A programming puzzle kept me awake one night: If a structure allows for any variable type to be a member, and a function is a valid variable type, why not have a structure with a function as one of its members? Am I nuts?
\n
\nYes, I’m nuts.<\/p>\n
Referring back to a Lesson from 2015<\/a>, functions are variables. All you must do is declare them properly — Never mind the why, though I can think of some intriguing reasons, but here is such a structure definition:<\/p>\n Structure f<\/em> contains two functions as members, f1<\/em> and f2<\/em>. Like any structure definition, these members are placeholders. The next step is to declare a struct f<\/em> variable and then assign functions to the f1<\/em> and f2<\/em> members. The following code demonstrates:<\/p>\n The silly functions left()<\/em> and right()<\/em> output strings. These functions are assigned to the members f1<\/em> and f2<\/em> of struct f<\/em> variable At Lines 24 and 25, the functions are called — indirectly via the structure members:<\/p>\n And it works! Here’s the output:<\/p>\n Keeping in mind that a structure is yet another type of C language variable (I refer to is as a “multi-variable” in my books), the potential of mixing data with functions is nifty. In a way, it opens a portal through which you could create almost object-oriented like features in the C language: By attaching functions to a data type, you can access them similar to the way “methods” are accessed in object-oriented languages. Of course, a lot of overhead would be involved, but the potential exists.<\/p>\n","protected":false},"excerpt":{"rendered":" A structure can hold any valid variable type as a member, so why not have a function as a structure member? Continue reading type<\/em> (*name<\/em>)(arguments<\/em>)<\/code> — and you can have a function as a member in a structure.<\/p>\n\tstruct f {
\n void (*f1)();
\n void (*f2)();
\n};<\/code><\/p>\n\r\n#include <stdio.h>\r\n\r\nvoid left(void)\r\n{\r\n puts(\"left\");\r\n}\r\n\r\nvoid right(void)\r\n{\r\n puts(\"right\");\r\n}\r\n\r\nint main()\r\n{\r\n struct f {\r\n void (*f1)();\r\n void (*f2)();\r\n };\r\n struct f func;\r\n\r\n func.f1 = &left;\r\n func.f2 = &right;\r\n\r\n func.f1();\r\n func.f2();\r\n\r\n return(0);\r\n}<\/pre>\nfunc<\/code> at Lines 21 and 22. Only the functions’ names are used, prefixed by the address-of (& ampersand) operator. Remember that functions are referenced internally by their addresses.<\/p>\nfunc.f1();
\nfunc.f2();<\/code><\/p>\nleft
\nright<\/code><\/p>\n