jmartin@secola.Columbia.NCR.COM (John Martinez) (04/02/91)
I seem to recall a thread of discussion a couple of weeks ago
about how to "stringize" a token using the K&R C preprocessor (ANSI C
provides the # operator to do this.) Anyway, as I recall, the concensus
was that it couldn't be done. (At least, no one submitted a method, as
far as I know...)
Well, now I had to try (the Mt. Everest principle, you know,)
so I locked myself in my laboratory with a six-pack of Mountain Dew and
a cold pizza, until I discovered a magic spell that (sort of) did the
trick:
#define quote "
#define stringize(A) quote A"
this is _almost_ right; stringize(foobar) --(magic)--> " foobar"
The next problem is to remove the leading space (if you don't
put a space between 'quote' and 'A' in the macro, it sees 'quoteA' as a
token, which doesn't resolve to anything interesting..)
Another slice of cold pizza later.....
-(POOF)-
*********
static char *global_char_ptr; /* 'static' is just a personal preference. */
#define quote "
#define stringize(a) (global_char_ptr=quote a",++global_char_ptr)
main()
{
int foo=5;
printf("%s = %d\n",stringize(foo),foo);
}
*********
_my_ system's preprocessor converts the printf line to:
printf("%s = %d\n",(global_char_ptr=" foo",++global_char_ptr),foo);
which outputs:
foo = 5
Voila!
NOTE: I'm sure there are lots of netters who:
a) already knew about this
b) saw a solution just like it in a posting I never saw
_or_
c) think the whole discussion is dumb since "the whole world should have
bought an ANSI C compiler 4.7 microseconds after the standards were
signed, no matter what the cost, and if you can't find a compiler you
can just write your own in 2 hours, otherwise you're not a real programmer"
I have a established a special account, devnull@TheAbyss for
you to reply by mail. Seriously - I posted this because I thought it
would be of use to someone, and I am pretty darn happy about it, so
don't rain on my parade, ok?
---
John V.Martinez "Yeah, I got a college degree, but somehow I
NCR Network Products Division managed to learn something in spite of it."
jmartin@secola.Columbia.NCR.COM
(803)772-8854 vplus:633-8854volpe@camelback.crd.ge.com (Christopher R Volpe) (04/02/91)
I thought just about all K&R1 based compilers expanded formal arguments within quotes, so the following would work: #define STRINGIZE(x) "x" STRINGIZE(foo) ==> "foo" At least it works on SunOS cc. ================== Chris Volpe G.E. Corporate R&D volpecr@crd.ge.com
gwyn@smoke.brl.mil (Doug Gwyn) (04/03/91)
In article <18148@crdgw1.crd.ge.com> volpe@camelback.crd.ge.com (Christopher R Volpe) writes: >I thought just about all K&R1 based compilers expanded formal arguments >within quotes, so the following would work: >#define STRINGIZE(x) "x" No, in fact K&R1 says that the argument must not be replaced within string literals in the replacement string. The behavior you cite is a long-standing bug in the Reiser CPP that was shipped with UNIX for many years.
shaunc@gold.gvg.tek.com (Shaun Case) (04/03/91)
In article <323@secola.Columbia.NCR.COM> jmartin@secola.Columbia.NCR.COM (John V. Martinez) writes: > > I seem to recall a thread of discussion a couple of weeks ago >about how to "stringize" a token using the K&R C preprocessor (ANSI C >provides the # operator to do this.) Anyway, as I recall, the concensus >was that it couldn't be done. (At least, no one submitted a method, as >far as I know...) > > Well, now I had to try (the Mt. Everest principle, you know,) >so I locked myself in my laboratory with a six-pack of Mountain Dew and >a cold pizza, until I discovered a magic spell that (sort of) did the >trick: > >#define quote " >#define stringize(A) quote A" > >this is _almost_ right; stringize(foobar) --(magic)--> " foobar" > Hey, I like that! It gave me an idea, too... I tried the following, and it worked under some random version of Ultrix: --- cut here--- #include <stdio.h> #define quote " #define stringize(a) (quote a"+1) main() { printf("[%s]\n",stringize(foo)); } --- cut here --- Is there some reason why this is a bad idea? // Shaun //
cnwietho@immd4.informatik.uni-erlangen.de (Carsten Wiethoff) (04/04/91)
I prefer #define QUOTE(x) "x #define STRINGINIZE(x) QUOTE(x)" which does not have the problem with spaces before x. Carsten Wiethoff | cnwietho@immd4.informatik.uni-erlangen.de
barnettr@snaggle.rtp.dg.com (Richard Barnette) (04/04/91)
In article <323@secola.Columbia.NCR.COM> jmartin@secola.Columbia.NCR.COM (John V. Martinez) writes: > > I seem to recall a thread of discussion a couple of weeks ago >about how to "stringize" a token using the K&R C preprocessor (ANSI C >provides the # operator to do this.) Anyway, as I recall, the concensus >was that it couldn't be done. (At least, no one submitted a method, as >far as I know...) > Some preprocessors will accept #define stringize(A) "A" GNU C will do this if you say -traditional. The AT&T unix compilers tend to do this (but I reckon it depends on your system). Richard Barnette | Data General Corporation | obligatory (in)famous quote: (919) 248-6225 | RTP, NC 27709 | You wascal wabbit! Commercial Languages | 62 T.W. Alexander Drive | Wandering wizards won't barnettr@dg-rtp.dg.com | win! - /usr/lib/sendmail
timr@gssc.UUCP (Tim Roberts) (04/05/91)
In article <1991Apr3.193750.14282@informatik.uni-erlangen.de> cnwietho@immd4.informatik.uni-erlangen.de (Carsten Wiethoff) writes: >I prefer > >#define QUOTE(x) "x >#define STRINGINIZE(x) QUOTE(x)" > >which does not have the problem with spaces before x. If that construct works for some preprocessor, then wouldn't the much simpler #define STRINGINIZE(x) "x" also work? -- timr@gssc.gss.com Tim N Roberts, CCP Graphic Software Systems Beaverton, OR This is a very long palindrome. .emordnilap gnol yrev a si sihT