Overloading Macro on Number of Arguments
Asked Answered
S

11

250

I have two macros FOO2 and FOO3:

#define FOO2(x,y) ...
#define FOO3(x,y,z) ...

I want to define a new macro FOO as follows:

#define FOO(x,y) FOO2(x,y)
#define FOO(x,y,z) FOO3(x,y,z)

But this doesn't work because macros do not overload on number of arguments.

Without modifying FOO2 and FOO3, is there some way to define a macro FOO (using __VA_ARGS__ or otherwise) to get the same effect of dispatching FOO(x,y) to FOO2, and FOO(x,y,z) to FOO3?

Stephanistephania answered 1/8, 2012 at 14:42 Comment(6)
I have a very strong feeling that this has been asked several times before... [update] e.g. here.Faradize
@KerrekSB: That may be related, must it is most certainly not a dupe.Stephanistephania
No, maybe not that one, but something like this comes up about once a month...Faradize
Same for C++: #3047389 Should be the same since the preprocessors are basically the same: #5086033Hamm
Related: #11317974Rank
This one too: #2124839Rank
C
372

Simple as:

#define GET_MACRO(_1,_2,_3,NAME,...) NAME
#define FOO(...) GET_MACRO(__VA_ARGS__, FOO3, FOO2)(__VA_ARGS__)

So if you have these macros, they expand as described:

FOO(World, !)         // expands to FOO2(World, !)
FOO(foo,bar,baz)      // expands to FOO3(foo,bar,baz)

If you want a fourth one:

#define GET_MACRO(_1,_2,_3,_4,NAME,...) NAME
#define FOO(...) GET_MACRO(__VA_ARGS__, FOO4, FOO3, FOO2)(__VA_ARGS__)

FOO(a,b,c,d)          // expands to FOO4(a,b,c,d)

Naturally, if you define FOO2, FOO3 and FOO4, the output will be replaced by those of the defined macros.

Ciro answered 1/8, 2012 at 16:8 Comment(13)
Is this possible, when there is an additional FOO0 that takes 0 zero arguments?Tittletattle
No. It has to have at least one argument.Ciro
@Tittletattle Adding a 0-argument macro to the list is possible, see my answer.Grillparzer
Does not work on Microsoft Visual Studio 2010, VA_ARGS seems to be expanded into a single macro argument.Rajasthan
Found this answer to make it work under MSVC 2010.Rajasthan
If anybody's confused as to how to use the EXPAND mentioned in @Étienne's link, you basically invoke it on GET_MACRO like so #define FOO(...) EXPAND(GET_MACRO(__VA_ARGS__, FOO3, FOO2, FOO1)(__VA_ARGS__)) and it should expand to the right number of arguments in msvc.Seppuku
What if you are using VAR_ARGS in your base macros? e.g. #define GET_MACRO(1,_2,NAME,...) NAME #define FOO_V(format, ...) printf(format, _VA_ARGS) #define FOO_1(format) printf(format) #define FOO(...) GET_MACRO(VA_ARGS, FOO_V, FOO_1)(VA_ARGS) Works: FOO("bar"); FOO("bar %d", 42); Doesn't Work: FOO("bar %d %d", 42, 69); GCC Error: Called object is not a function or function pointer FOO("boo %d %d", 42, 69);Aspen
Note that on C++11, you'll get a warning: ISO C++11 requires at least one argument for the "..." in a variadic macro. To fix this, add an unused argument (or even just a comma) after the last param in the definition of FOO(...): #define FOO(...) GET_MACRO(__VA_ARGS__, FOO3, FOO2, UNUSED)(__VA_ARGS__) (See it run on Coliru).Fireside
@Ciro can we use same to overload strcpy_s using macros. I want to replace current 'strcpy' occurrences with 'strcpy_s'. Don't get me wrong i am just curious about it. Is it possible practically? Actially 'strcpy' takes two parameters and 'strcpy_s' takes three parameters. I was thinking about defining macro:" #define strcpy(x,y) strcpy_s(x,sizeof(x),y) ". Ultimately its gona work for static array and not for dynamic array. so i was thinking about introducing another macro as : " #define strcpy(x,y,z) strcpy_s(x,y,z) "Decalcomania
is it also possible to achieve default argument in this?Elephant
I also run into problems with MSVC and __VA_ARGS__. I believe this may be the solution that @Étienne alluded to: #define GET_MACRO(_1, _2, _3, NAME, ...) NAME ...NEW:... #define EXPAND(x) x ...and FINALLY:... #define FOO(...) EXPAND(GET_MACRO(__VA_ARGS__, FOO3, FOO2, FOO1))EXPAND((__VA_ARGS__))Hopi
Nice, although it's hard to understand unless you are familiar with C macros. This answer could be improved by 1-2 sentences elaborating how this works.Neurotomy
Those wondering how it works, GET_MACRO is sort of "sliding-window" to select name and is defined such that it always expands to 4th argument passed to it. So in example above, it must always be called with at least 4 arguments and thus won't work for calls like FOO0 or FOO1.Shoer
G
74

To add on to netcoder's answer, you CAN in fact do this with a 0-argument macro, with the help of the GCC ##__VA_ARGS__ extension:

#define GET_MACRO(_0, _1, _2, NAME, ...) NAME
#define FOO(...) GET_MACRO(_0, ##__VA_ARGS__, FOO2, FOO1, FOO0)(__VA_ARGS__)
Grillparzer answered 27/1, 2014 at 0:58 Comment(6)
is it possible to allow FOO1 and FOO2 but not FOO0 without doing #define FOO0 _Pragma("error FOO0 not allowed")?Hornswoggle
FOO0 not working in qt + mingw32, call FOO0 will invoke the FOO1Shoon
Very promising and simple. But doesn't work for FOO0 with -std=c++11...:-(Sorority
@Sorority That's because ##__VA_ARGS__ is a GNU feature which differs from the C++11 standard. It works with -std=gnu++11.Grillparzer
Same problem if you're doing this in C and you try to use -std=c99 or -std=c11. You need to use -std=gnu99 or -std=gnu11 insteadEnclosure
It appears that replacing _0, ##__VA_ARGS__ with _0 __VA_OPT__(,) __VA_ARGS__ is the new way to do this.Brody
N
56

Here is a more general solution:

// get number of arguments with __NARG__
#define __NARG__(...)  __NARG_I_(__VA_ARGS__,__RSEQ_N())
#define __NARG_I_(...) __ARG_N(__VA_ARGS__)
#define __ARG_N( \
      _1, _2, _3, _4, _5, _6, _7, _8, _9,_10, \
     _11,_12,_13,_14,_15,_16,_17,_18,_19,_20, \
     _21,_22,_23,_24,_25,_26,_27,_28,_29,_30, \
     _31,_32,_33,_34,_35,_36,_37,_38,_39,_40, \
     _41,_42,_43,_44,_45,_46,_47,_48,_49,_50, \
     _51,_52,_53,_54,_55,_56,_57,_58,_59,_60, \
     _61,_62,_63,N,...) N
#define __RSEQ_N() \
     63,62,61,60,                   \
     59,58,57,56,55,54,53,52,51,50, \
     49,48,47,46,45,44,43,42,41,40, \
     39,38,37,36,35,34,33,32,31,30, \
     29,28,27,26,25,24,23,22,21,20, \
     19,18,17,16,15,14,13,12,11,10, \
     9,8,7,6,5,4,3,2,1,0

// general definition for any function name
#define _VFUNC_(name, n) name##n
#define _VFUNC(name, n) _VFUNC_(name, n)
#define VFUNC(func, ...) _VFUNC(func, __NARG__(__VA_ARGS__)) (__VA_ARGS__)

// definition for FOO
#define FOO(...) VFUNC(FOO, __VA_ARGS__)

Define your functions:

#define FOO2(x, y) ((x) + (y))
#define FOO3(x, y, z) ((x) + (y) + (z))

// it also works with C functions:
int FOO4(int a, int b, int c, int d) { return a + b + c + d; }

Now you can use FOO with 2, 3 and 4 arguments:

FOO(42, 42) // will use makro function FOO2
FOO(42, 42, 42) // will use makro function FOO3
FOO(42, 42, 42, 42) // will call FOO4 function

Limitations

  • Only up to 63 arguments (but expandable)
  • Function for no argument only in GCC possible

Ideas

Use it for default arguments:

#define func(...) VFUNC(func, __VA_ARGS__)
#define func2(a, b) func4(a, b, NULL, NULL)
#define func3(a, b, c) func4(a, b, c, NULL)

// real function:
int func4(int a, int b, void* c, void* d) { /* ... */ }

Use it for functions with possible infinite number of arguments:

#define SUM(...) VFUNC(SUM, __VA_ARGS__)
#define SUM2(a, b) ((a) + (b))
#define SUM3(a, b, c) ((a) + (b) + (c))
#define SUM4(a, b, c) ((a) + (b) + (c) + (d))
// ...

PS: __NARG__ is copied from Laurent Deniau & Roland Illig here: https://groups.google.com/group/comp.std.c/browse_thread/thread/77ee8c8f92e4a3fb/346fc464319b1ee5?pli=1

Nordstrom answered 16/10, 2014 at 15:25 Comment(6)
Related: #11317974Rank
This one too: #2124839Rank
The macro __NARG_I_ seems completely unnecessary and superfluous. It just adds an extra step, and confusion. I recommend deleting it entirely and just defining __NARG__ instead as: #define __NARG__(...) __ARG_N(__VA_ARGS__,__RSEQ_N()).Rank
Or will that somehow break pre-processing? Am I missing something?Rank
Same with _VFUNC_: just delete it. Then, define _VFUNC as: #define _VFUNC(name, n) name##n instead of #define _VFUNC(name, n) _VFUNC_(name, n).Rank
@GabrielStaples This would break macro expansion of the arguments. See for example this answer where a extra _NX function is used: https://mcmap.net/q/67632/-c-c-macro-string-concatenationNordstrom
N
22

I was just researching this myself, and I came across this here. The author added default argument support for C functions via macros.

I'll try to briefly summarize the article. Basically, you need to define a macro that can count arguments. This macro will return 2, 1, 0, or whatever range of arguments it can support. Eg:

#define _ARG2(_0, _1, _2, ...) _2
#define NARG2(...) _ARG2(__VA_ARGS__, 2, 1, 0)

With this, you need to create another macro that takes a variable number of arguments, counts the arguments, and calls the appropriate macro. I've taken your example macro and combined it with the article's example. I have FOO1 call function a() and FOO2 call function a with argument b (obviously, I'm assuming C++ here, but you can change the macro to whatever).

#define FOO1(a) a();
#define FOO2(a,b) a(b);

#define _ARG2(_0, _1, _2, ...) _2
#define NARG2(...) _ARG2(__VA_ARGS__, 2, 1, 0)

#define _ONE_OR_TWO_ARGS_1(a) FOO1(a)
#define _ONE_OR_TWO_ARGS_2(a, b) FOO2(a,b)

#define __ONE_OR_TWO_ARGS(N, ...) _ONE_OR_TWO_ARGS_ ## N (__VA_ARGS__)
#define _ONE_OR_TWO_ARGS(N, ...) __ONE_OR_TWO_ARGS(N, __VA_ARGS__)

#define FOO(...) _ONE_OR_TWO_ARGS(NARG2(__VA_ARGS__), __VA_ARGS__)

So if you have

FOO(a)
FOO(a,b)

The preprocessor expands that to

a();
a(b);

I would definitely read the article that I linked. It's very informative and he mentions that NARG2 won't work on empty arguments. He follows this up here.

Nose answered 1/8, 2012 at 16:4 Comment(0)
P
10

Here is a more compact version of the answer above. With example.

#include <iostream>
using namespace std;

#define OVERLOADED_MACRO(M, ...) _OVR(M, _COUNT_ARGS(__VA_ARGS__)) (__VA_ARGS__)
#define _OVR(macroName, number_of_args)   _OVR_EXPAND(macroName, number_of_args)
#define _OVR_EXPAND(macroName, number_of_args)    macroName##number_of_args

#define _COUNT_ARGS(...)  _ARG_PATTERN_MATCH(__VA_ARGS__, 9,8,7,6,5,4,3,2,1)
#define _ARG_PATTERN_MATCH(_1,_2,_3,_4,_5,_6,_7,_8,_9, N, ...)   N


//Example:
#define ff(...)     OVERLOADED_MACRO(ff, __VA_ARGS__)
#define ii(...)     OVERLOADED_MACRO(ii, __VA_ARGS__)

#define ff3(c, a, b) for (int c = int(a); c < int(b); ++c)
#define ff2(c, b)   ff3(c, 0, b)

#define ii2(a, b)   ff3(i, a, b)
#define ii1(n)      ii2(0, n)


int main() {
    ff (counter, 3, 5)
        cout << "counter = " << counter << endl;
    ff (abc, 4)
        cout << "abc = " << abc << endl;
    ii (3)
        cout << "i = " << i << endl;
    ii (100, 103)
        cout << "i = " << i << endl;


    return 0;
}

Run:

User@Table 13:06:16 /c/T
$ g++ test_overloaded_macros.cpp 

User@Table 13:16:26 /c/T
$ ./a.exe
counter = 3
counter = 4
abc = 0
abc = 1
abc = 2
abc = 3
i = 0
i = 1
i = 2
i = 100
i = 101
i = 102

Note that having both _OVR and _OVR_EXPAND may look redundant, but it's necessary for the preprocessor to expand the _COUNT_ARGS(__VA_ARGS__) part, which otherwise is treated as a string.

Pratfall answered 1/6, 2015 at 5:23 Comment(1)
I like this solution. Can it be modified to handle an overloaded macro that takes zero arguments?Kinnikinnick
A
5

Here's a spin off from Evgeni Sergeev's answer. This one supports zero argument overloads as well!

I tested this with GCC and MinGW. It ought to work with old and new versions of C++. Note that I wouldn't guarantee it for MSVC... But with some tweaking, I'm confident it could be made to work with that too.

I also formatted this to be pasted into a header file (which I called macroutil.h). If you do that, you can just include this header whatever you need the feature, and not look at the nastiness involved in the implementation.

#ifndef MACROUTIL_H
#define MACROUTIL_H

//-----------------------------------------------------------------------------
// OVERLOADED_MACRO
//
// used to create other macros with overloaded argument lists
//
// Example Use:
// #define myMacro(...) OVERLOADED_MACRO( myMacro, __VA_ARGS__ )
// #define myMacro0() someFunc()
// #define myMacro1( arg1 ) someFunc( arg1 )
// #define myMacro2( arg1, arg2 ) someFunc( arg1, arg2 )
//
// myMacro();
// myMacro(1);
// myMacro(1,2);
//
// Note the numerical suffix on the macro names,
// which indicates the number of arguments.
// That is the REQUIRED naming convention for your macros.
//
//-----------------------------------------------------------------------------

// OVERLOADED_MACRO
// derived from: https://mcmap.net/q/67473/-overloading-macro-on-number-of-arguments
// replaced use of _COUNT_ARGS macro with VA_NUM_ARGS defined below
// to support of zero argument overloads
#define OVERLOADED_MACRO(M, ...) _OVR(M, VA_NUM_ARGS(__VA_ARGS__)) (__VA_ARGS__)
#define _OVR(macroName, number_of_args)   _OVR_EXPAND(macroName, number_of_args)
#define _OVR_EXPAND(macroName, number_of_args)    macroName##number_of_args
//#define _COUNT_ARGS(...)  _ARG_PATTERN_MATCH(__VA_ARGS__, 15,14,13,12,11,10,9,8,7,6,5,4,3,2,1)
#define _ARG_PATTERN_MATCH(_1,_2,_3,_4,_5,_6,_7,_8,_9,_10,_11,_12,_13,_14,_15, N, ...)   N

// VA_NUM_ARGS
// copied from comments section of:
// http://efesx.com/2010/07/17/variadic-macro-to-count-number-of-arguments/
// which itself was derived from:
// https://gustedt.wordpress.com/2010/06/08/detect-empty-macro-arguments/
#define _ARG16(_0, _1, _2, _3, _4, _5, _6, _7, _8, _9, _10, _11, _12, _13, _14, _15, ...) _15
#define HAS_COMMA(...) _ARG16(__VA_ARGS__, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0)
#define HAS_NO_COMMA(...) _ARG16(__VA_ARGS__, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1)
#define _TRIGGER_PARENTHESIS_(...) ,

#define HAS_ZERO_OR_ONE_ARGS(...) \
    _HAS_ZERO_OR_ONE_ARGS( \
    /* test if there is just one argument, eventually an empty one */ \
    HAS_COMMA(__VA_ARGS__), \
    /* test if _TRIGGER_PARENTHESIS_ together with the argument adds a comma */ \
    HAS_COMMA(_TRIGGER_PARENTHESIS_ __VA_ARGS__), \
    /* test if the argument together with a parenthesis adds a comma */ \
    HAS_COMMA(__VA_ARGS__ (~)), \
    /* test if placing it between _TRIGGER_PARENTHESIS_ and the parenthesis adds a comma */ \
    HAS_COMMA(_TRIGGER_PARENTHESIS_ __VA_ARGS__ (~)) \
    )

#define PASTE5(_0, _1, _2, _3, _4) _0 ## _1 ## _2 ## _3 ## _4
#define _HAS_ZERO_OR_ONE_ARGS(_0, _1, _2, _3) HAS_NO_COMMA(PASTE5(_IS_EMPTY_CASE_, _0, _1, _2, _3))
#define _IS_EMPTY_CASE_0001 ,

#define _VA0(...) HAS_ZERO_OR_ONE_ARGS(__VA_ARGS__)
#define _VA1(...) HAS_ZERO_OR_ONE_ARGS(__VA_ARGS__)
#define _VA2(...) 2
#define _VA3(...) 3
#define _VA4(...) 4
#define _VA5(...) 5
#define _VA6(...) 6
#define _VA7(...) 7
#define _VA8(...) 8
#define _VA9(...) 9
#define _VA10(...) 10
#define _VA11(...) 11
#define _VA12(...) 12
#define _VA13(...) 13
#define _VA14(...) 14
#define _VA15(...) 15
#define _VA16(...) 16

#define VA_NUM_ARGS(...) VA_NUM_ARGS_IMPL(__VA_ARGS__, PP_RSEQ_N(__VA_ARGS__) )
#define VA_NUM_ARGS_IMPL(...) VA_NUM_ARGS_N(__VA_ARGS__)

#define VA_NUM_ARGS_N( \
    _1, _2, _3, _4, _5, _6, _7, _8, _9,_10, \
    _11,_12,_13,_14,_15,_16,N,...) N

#define PP_RSEQ_N(...) \
    _VA16(__VA_ARGS__),_VA15(__VA_ARGS__),_VA14(__VA_ARGS__),_VA13(__VA_ARGS__), \
    _VA12(__VA_ARGS__),_VA11(__VA_ARGS__),_VA10(__VA_ARGS__), _VA9(__VA_ARGS__), \
    _VA8(__VA_ARGS__),_VA7(__VA_ARGS__),_VA6(__VA_ARGS__),_VA5(__VA_ARGS__), \
    _VA4(__VA_ARGS__),_VA3(__VA_ARGS__),_VA2(__VA_ARGS__),_VA1(__VA_ARGS__), \
    _VA0(__VA_ARGS__)

//-----------------------------------------------------------------------------

#endif // MACROUTIL_H
Amoebaean answered 22/11, 2017 at 1:3 Comment(0)
T
4

This seems to work fine on GCC, Clang and MSVC. It's a cleaned up version of some of the answers here

#define _my_BUGFX(x) x

#define _my_NARG2(...) _my_BUGFX(_my_NARG1(__VA_ARGS__,_my_RSEQN()))
#define _my_NARG1(...) _my_BUGFX(_my_ARGSN(__VA_ARGS__))
#define _my_ARGSN(_1,_2,_3,_4,_5,_6,_7,_8,_9,_10,N,...) N
#define _my_RSEQN() 10,9,8,7,6,5,4,3,2,1,0

#define _my_FUNC2(name,n) name ## n
#define _my_FUNC1(name,n) _my_FUNC2(name,n)
#define GET_MACRO(func,...) _my_FUNC1(func,_my_BUGFX(_my_NARG2(__VA_ARGS__))) (__VA_ARGS__)

#define FOO(...) GET_MACRO(FOO,__VA_ARGS__)
Twopence answered 9/8, 2017 at 20:59 Comment(1)
@RianQuinn How to adjust this macro so that it works with zero argument #define func0() foo? The current version does not handle this case unfortunately.Indifferentism
C
4

Based on @netcoder's answer and @vexe's suggestion about the Visual Studio compiler support, I found this code working pretty well across various platforms:

#define FOO1(a) func1(a)
#define FOO2(a, b) func2(a, b)
#define FOO3(a, b, c) func3(a, b, c)

#define EXPAND(x) x
#define GET_MACRO(_1, _2, _3, NAME, ...) NAME
#define FOO(...) EXPAND(GET_MACRO(__VA_ARGS__, FOO3, FOO2, FOO1)(__VA_ARGS__))

, where func1(), func2(), func3() are just normal functions that accept different number of parameters.

Campy answered 12/11, 2021 at 15:24 Comment(0)
S
3

Maybe you can use this macro to count the number of arguments.

#define VA_NUM_ARGS(...) VA_NUM_ARGS_IMPL(__VA_ARGS__, 5,4,3,2,1)
#define VA_NUM_ARGS_IMPL(_1,_2,_3,_4,_5,N,...) N
Sirius answered 1/8, 2012 at 15:55 Comment(0)
I
0

Based on R1tschY's reply, here's what I did to implement variable arguments and default arguments, using the Deniau-Illig construction.

// ----------------------------------------------------------------
// library
#define __nargs100__(a00,a01,a02,a03,a04,a05,a06,a07,a08,a09,a0a,a0b,a0c,a0d,a0e,a0f,a10,a11,a12,a13,a14,a15,a16,a17,a18,a19,a1a,a1b,a1c,a1d,a1e,a1f,a20,a21,a22,a23,a24,a25,a26,a27,a28,a29,a2a,a2b,a2c,a2d,a2e,a2f,a30,a31,a32,a33,a34,a35,a36,a37,a38,a39,a3a,a3b,a3c,a3d,a3e,a3f,a40,a41,a42,a43,a44,a45,a46,a47,a48,a49,a4a,a4b,a4c,a4d,a4e,a4f,a50,a51,a52,a53,a54,a55,a56,a57,a58,a59,a5a,a5b,a5c,a5d,a5e,a5f,a60,a61,a62,a63,a64,a65,a66,a67,a68,a69,a6a,a6b,a6c,a6d,a6e,a6f,a70,a71,a72,a73,a74,a75,a76,a77,a78,a79,a7a,a7b,a7c,a7d,a7e,a7f,a80,a81,a82,a83,a84,a85,a86,a87,a88,a89,a8a,a8b,a8c,a8d,a8e,a8f,a90,a91,a92,a93,a94,a95,a96,a97,a98,a99,a9a,a9b,a9c,a9d,a9e,a9f,aa0,aa1,aa2,aa3,aa4,aa5,aa6,aa7,aa8,aa9,aaa,aab,aac,aad,aae,aaf,ab0,ab1,ab2,ab3,ab4,ab5,ab6,ab7,ab8,ab9,aba,abb,abc,abd,abe,abf,ac0,ac1,ac2,ac3,ac4,ac5,ac6,ac7,ac8,ac9,aca,acb,acc,acd,ace,acf,ad0,ad1,ad2,ad3,ad4,ad5,ad6,ad7,ad8,ad9,ada,adb,adc,add,ade,adf,ae0,ae1,ae2,ae3,ae4,ae5,ae6,ae7,ae8,ae9,aea,aeb,aec,aed,aee,aef,af0,af1,af2,af3,af4,af5,af6,af7,af8,af9,afa,afb,afc,afd,afe,aff,a100,...)  a100
#define __nargs__(...)   __nargs100__(,##__VA_ARGS__, ff,fe,fd,fc,fb,fa,f9,f8,f7,f6,f5,f4,f3,f2,f1,f0,ef,ee,ed,ec,eb,ea,e9,e8,e7,e6,e5,e4,e3,e2,e1,e0,df,de,dd,dc,db,da,d9,d8,d7,d6,d5,d4,d3,d2,d1,d0,cf,ce,cd,cc,cb,ca,c9,c8,c7,c6,c5,c4,c3,c2,c1,c0,bf,be,bd,bc,bb,ba,b9,b8,b7,b6,b5,b4,b3,b2,b1,b0,af,ae,ad,ac,ab,aa,a9,a8,a7,a6,a5,a4,a3,a2,a1,a0,9f,9e,9d,9c,9b,9a,99,98,97,96,95,94,93,92,91,90,8f,8e,8d,8c,8b,8a,89,88,87,86,85,84,83,82,81,80,7f,7e,7d,7c,7b,7a,79,78,77,76,75,74,73,72,71,70,6f,6e,6d,6c,6b,6a,69,68,67,66,65,64,63,62,61,60,5f,5e,5d,5c,5b,5a,59,58,57,56,55,54,53,52,51,50,4f,4e,4d,4c,4b,4a,49,48,47,46,45,44,43,42,41,40,3f,3e,3d,3c,3b,3a,39,38,37,36,35,34,33,32,31,30,2f,2e,2d,2c,2b,2a,29,28,27,26,25,24,23,22,21,20,1f,1e,1d,1c,1b,1a,19,18,17,16,15,14,13,12,11,10,f,e,d,c,b,a,9,8,7,6,5,4,3,2,1,0)
#define __vfn(name, n)  name##n
#define _vfn( name, n)  __vfn(name, n)
#define vfn(  fn, ...)  _vfn(fn, __nargs__(__VA_ARGS__))(__VA_ARGS__)

// ----------------------------------------------------------------
// example

// backend: actual implementation, 2 mandatory args, 3 optional args
#define _bshow(bdim,data, ncols,nbits,base)({  \
  /* do stuff here */  \
})

// "frontend", default arguments get implemented here. the suffix is the number of arguments, in hexadecimal base
#define bshow2(...)  _bshow(__VA_ARGS__, 16,32,16)
#define bshow3(...)  _bshow(__VA_ARGS__, 32,16)
#define bshow4(...)  _bshow(__VA_ARGS__, 16)
#define bshow5(...)  _bshow(__VA_ARGS__)
#define bshow(...)  vfn(bshow,__VA_ARGS__)

// test
bshow(0x100,data0);
bshow(0x100,data0, 14);
bshow(0x100,data0, 12,16);
bshow(0x100,data0, 10, 8,2);
Impartial answered 7/6, 2022 at 15:15 Comment(0)
P
0

I came up with a generic solution by myself. Though it is a little verbose, it supports default overload, and provides a template to create any overloading macro as long as it needs.

#define GET_OVERLOAD_RESULT_(                                                  \
    _0, _1, _2, _3, _4, _5, _6, _7, _8, _9, _10, _11, _12, _13, _14, _15, _16, \
    _17, _18, _19, _20, _21, _22, _23, _24, _25, _26, _27, _28, _29, _30, _31, \
    _32, _33, _34, _35, _36, _37, _38, _39, _40, _41, _42, _43, _44, _45, _46, \
    _47, _48, _49, _50, _51, _52, _53, _54, _55, _56, _57, _58, _59, _60, _61, \
    _62, _63, _64, _65, _66, _67, _68, _69, _70, _71, _72, _73, _74, _75, _76, \
    _77, _78, _79, _80, _81, _82, _83, _84, _85, _86, _87, _88, _89, _90, _91, \
    _92, _93, _94, _95, _96, _97, _98, _99, _100, result, ...)                 \
  result

#define OVERLOADER_TAKES_AT_MOST_100_ARGS_(func, ...)                          \
  GET_OVERLOAD_RESULT_(                                                        \
      __VA_ARGS__, func(100), func(99), func(98), func(97), func(96),          \
      func(95), func(94), func(93), func(92), func(91), func(90), func(89),    \
      func(88), func(87), func(86), func(85), func(84), func(83), func(82),    \
      func(81), func(80), func(79), func(78), func(77), func(76), func(75),    \
      func(74), func(73), func(72), func(71), func(70), func(69), func(68),    \
      func(67), func(66), func(65), func(64), func(63), func(62), func(61),    \
      func(60), func(59), func(58), func(57), func(56), func(55), func(54),    \
      func(53), func(52), func(51), func(50), func(49), func(48), func(47),    \
      func(46), func(45), func(44), func(43), func(42), func(41), func(40),    \
      func(39), func(38), func(37), func(36), func(35), func(34), func(33),    \
      func(32), func(31), func(30), func(29), func(28), func(27), func(26),    \
      func(25), func(24), func(23), func(22), func(21), func(20), func(19),    \
      func(18), func(17), func(16), func(15), func(14), func(13), func(12),    \
      func(11), func(10), func(9), func(8), func(7), func(6), func(5),         \
      func(4), func(3), func(2), func(1), func(0))

#define GET_ARGS_COUNT_(n) n

#define ARGS_COUNT(...)                                                        \
  OVERLOADER_TAKES_AT_MOST_100_ARGS_(GET_ARGS_COUNT_, _, ##__VA_ARGS__)

#define CONCAT_TOKEN(a, ...) a##__VA_ARGS__

#define CHOOSE_OVERLOAD_(_, overload, ...) overload
#define DISPATCH_OVERLOAD_INTERNAL_(default_overload, ...)                     \
  CHOOSE_OVERLOAD_(__VA_ARGS__, default_overload)

#define DISPATCH_OVERLOAD(overload_func, n)                                    \
  DISPATCH_OVERLOAD_INTERNAL_(overload_func##_default,                         \
                              CONCAT_TOKEN(overload_func##_, n))

#define MAKE_OVERLOAD(func) _, func

#define DEFINE_OVERLOAD_MACRO(overloader, ...)                                 \
  OVERLOADER_TAKES_AT_MOST_100_ARGS_(overloader, _, ##__VA_ARGS__)(__VA_ARGS__)

ARGS_COUNT is used to determine the number of the arguments it takes. And there are several macro for creating a custom overload. For example:

template <typename... Args>
std::string default_printer(Args&&... args) {
  return "default";
}

#define PRINT_1_arg(a) ("print one argument. " #a ": " + std::to_string(a))
#define PRINT_2_args(a, b) \
  ("print two arguments. " #a ": " + std::to_string(a) + ", " #b ": " + std::to_string(b))

// define the overload of one argument
#define PRINT_IMPL_1 MAKE_OVERLOAD(PRINT_1_arg)
// define the overload of two arguments
#define PRINT_IMPL_2 MAKE_OVERLOAD(PRINT_2_args)
// define the default overload
#define PRINT_IMPL_default(...) default_printer(__VA_ARGS__)
// define a helper (I hope it can be ommited, yet I didn't find a way so far)
#define PRINT_OVERLOADER(n) DISPATCH_OVERLOAD(PRINT_IMPL, n)

// the custom overload macro is named `PRINT`
#define PRINT(...) DEFINE_OVERLOAD_MACRO(PRINT_OVERLOADER, ##__VA_ARGS__)


int main() {
  int a = 0;
  double b = 1.5;
  std::string c = "234";
  static_assert(ARGS_COUNT() == 0);
  static_assert(ARGS_COUNT(a) == 1);
  static_assert(ARGS_COUNT(a, b) == 2);
  static_assert(ARGS_COUNT(a, b, c) == 3);
  // the output of following codes is:
  // default
  // print one argument. a: 0
  // print two arguments. a: 0, b: 1.500000
  // default
  std::cout << PRINT() << '\n';
  std::cout << PRINT(a) << '\n';
  std::cout << PRINT(a, b) << '\n';
  std::cout << PRINT(a, b, c) << '\n';
}

Platino answered 15/4 at 18:58 Comment(0)

© 2022 - 2024 — McMap. All rights reserved.