当我们用C++开发一些C++控制台小工具时，会需要一些用户输入的参数来决定程序如何工作和执行，而用户输入参数的方式大部分都是采用命令行参数的方式。
比如上一篇文章 玩转Windows服务系列——命令行管理Windows服务 中介绍的sc和net工具。
既然命令行参数这么普遍也这么有用，那么就有必要学习一下如何解析命令行参数。
如何解析命令行参数
那么命令行参数要如何解析呢。
最最最笨的办法就是每次程序中需要解析命令行参数时，就写一堆解析的代码，但是这样的效率其实很低的，不如使用开源库中的帮助类来解析。
我所了解的开源库的帮助类有
- ACE库的ACE_Get_Opt
- boost的Program_options
ACE_Get_Opt类的主要使用方法介绍
那么我们主要看一下ACE_Get_Opt类
先看它的构造方法
ACE_Get_Opt (int argc, ACE_TCHAR **argv, const ACE_TCHAR *optstring = ACE_TEXT (""), int skip_args = 1, int report_errors = 0, int ordering = PERMUTE_ARGS, int long_only = 0);
下面是代码中的注释
/** * Constructor initializes the command line to be parsed. All information * for parsing must be supplied to this constructor. * * @param argc The number of @a argv elements to parse. * @param argv Command line tokens, such as would be passed * to @c main(). * @param optstring Nul-terminated string containing the legitimate * short option characters. A single colon ":" * following an option character means the option * requires an argument. A double colon "::" following * an option character means the argument is optional. * The argument is taken from the rest of the current * @a argv element, or from the following @a argv * element (only valid for required arguments; * optional arguments must always reside in the same * @a argv element). The argument value, if any is * returned by the @c opt_arg() method. * @a optstring can be extended by adding long options * with corresponding short options via the * @c long_option() method. If the short option * already appears in @a optstring, the argument * characteristics must match, otherwise it is added. * See @c long_option() for more information. * If 'W', followed by a semi-colon ';' appears in * @a optstring, then any time a 'W' appears on the * command line, the following argument is treated as * a long option. For example, if the command line * contains "program -W foo", "foo" is treated as a * long option, that is, as if "program --foo" had * been passed. * The following characters can appear in @a optstring * before any option characters, with the described * effect: * - '+' changes the @a ordering to @a REQUIRE_ORDER. * - '-' changes the @a ordering to @a RETURN_IN_ORDER. * - ':' changes the return value from @c operator() * and get_opt() from '?' to ':' when an option * requires an argument but none is specified. * * @param skip_args Optional (default 1). The specified number of * initial elements in @a argv are skipped before * parsing begins. Thus, the default prevents * @a argv[0] (usually the command name) from being * parsed. @a argc includes all @a argv elements, * including any skipped elements. * @param report_errors Optional, if non-zero then parsing errors cause * an error message to be displayed from the * @c operator() method before it returns. The * error message is suppressed if this argument is 0. * This setting also controls whether or not an error * message is displayed in @c long_option() encounters * an error. * @param ordering Optional (default is @c PERMUTE_ARGS); determines * how the @a argv elements are processed. This argument * is overridden by two factors: * -# The @c POSIXLY_CORRECT environment variable. If * this environment variable is set, the ordering * is changed to @c REQUIRE_ORDER. * -# Leading characters in @a optstring (see above). * Any leading ordering characters override both * the @a ordering argument and any effect of the * @c POSIXLY_CORRECT environment variable. * @param long_only Optional. If non-zero, then long options can be * specified using a single '-' on the command line. * If the token is not a long option, it is processed * as usual, that is, as a short option or set of * short options. * * Multiple short options can be combined as long as only the last * one can takes an argument. For example, if @a optstring is defined as * @c "abc:" or @c "abc::" then the command line @e "program -abcxxx" short * options @e a, @e b, and @e c are found with @e "xxx" as the argument for * @e c. * However, if the command line is specified as @e "program -acb" only * options @e a and @e c are found with @e "b" as the argument for @e c. * Also, for options with optional arguments, that is, those followed by * "::", the argument must be in the same @a argv element, so "program -abc * xxx" will only find "xxx" as the argument for @e c if @a optstring is * specified as @c "abc:" not @c "abc::". */
构造函数共有七个参数，
argc、argv为程序入口函数main方法的参数，也就是程序的命令行。
optstring 为指定的参数选项，并且是一个字符作为一个选项，主要包含三种形式
- 单纯的一个字符选项，比如 s，表示此选项后面不能添加选项的参数
- 一个字符选项后跟一个冒号，比如 s:，表示此选项后面会有一个参数
- 一个字符后面跟两个冒号，比如 s::，表示此选项后面即可以有参数也可以无参数
skip_args 表示从argv的第几个元素开始，默认为1，一般情况下argv[0]为程序的path
report_errors遇到不识别的参数时是否提示错误
long_only表示是否只包含字符串的选项参数。
下面解释一下 字符选项和字符串选项，也就是 short option 和 long option.
- short option 以 字符’-’开始，比如 -s
- long option 以两个字符’-’开始，日入 --server
这是在默认的情况下，也就是long_only = 0的情况下。而当long_only不等于0的时候，就可以通过-server来表示long option了。
那么来看一个定义的例子
ACE_Get_Opt opt(argc, argv, "s:u:p:", 1, 0);
三个重载方法long_option的介绍
由于构造方法中的optstring只能制定short option，所以必须通过long_option方法来指定long option。
ACE_Get_Opt类中有三个重载的long_option方法，下面来一一介绍
long_option方法一：
/// Adds a long option with no corresponding short option. /** * If the @a name option is seen, @c operator() returns 0. * * @param name The long option to add. * @param has_arg Defines the argument requirements for * the new option. * * @retval 0 Success * @retval -1 The long option can not be added. */ int long_option (const ACE_TCHAR *name, OPTION_ARG_MODE has_arg = NO_ARG);
这个方法可以增加long_option，第一个参数name表示要添加的option的名字，第二个参数表示这个option是否包含参数，下面是OPTION_ARG_MODE的定义：
/// Mutually exclusive option argument mode used by long options. enum OPTION_ARG_MODE { /// Doesn't take an argument. NO_ARG = 0, /// Requires an argument, same as passing ":" after a short option /// character in @a optstring. ARG_REQUIRED = 1, /// Argument is optional, same as passing "::" after a short /// option character in @a optstring. ARG_OPTIONAL = 2 };
- NO_ARG表示没有参数，对应short option中的没有冒号的情况
- ARG_REQUIRED表示必须有参数，对应short option中有一个冒号的情况
- ARG_OPTIONAL表示参数可有可无，对应short option中有两个冒号的情况
此方法可以这样使用
if(opt.long_option("server", ACE_Get_Opt::ARG_REQUIRED) == -1) { cout << "server option can not be added" << endl; }
long_option方法二：
/// Adds a long option with a corresponding short option. /** * @param name The long option to add. * @param short_option A character, the short option that corresponds * to @a name. * @param has_arg Defines the argument requirements for * the new option. If the short option has already * been supplied in the @a optstring, @a has_arg * must match or an error is returned; otherwise, the * new short option is added to the @a optstring. * * @retval 0 Success * @retval -1 The long option can not be added. */ int long_option (const ACE_TCHAR *name, int short_option, OPTION_ARG_MODE has_arg = NO_ARG);
这个方法中增加了一个short_option参数，表示这个long_option与对应的short_option所表达的意义是一致的，当解析命令行时，遇到这个long_option时，就按此处的short option进行处理。
此方法可以这样使用
if(opt.long_option("server", 's', ACE_Get_Opt::ARG_REQUIRED) == -1) { cout << "server option can not be added" << endl; }
long_option方法三：
/// Returns the name of the long option found on the last call to /// @c operator() or 0 if none was found. const ACE_TCHAR *long_option (void) const;
这个方法表示解析命令行时，如果遇到long_option，则可以通过这个方法来获取是哪一个long_option。
命令行的解析
命令行的解析就要依赖 operator()()方法，和opt_arg方法，前者解析遇到的是哪个short option 或者 long option，而 opt_arg方法则获取option的参数。
下面看一下方法的声明
/** * Scan elements of @a argv (whose length is @a argc) for short option * characters given in @a optstring or long options (with no short * option equivalents). * * If an element of @a argv starts with '-', and is not exactly "-" * or "--", then it is a short option element. The characters of this * element (aside from the initial '-') are option characters. If * it starts with "--" followed by other characters it is treated as * a long option. If @c operator() is called repeatedly, it returns * each of the option characters from each of the option elements. * * @return The parsed option character. The following characters have * special significance. * @retval 0 A long option was found * @retval '\?' Either an unknown option character was found, or the * option is known but requires an argument, none was * specified, and @a optstring did not contain a leading * colon. * @retval ':' A known option character was found but it requires an * argument and none was supplied, and the first character * of @a optstring was a colon. @c opt_opt() indicates * which option was specified. * @retval '1' @c RETURN_IN_ORDER was specified and a non-option argument * was found. * @retval EOF No more option characters were found. @c opt_ind() will * return the index in @a argv of the first @a argv element * that is not an option. If @c PERMUTE_ARGS was * specified, the @a argv elements have been permuted so that * those that are not options now come last. * * @note The standards are unclear with respect to the conditions under * which '?' and ':' are returned, so we scan the initial characters of * @a optstring up unto the first short option character for '+', '-', * and ':' in order to determine ordering and missing argument behavior. */ int operator () (void);
/** * For communication from @c operator() to the caller. When * @c operator() finds an option that takes an argument, the argument * value is returned from this method, otherwise it returns 0. */ ACE_TCHAR *opt_arg (void) const;
operator()方法有几种返回值：
- 具体的option 字符，表示命令行中遇到了此 short option，或者关联了short option的 long option
- 0，表示遇到了一个long option，可以通过上面介绍的long_option的方法三来获取一下具体是哪一个long option
- EOF，也就是-1，表示解析结束。
当operator()方法返回一个short option或者 long option时，如果这个option可以有参数，则通过opt_arg方法来获取具体的参数。
完整的命令行解析代码
下面展示一个完整但是比较简单的命令行解析的示例程序代码
#include "ace/Get_Opt.h" #include <iostream> using namespace std; int _tmain(int argc, _TCHAR* argv[]) { //定义了三个short option, 类型为 ARG_REQUIRED ACE_Get_Opt opt(argc, argv, "s:u:p:", 1, 0); //定义三个long option, 并与相应的short option 进行关联 if(opt.long_option("server", 's', ACE_Get_Opt::ARG_REQUIRED) == -1) { cout << "server option can not be added" << endl; } if(opt.long_option("user", 'u', ACE_Get_Opt::ARG_REQUIRED) == -1) { cout << "user option can not be added" << endl; } if(opt.long_option("pwd", 'p', ACE_Get_Opt::ARG_REQUIRED) == -1) { cout << "pwd option can not be added" << endl; } //开始解析 for(int arg = 0; (arg = opt()) != -1;) { switch (arg) { case 's': cout << "server is " << opt.opt_arg() << endl; break; case 'u': cout << "user is " << opt.opt_arg() << endl; break; case 'p': cout << "pwd is " << opt.opt_arg() << endl; break; //解析 long option case 0: if(_stricmp(opt.long_option(), "server") == 0) { cout << "long option server is" << opt.opt_arg() << endl; } break; default: break; } } //如果直接运行，或者使用方式不对，则显示帮助 if(opt.opt_ind() + 2 < argc) { cout << "the usage is : GetOptTest.exe [option]" << endl; cout << "\t [-s] server address" << endl; cout << "\t [-u] user name" << endl; cout << "\t [-p] password of user" << endl << endl; } return 0; }
命令行解析实战
至此命令行解析的基本功能已经都介绍了，而且相应的代码已经完成，那么就让我们来欣赏一下解析的效果吧。
具体的命令行
-s 127.0.0.1 --server 192.168.0.1 -u hbccdf -p pwdfortest
运行效果图
其实功能很简单，ACE_Get_Opt的使用也很简单，但是一点点学习，并进行总结，是我一直需要提高的地方。以后会坚持把学到的东西进行总结，并分享到博客上。
坚持！！！
