FileWatch，观测文件变化，源码地址：https://github.com/ThomasMonkman/filewatch
nlohmann::json，json封装解析，源码地址：https://github.com/nlohmann/json
optionparser，解析选项，源码地址：https://optionparser.sourceforge.net/

以上介绍的这三个工具都是我在学习FastDDS过程中看到的。他们的使用非常简单，都只有一个头文件。下面简单介绍一下如何使用：
FileWatch

#include <functional>
#include "FileWatch.hpp"using FileWatchHandle = std::unique_ptr<filewatch::FileWatch<std::string>>;void watch()
{std::cout << "watch" << std::endl;
}int main()
{std::function<void()> callback = watch;(new filewatch::FileWatch<std::string>("d:\\test.txt",[callback](const std::string& /*path*/, const filewatch::Event change_type){switch (change_type){case filewatch::Event::modified:callback();break;default:// No-opbreak;}}));getchar();return 0;
}

只需要设置一个文件路径和一个回调函数，当这个文件改动时，就会触发回调函数。这个可以做一些热更新配置的功能，当配置文件变动时，无需重启软件，就能读取新的配置。
nlohmann::json

#include "json.hpp"
#include <iostream>using Info = nlohmann::json;int main()
{Info info;std::cout << info.size() << std::endl;info["a"] = "b";std::cout << info["a"] << std::endl;auto iter = info.find("a");if (iter == info.end()) {std::cout << "not found" << std::endl;}else {std::cout << *iter << std::endl;}std::string s = R"({"name" : "nick","credits" : "123","ranking" : 1})";auto j = nlohmann::json::parse(s);std::cout << j["name"] << std::endl;std::string ss = j.dump();std::cout << "ss : " << ss << std::endl;Info j1;Info j2 = nlohmann::json::object();Info j3 = nlohmann::json::array();std::cout << j1.is_object() << std::endl;std::cout << j1.type_name() << std::endl;std::cout << j2.is_object() << std::endl;std::cout << j2.is_array() << std::endl;Info infoo{{"name", "darren"},{"credits", 123},{"ranking", 1}};std::cout << infoo["name"] << std::endl;std::cout << infoo.type_name() << std::endl;//遍历for (auto iter = infoo.begin(); iter != infoo.end(); iter++) {std::cout << iter.key() << " : " << iter.value() << std::endl;//std::cout << iter.value() << std::endl;//std::cout << *iter << std::endl;}system("pause");return 0;
}

更多内容可以参考csdn：https://blog.csdn.net/gaoyuelon/article/details/131482372?fromshare=blogdetail
optionparser

#include "optionparser.h"struct Arg : public option::Arg
{static void print_error(const char* msg1,const option::Option& opt,const char* msg2){fprintf(stderr, "%s", msg1);fwrite(opt.name, opt.namelen, 1, stderr);fprintf(stderr, "%s", msg2);}static option::ArgStatus Unknown(const option::Option& option,bool msg){if (msg){print_error("Unknown option '", option, "'\n");}return option::ARG_ILLEGAL;}static option::ArgStatus Required(const option::Option& option,bool msg){if (option.arg != 0 && option.arg[0] != 0){return option::ARG_OK;}if (msg){print_error("Option '", option, "' requires an argument\n");}return option::ARG_ILLEGAL;}static option::ArgStatus Numeric(const option::Option& option,bool msg){char* endptr = 0;if (option.arg != nullptr){strtol(option.arg, &endptr, 10);if (endptr != option.arg && *endptr == 0){return option::ARG_OK;}}if (msg){print_error("Option '", option, "' requires a numeric argument\n");}return option::ARG_ILLEGAL;}template<long min = 0, long max = std::numeric_limits<long>::max()>static option::ArgStatus NumericRange(const option::Option& option,bool msg){static_assert(min <= max, "NumericRange: invalid range provided.");char* endptr = 0;if (option.arg != nullptr){long value = strtol(option.arg, &endptr, 10);if (endptr != option.arg && *endptr == 0 &&value >= min && value <= max){return option::ARG_OK;}}if (msg){std::ostringstream os;os << "' requires a numeric argument in range ["<< min << ", " << max << "]" << std::endl;print_error("Option '", option, os.str().c_str());}return option::ARG_ILLEGAL;}static option::ArgStatus String(const option::Option& option,bool msg){if (option.arg != 0){return option::ARG_OK;}if (msg){print_error("Option '", option, "' requires an argument\n");}return option::ARG_ILLEGAL;}};enum  optionIndex
{UNKNOWN_OPT,HELP,SAMPLES,INTERVAL,ENVIRONMENT
};const option::Descriptor usage[] = {{ UNKNOWN_OPT, 0, "", "",                Arg::None,"Usage: HelloWorldExample <publisher|subscriber>\n\nGeneral options:" },{ HELP,    0, "h", "help",               Arg::None,      "  -h \t--help  \tProduce help message." },{ UNKNOWN_OPT, 0, "", "",                Arg::None,      "\nPublisher options:"},{ SAMPLES, 0, "s", "samples",            Arg::NumericRange<>,"  -s <num>, \t--samples=<num>  \tNumber of samples (0, default, infinite)." },{ INTERVAL, 0, "i", "interval",          Arg::NumericRange<>,"  -i <num>, \t--interval=<num>  \tTime between samples in milliseconds (Default: 100)." },{ ENVIRONMENT, 0, "e", "env",            Arg::None,       "  -e \t--env   \tLoad QoS from environment." },{ 0, 0, 0, 0, 0, 0 }
};int main(int argc, char **argv)
{argc -= (argc > 0);argv += (argc > 0); // skip program name argv[0] if presentoption::Stats stats(true, usage, argc, argv);std::vector<option::Option> options(stats.options_max);std::vector<option::Option> buffer(stats.buffer_max);option::Parser parse(true, usage, argc, argv, &options[0], &buffer[0]);try{if (parse.error()){throw 1;}if (options[HELP] || options[UNKNOWN_OPT]){throw 1;}// For backward compatibility count and sleep may be given positionallyif (parse.nonOptionsCount() > 3 || parse.nonOptionsCount() == 0){throw 2;}// Decide between publisher or subscriberconst char* type_name = parse.nonOption(0);// make sure is the first option.// type_name and buffer[0].name reference the original command line char array// type_name must precede any other arguments in the array.// Note buffer[0].arg may be null for non-valued options and is not reliable for// testing purposes.if (parse.optionsCount() && type_name >= buffer[0].name){throw 2;}if (strcmp(type_name, "publisher") == 0){std::cout << "publisher" << std::endl;}else if (strcmp(type_name, "subscriber") == 0){std::cout << "subscriber" << std::endl;}else{throw 2;}}catch (int error){if (error == 2){std::cerr << "ERROR: first argument must be <publisher|subscriber> followed by - or -- options"<< std::endl;}option::printUsage(fwrite, stdout, usage);return error;}getchar();return 0;
}

简单好用的工具能给我们工作带来很多便利，希望这些工具对你有用~