	demo_sha1.cpp
+)
+
+target_link_libraries(
+	"${TEST}"
+	PRIVATE
+	SHA1_lib
+)
+
+#---------------------------------------------------------------------

Makefile

@@ -5,10 +5,10 @@ RM    = rm -f
 all: demo_sha1 test_sha1
 
 demo_sha1: demo_sha1.cpp sha1.cpp sha1.hpp
-	$(CXX) $(CXXFLAGS) $(CPPFLAGS) -Wall -Wextra -std=c++11 -o $@ demo_sha1.cpp sha1.cpp
+	$(CXX) $(CXXFLAGS) $(CPPFLAGS) -Wall -Wextra -std=c++11 -g -O3 -o $@ demo_sha1.cpp
 
 test_sha1: test_sha1.cpp test_sha1_file.cpp sha1.hpp
-	$(CXX) $(CXXFLAGS) $(CPPFLAGS) -Wall -Wextra -std=c++11 -o $@ test_sha1.cpp test_sha1_file.cpp sha1.cpp
+	$(CXX) $(CXXFLAGS) $(CPPFLAGS) -Wall -Wextra -std=c++11 -g -O3 -o $@ test_sha1.cpp test_sha1_file.cpp
 
 check: test_sha1
 	./test_sha1

README.md

@@ -0,0 +1,36 @@
+# SHA-1 implementation in C++
+
+## Warning
+
+Do not use SHA-1 unless you have to. [SHA-1 is practically broken](https://en.wikipedia.org/wiki/SHA-1#Birthday-Near-Collision_Attack_%E2%80%93_first_practical_chosen-prefix_attack). 
+Use a hash function from the [SHA-2](https://en.wikipedia.org/wiki/SHA-2) or [SHA-3](https://en.wikipedia.org/wiki/SHA-3) family instead.
+
+- Despite been proven that SHA-1 has collision attacks, SHA-1 is still extensively used, for example in
+websocket protocol.
+
+- For security related purposes one should use a combination of at least two hashes. For example in pseudocode:
+   - sha2(sha1(data)) not very good
+   - sha1(data)+sha2(data) better
+   - sha1(data.part1)+sha2(data.part2)+sha3(data) strong (where data=data.part1+data.part2)
+
+
+## To build the tests
+```
+cd tests
+mkdir build
+cd build
+cmake .. 
+```
+
+## License
+
+100% Public Domain
+
+## Authors
+
+- Steve Reid (Original C Code)
+- [Bruce Guenter](http://untroubled.org/) (Small changes to fit into bglibs)
+- [Volker Diels-Grabsch](https://njh.eu/) (Translation to simpler C++ Code)
+- [Eugene Hopkinson](https://riot.so/) (Safety improvements)
+- [Zlatko Michailov](http://zlatko.michailov.org) (Header-only library)
+- [Dan Machado](dan-machado@yandex.com) (Optimization)

performance/CMakeLists.txt

@@ -0,0 +1,88 @@
+cmake_minimum_required(VERSION 3.18)
+project("SHA1_tests")
+
+set(CMAKE_CXX_STANDARD 17)
+set(CMAKE_CXX_FLAGS "-Wall")
+set(CMAKE_CXX_STANDARD_REQUIRED ON)
+set(CMAKE_CXX_EXTENSIONS OFF)
+
+set(EXECUTABLE_OUTPUT_PATH ${CMAKE_SOURCE_DIR}/bin)
+
+get_filename_component(PARENT_DIR ../ ABSOLUTE)
+
+
+if(${CMAKE_SYSTEM_NAME} STREQUAL "Windows")
+	add_definitions(-D_WIN32)
+	add_compile_options("/O2")
+else()
+	add_definitions(-D_LINUX)
+   if(${CMAKE_SYSTEM_NAME} STREQUAL "Linux")
+		message(WARNING "Assuming system in Linux")
+	endif()
+	add_compile_options("-O3")
+	add_compile_options("-g")
+endif()
+
+#---------------------------------------------------------------------
+
+include(FetchContent)
+
+FetchContent_Declare(TimeProfiler
+    GIT_REPOSITORY "https://github.com/volatilflerovium/time_profiler_visualizer"
+    GIT_TAG "origin/main"
+)
+
+FetchContent_MakeAvailable(TimeProfiler)
+
+#---------------------------------------------------------------------
+
+add_library(SHA1_lib INTERFACE)
+
+target_include_directories(
+	SHA1_lib
+	INTERFACE
+	"${CMAKE_SOURCE_DIR}/.."
+)
+
+#---------------------------------------------------------------------
+
+if(${CMAKE_SYSTEM_NAME} STREQUAL "Linux")
+
+	execute_process(COMMAND ls /tmp/1G_sha1_test_file.txt
+		ERROR_VARIABLE error
+		OUTPUT_VARIABLE output
+	)
+
+	if(NOT ("${error}" STREQUAL ""))
+		message("Creating a 1G file of random data in /tmp directory...")
+		execute_process(COMMAND dd if=/dev/urandom of=/tmp/1G_sha1_test_file.txt bs=1G count=1 iflag=fullblock
+			ERROR_VARIABLE error
+			OUTPUT_VARIABLE output
+		)
+	endif()
+	add_compile_definitions(TEST_FILE="/tmp/1G_sha1_test_file.txt")
+endif()
+
+#---------------------------------------------------------------------
+
+set(TEST "big_file_test")
+
+add_executable(
+	"${TEST}"
+	big_file_test.cpp
+)
+
+target_link_libraries(
+	"${TEST}"
+	PRIVATE
+	TimeProfiler
+	SHA1_lib
+)
+
+target_compile_definitions(
+	"${TEST}"
+	PRIVATE
+	-DENABLE_STOPWATCH
+)
+
+#---------------------------------------------------------------------

performance/README.md

@@ -0,0 +1,20 @@
+# Profiling
+
+Charts were generated using (time_profiler_visualizer)[https://github.com/volatilflerovium/time_profiler_visualizer]
+
+## Performance
+
+Performace tests run under Debian 11 with AMD Ryzen7 processor.
+
+For a file of 1GB
+   - SHA1 previous version: ~1870ms
+   - Linux command sha1sum: ~1471ms
+   - SHA1 optimized: ~1340ms (under Linux with mmap) and ~1530ms (for Linux and Windows using std::fread)
+
+On a 1 GB file (in seconds)
+
+[![performance-1GB-file.png](https://i.postimg.cc/GpPT4tPZ/performance-1GB-file.png)](https://postimg.cc/7C6ZcxWV)
+
+On incremental string (in microseconds)
+
+[![performance.png](https://i.postimg.cc/J4dN8sK0/performance.png)](https://postimg.cc/YhYGfq5B)

performance/big_file_test.cpp

@@ -0,0 +1,47 @@
+#include <iostream>
+
+#include "sha1.hpp"
+
+#include "time_profiler/time_profiler.h"
+
+using Profiler=tprofiler::TimeProfiler<std::chrono::seconds>;
+
+//====================================================================
+
+int main()
+{
+	const char* testFile="";
+	#ifndef TEST_FILE
+		std::cout<<TEST_FILE<<" no found.\n";
+		return 0;
+	#else
+		testFile= TEST_FILE;
+	#endif
+
+	Profiler watch("sha1_new_1file", "#5400AA");
+
+	SHA1 sha1;
+
+	std::string hash;
+
+	for(int i=0; i<10; i++){
+		std::cout<<"sha1sum command: ";
+
+		watch.start();
+			system("sha1sum  /tmp/1G_sha1_test_file.txt");
+		watch.takeSample(true);
+
+		std::cout<<"\nSHA1 class: ";
+		watch.start();
+			if(sha1(testFile, hash)){
+				std::cout<<hash<<"\n";
+			}
+			else{
+				std::cout<<"Error: "<<sha1.getError()<<"\n";
+			}
+		watch.takeSample(true);
+		std::cout<<"\n";
+	}
+
+	return 0;
+}