comp.std.c-__VA_NARG__.txt

Path: g2news1.google.com!news1.google.com!news3.google.com!news.glorb.com!news2.euro.net!newsfeed.vmunix.org!newsfeed.cw.net!cw.net!news-FFM2.ecrc.de!news0.de.colt.net!news-fra1.dfn.de!kanaga.switch.ch!ezmp3.switch.ch!news-zh.switch.ch!switch.ch!cernne03.cern.ch!cern.ch!news
From: Laurent Deniau <laurent...@cern.ch>
Newsgroups: comp.std.c
Subject: __VA_NARG__
Date: Mon, 16 Jan 2006 18:43:40 +0100
Organization: CERN - European Laboratory for Particle Physics
Lines: 39
Message-ID: <dqgm2f$ije$1@sunnews.cern.ch>
NNTP-Posting-Host: pb-d-128-141-34-253.cern.ch
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Trace: sunnews.cern.ch 1137433487 19054 (None) 128.141.34.253
X-Complaints-To: ne...@sunnews.cern.ch
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.12) Gecko/20051007 Debian/1.7.12-1
X-Accept-Language: en
Xref: g2news1.google.com comp.std.c:1121

Hi all,

A year ago, I was asking here for an equivalent of __VA_NARG__ which 
would return the number of arguments contained in __VA_ARGS__ before its 
expansion. In fact my problem at that time (detecting for a third 
argument) was solved by the solution of P. Mensonides. But I was still 
thinking that the standard should have provided such a facilities rather 
easy to compute for cpp.

This morning I had to face again the same problem, that is knowing the 
number of arguments contained in __VA_ARGS__ before its expansion (after 
its expansion can always be achieved if you can do it before). I found a 
simple non-iterative solution which may be of interest here as an answer 
to who will ask in the future for a kind of __VA_NARG__ in the standard 
and I post it for archiving. May be some more elegant-efficient solution 
exists?

regards,

ld.


Returns NARG, the number of arguments contained in __VA_ARGS__ before 
expansion as far as NARG is >0 and <64 (cpp limits):

#define PP_NARG( ...) PP_NARG_(__VA_ARGS__,PP_RSEQ_N())
#define PP_NARG_(...) PP_ARG_N(__VA_ARGS__)
#define PP_ARG_N(_1,_2,_3,_4,_5,_6,_7,_8,_9,[..],_61,_62,_63,N,...) N
#define PP_RSEQ_N() 63,62,61,60,[..],9,8,7,6,5,4,3,2,1,0

[..] stands for the continuation of the sequence omitted here for 
lisibility.

PP_NARG(A) -> 1
PP_NARG(A,B) -> 2
PP_NARG(A,B,C) -> 3
PP_NARG(A,B,C,D) -> 4
PP_NARG(A,B,C,D,E) -> 5
PP_NARG(A1,A2,[..],A62,A63) -> 63


Path: g2news1.google.com!news1.google.com!proxad.net!news.tele.dk!news.tele.dk!small.news.tele.dk!news-fra1.dfn.de!news-ham1.dfn.de!news.uni-hamburg.de!not-for-mail
From: Roland Illig <roland...@gmx.de>
Newsgroups: comp.std.c
Subject: Re: __VA_NARG__
Date: Fri, 20 Jan 2006 12:58:41 +0100
Organization: University of Hamburg -- Germany
Lines: 53
Message-ID: <dqqj9j$31l$1@rzsun03.rrz.uni-hamburg.de>
References: <dqgm2f$ije$1@sunnews.cern.ch>
NNTP-Posting-Host: rzdspc3.informatik.uni-hamburg.de
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Trace: rzsun03.rrz.uni-hamburg.de 1137758323 3125 134.100.9.63 (20 Jan 2006 11:58:43 GMT)
X-Complaints-To: use...@news.uni-hamburg.de
NNTP-Posting-Date: 20 Jan 2006 11:58:43 GMT
User-Agent: Mozilla/5.0 (X11; U; NetBSD i386; en-US; rv:1.7.12) Gecko/20051220
X-Accept-Language: de-de, de, en-us, en
In-Reply-To: <dqgm2f$ije$1@sunnews.cern.ch>
Xref: g2news1.google.com comp.std.c:1146

Laurent Deniau wrote:
> This morning I had to face again the same problem, that is knowing the 
> number of arguments contained in __VA_ARGS__ before its expansion (after 
> its expansion can always be achieved if you can do it before). I found a 
> simple non-iterative solution which may be of interest here as an answer 
> to who will ask in the future for a kind of __VA_NARG__ in the standard 
> and I post it for archiving. May be some more elegant-efficient solution 
> exists?

Thanks for this idea. I really like it.

For those that only want to copy and paste it, here is the expanded version:

/* The PP_NARG macro returns the number of arguments that have been
  * passed to it.
  */

#define PP_NARG(...) \
         PP_NARG_(__VA_ARGS__,PP_RSEQ_N())
#define PP_NARG_(...) \
         PP_ARG_N(__VA_ARGS__)
#define PP_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 PP_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

/* Some test cases */
PP_NARG(A) -> 1
PP_NARG(A,B) -> 2
PP_NARG(A,B,C) -> 3
PP_NARG(A,B,C,D) -> 4
PP_NARG(A,B,C,D,E) -> 5
PP_NARG(1,2,3,4,5,6,7,8,9,0,
         1,2,3,4,5,6,7,8,9,0,
         1,2,3,4,5,6,7,8,9,0,
         1,2,3,4,5,6,7,8,9,0,
         1,2,3,4,5,6,7,8,9,0,
         1,2,3,4,5,6,7,8,9,0,
         1,2,3) -> 63

Note: using PP_NARG() without arguments would violate 6.10.3p4 of ISO C99.


Path: g2news1.google.com!news4.google.com!news.glorb.com!newsfeed00.sul.t-online.de!t-online.de!130.59.10.21.MISMATCH!kanaga.switch.ch!ezmp3.switch.ch!news-zh.switch.ch!switch.ch!cernne03.cern.ch!cern.ch!news
From: Laurent Deniau <laurent...@cern.ch>
Newsgroups: comp.std.c
Subject: Re: __VA_NARG__
Date: Fri, 20 Jan 2006 13:16:24 +0100
Organization: CERN - European Laboratory for Particle Physics
Lines: 64
Message-ID: <dqqkcp$fk1$1@sunnews.cern.ch>
References: <dqgm2f$ije$1@sunnews.cern.ch> <dqqj9j$31l$1@rzsun03.rrz.uni-hamburg.de>
NNTP-Posting-Host: pb-d-128-141-34-204.cern.ch
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Trace: sunnews.cern.ch 1137759449 16001 (None) 128.141.34.204
X-Complaints-To: ne...@sunnews.cern.ch
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.12) Gecko/20051007 Debian/1.7.12-1
X-Accept-Language: en
In-Reply-To: <dqqj9j$31l$1@rzsun03.rrz.uni-hamburg.de>
Xref: g2news1.google.com comp.std.c:1147

Roland Illig wrote:
> Laurent Deniau wrote:
> 
>> This morning I had to face again the same problem, that is knowing the 
>> number of arguments contained in __VA_ARGS__ before its expansion 
>> (after its expansion can always be achieved if you can do it before). 
>> I found a simple non-iterative solution which may be of interest here 
>> as an answer to who will ask in the future for a kind of __VA_NARG__ 
>> in the standard and I post it for archiving. May be some more 
>> elegant-efficient solution exists?
> 
> 
> Thanks for this idea. I really like it.
> 
> For those that only want to copy and paste it, here is the expanded 
> version:
> 
> /* The PP_NARG macro returns the number of arguments that have been
>  * passed to it.
>  */
> 
> #define PP_NARG(...) \
>         PP_NARG_(__VA_ARGS__,PP_RSEQ_N())
> #define PP_NARG_(...) \
>         PP_ARG_N(__VA_ARGS__)
> #define PP_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 PP_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
> 
> /* Some test cases */
> PP_NARG(A) -> 1
> PP_NARG(A,B) -> 2
> PP_NARG(A,B,C) -> 3
> PP_NARG(A,B,C,D) -> 4
> PP_NARG(A,B,C,D,E) -> 5
> PP_NARG(1,2,3,4,5,6,7,8,9,0,
>         1,2,3,4,5,6,7,8,9,0,
>         1,2,3,4,5,6,7,8,9,0,
>         1,2,3,4,5,6,7,8,9,0,
>         1,2,3,4,5,6,7,8,9,0,
>         1,2,3,4,5,6,7,8,9,0,
>         1,2,3) -> 63
> 
> Note: using PP_NARG() without arguments would violate 6.10.3p4 of ISO C99.

Right. I try to ensure that I always have at least one argument in my 
macros which are expected to accept arguments (like for variadic functions).

Thanks for the pretty expanded version.

ld.


Received: by 10.68.211.136 with SMTP id nc8mr498283pbc.6.1335094289119;
        Sun, 22 Apr 2012 04:31:29 -0700 (PDT)
Path: r9ni85324pbh.0!nntp.google.com!news1.google.com!postnews.google.com!glegroupsg2000goo.googlegroups.com!not-for-mail
From: arpad.g...@gmail.com
Newsgroups: comp.std.c
Subject: Re: __VA_NARG__
Date: Sun, 22 Apr 2012 04:31:28 -0700 (PDT)
Organization: http://groups.google.com
Lines: 75
Message-ID: <23410423.357.1335094288749.JavaMail.geo-discussion-forums@ynbv36>
References: <dqgm2f$ije$1@sunnews.cern.ch>
NNTP-Posting-Host: 178.48.71.7
Mime-Version: 1.0
X-Trace: posting.google.com 1335094289 11907 127.0.0.1 (22 Apr 2012 11:31:29 GMT)
X-Complaints-To: groups...@google.com
NNTP-Posting-Date: Sun, 22 Apr 2012 11:31:29 +0000 (UTC)
In-Reply-To: <dqgm2f$ije$1@sunnews.cern.ch>
Complaints-To: groups...@google.com
Injection-Info: glegroupsg2000goo.googlegroups.com; posting-host=178.48.71.7; posting-account=s3bX8goAAADykh5tnw0EN0ZoghWARt3z
User-Agent: G2/1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

2006. janu=C3=A1r 16., h=C3=A9tf=C5=91 18:43:40 UTC+1 id=C5=91pontban Laure=
nt Deniau a k=C3=B6vetkez=C5=91t =C3=ADrta:
> Hi all,
>=20
> A year ago, I was asking here for an equivalent of __VA_NARG__ which=20
> would return the number of arguments contained in __VA_ARGS__ before its=
=20
> expansion. In fact my problem at that time (detecting for a third=20
> argument) was solved by the solution of P. Mensonides. But I was still=20
> thinking that the standard should have provided such a facilities rather=
=20
> easy to compute for cpp.
>=20
> This morning I had to face again the same problem, that is knowing the=20
> number of arguments contained in __VA_ARGS__ before its expansion (after=
=20
> its expansion can always be achieved if you can do it before). I found a=
=20
> simple non-iterative solution which may be of interest here as an answer=
=20
> to who will ask in the future for a kind of __VA_NARG__ in the standard=
=20
> and I post it for archiving. May be some more elegant-efficient solution=
=20
> exists?
>=20
> regards,
>=20
> ld.
>=20
>=20
> Returns NARG, the number of arguments contained in __VA_ARGS__ before=20
> expansion as far as NARG is >0 and <64 (cpp limits):
>=20
> #define PP_NARG( ...) PP_NARG_(__VA_ARGS__,PP_RSEQ_N())
> #define PP_NARG_(...) PP_ARG_N(__VA_ARGS__)
> #define PP_ARG_N(_1,_2,_3,_4,_5,_6,_7,_8,_9,[..],_61,_62,_63,N,...) N
> #define PP_RSEQ_N() 63,62,61,60,[..],9,8,7,6,5,4,3,2,1,0
>=20
> [..] stands for the continuation of the sequence omitted here for=20
> lisibility.
>=20
> PP_NARG(A) -> 1
> PP_NARG(A,B) -> 2
> PP_NARG(A,B,C) -> 3
> PP_NARG(A,B,C,D) -> 4
> PP_NARG(A,B,C,D,E) -> 5
> PP_NARG(A1,A2,[..],A62,A63) -> 63

I extended this to also work correctly with 0 arguments (but now it will ac=
cept only 62 arguments):


#define __VA_NARG__(...) \
=09(__VA_NARG_(_0, ## __VA_ARGS__, __RSEQ_N()) - 1)
#define __VA_NARG_(...) \
=09__VA_ARG_N(__VA_ARGS__)
#define __VA_ARG_N( \
=09 _1, _2, _3, _4, _5, _6, _7, _8, _9,_10, \
=09_11,_12,_13,_14,_15,_16,_17,_18,_19,_20, \
=09_21,_22,_23,_24,_25,_26,_27,_28,_29,_30, \
=09_31,_32,_33,_34,_35,_36,_37,_38,_39,_40, \
=09_41,_42,_43,_44,_45,_46,_47,_48,_49,_50, \
=09_51,_52,_53,_54,_55,_56,_57,_58,_59,_60, \
=09_61,_62,_63,N,...) N
#define __RSEQ_N() \
=0963, 62, 61, 60,                         \
=0959, 58, 57, 56, 55, 54, 53, 52, 51, 50, \
=0949, 48, 47, 46, 45, 44, 43, 42, 41, 40, \
=0939, 38, 37, 36, 35, 34, 33, 32, 31, 30, \
=0929, 28, 27, 26, 25, 24, 23, 22, 21, 20, \
=0919, 18, 17, 16, 15, 14, 13, 12, 11, 10, \
=09 9,  8,  7,  6,  5,  4,  3,  2,  1,  0

Hope this helps.


Received: by 10.68.129.169 with SMTP id nx9mr11342325pbb.2.1335111054227;
        Sun, 22 Apr 2012 09:10:54 -0700 (PDT)
Path: r9ni86043pbh.0!nntp.google.com!news1.google.com!goblin2!goblin1!goblin.stu.neva.ru!news.osn.de!diablo2.news.osn.de!proxad.net!feeder2-2.proxad.net!cleanfeed4-a.proxad.net!nnrp4-2.free.fr!not-for-mail
Message-ID: <4F942D8D.7020203@loria.fr>
Date: Sun, 22 Apr 2012 18:10:53 +0200
From: Jens Gustedt <jens.g...@loria.fr>
User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; de; rv:1.9.2.28) Gecko/20120313 Lightning/1.0b2 Thunderbird/3.1.20
MIME-Version: 1.0
Newsgroups: comp.std.c
Subject: Re: __VA_NARG__
References: <dqgm2f$ije$1@sunnews.cern.ch> <23410423.357.1335094288749.JavaMail.geo-discussion-forums@ynbv36>
In-Reply-To: <23410423.357.1335094288749.JavaMail.geo-discussion-forums@ynbv36>
Lines: 16
Organization: Guest of ProXad - France
NNTP-Posting-Date: 22 Apr 2012 18:10:53 CEST
NNTP-Posting-Host: 81.56.96.245
X-Trace: 1335111053 news-1.free.fr 6629 81.56.96.245:45178
X-Complaints-To: ab...@proxad.net
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit

Hello,
it seems that you are using the ``, ##'' extension of gcc to achieve
your goal. This is not standard, and so in particular this news group
is probably not appropriate.

There are other, more standard ways around this problem of detecting
if a va_arg macro is called with an empty argument. Unfortunately none
of them is completely error prone, but at least some are usably in
practise.

In that sense it would be really nice, I agree, if there would be a
standard functionality to achieve that goal. For the moment the
preprocessing phase doesn't foresee much of "meta" functions. So this
would be a substatial change of the pattern, not likely to happen.

Jens