maart@cs.vu.nl (Maarten Litmaath) (12/03/88)
gwyn@smoke.BRL.MIL (Doug Gwyn ) writes: \In article <6625@csli.STANFORD.EDU> wagner@arisia.xerox.com (Juergen Wagner) \writes: \> proc/**/VERSION \>People relying on this bug should change their habits. There are better ways \>to concatenate tokens. \No, for Reiser-based preprocessors there aren't any better ways. \ANSI-style token pasting is fairly new, and many C implementations \in current use do not support it. Why didn't the ANSI committee legalize that current practice, instead of defining a new paste operator (##)? -- fcntl(fd, F_SETFL, FNDELAY): |Maarten Litmaath @ VU Amsterdam: let's go weepin' in the corner! |maart@cs.vu.nl, mcvax!botter!maart
gwyn@smoke.BRL.MIL (Doug Gwyn ) (12/04/88)
In article <1747@solo3.cs.vu.nl> maart@cs.vu.nl (Maarten Litmaath) writes: >Why didn't the ANSI committee legalize that current practice, instead of >defining a new paste operator (##)? Because there were probably more C implementations that followed the explicit rule stated in the base document (K&R 1st Ed.) than that violated it. Although X3J11 couldn't legalize this behavior, it did recognize the utility of token pasting, which is why it designed a legal solution to it. I suggest adding a paste(a,b) macro to your equivalent of <std.h>, and define paste(a,b) according to the specific implementation. Be advised that many pre-ANSI C implementations cannot implement this macro.
henry@utzoo.uucp (Henry Spencer) (12/06/88)
In article <1747@solo3.cs.vu.nl> maart@cs.vu.nl (Maarten Litmaath) writes: >Why didn't the ANSI committee legalize that current practice, instead of >defining a new paste operator (##)? Because that current practice was current practice only on some compilers; on others it did not work and could not easily be implemented. In particular, it is hard to make it work under a token-based preprocessor. It's also hard to define in a precise way that doesn't cause trouble with other things. (He who does not value a precise definition of this stuff has never tried to implement it.) -- SunOSish, adj: requiring | Henry Spencer at U of Toronto Zoology 32-bit bug numbers. | uunet!attcan!utzoo!henry henry@zoo.toronto.edu
orr@cs.glasgow.ac.uk (Fraser Orr) (11/09/89)
I would like to define a cpp macro along the lines of
#define list(name,type) \
typedef _${name} { \
struct _${name} _${name}_Link \
${type} _${name}_Data } ${name} ;
(where I've emphasised where I want param expansion with ${...})
The problem is joining up the bits. "_name" does not expand, because it
is a whole token. Ideally what is required is something to join together
the tokens - perhaps PASTE(_,name). I seem to remember seeing this
somewhere but I've forgotten where. Any ideas? (e-mail => summary).
==Fraser Orr ( Dept C.S., Univ. Glasgow, Glasgow, G12 8QQ, UK)
UseNet: {uk}!cs.glasgow.ac.uk!orr JANET: orr@uk.ac.glasgow.cs
ARPANet(preferred xAtlantic): orr%cs.glasgow.ac.uk@nsfnet-relay.ac.ukhurwitz@enuxha.eas.asu.edu (Roger A. Hurwitz) (02/10/91)
I have heard about token pasting, and I have fooled around with the C ## preprocessor operator, but I do not understand the need for this capability. Could someone provide an example of when/where/how token pasting comes in handy? Thanks, Roger Hurwitz hurwitz@enuxha.eas.asu.edu
bliss@sp64.csrd.uiuc.edu (Brian Bliss) (02/12/91)
In article <2261@enuxha.eas.asu.edu>, hurwitz@enuxha.eas.asu.edu (Roger A. Hurwitz) writes: |> |> I have heard about token pasting, and I have fooled around with |> the C ## preprocessor operator, but I do not understand the |> need for this capability. Could someone provide an example |> of when/where/how token pasting comes in handy? |> |> Thanks, |> Roger Hurwitz |> hurwitz@enuxha.eas.asu.edu The last time I needed to do it was when writing a debugger. For each simple type, I had an associated type descriptor and symbol. so I needed to say: #define init(keyword) keyword ## _symbol = \ symbol_alloc ("keyword", keyword ## _type) init (int) init (short) init (long) etc. when I tried to port the thing, the token pasting didn't work, putting a macro argument inside a string didn't result in the correct substitution, etc., so I gave up and hard-coded the whole damn thing. bb
sanders@peyote.cactus.org (Tony Sanders) (02/12/91)
In article <1991Feb11.195439.17787@csrd.uiuc.edu> bliss@sp64.csrd.uiuc.edu (Brian Bliss) writes: > #define init(keyword) keyword ## _symbol = \ > symbol_alloc ("keyword", keyword ## _type) should be: #ifdef __STDC__ #define init(keyword) keyword ## _symbol = \ symbol_alloc (# keyword, keyword ## _type) #else #define init(keyword) keyword ## _symbol = \ symbol_alloc ("keyword", keyword ## _type) #endif Which I use like this: #ifdef __STDC__ #define DPV(variable,type) if (debug) fprintf(stderr, \ "%s:%d, " # variable " = %" # type "\n",__FILE__,__LINE__,variable) #else #define DPV(variable,type) if (debug) fprintf(stderr, \ "%s:%d, variable = %type\n",__FILE__,__LINE__,variable) #endif It's quite ugly but it works great. Your mileage may vary. This works because # keyword converts to "expaned_keyword" and "foo" "bar" "bat" "baz" converts to "foobatbatbaz". -- sanders@peyote.cactus.org First rule of software: Throw the first one away. and so on... I am not an IBM representative and I speak only for myself.