pasting "." and "something" does not give a valid preprocessing token..

pasting "." and "something" does not give a valid preprocessing token..

[mskhan]

I've just installed gcc 3.2 and I get the warning from the subject line when
compiling code that worked just fine with 2.95.3. Can anyone tell me what the
 warning means and how it can be rectified.

The offending code is a macro :

#define EMUL_TRACE(level, format, args...)     proc::theTrace->logTrace((level,  ##format , ##args)


where "Proc" is a class defined as

class Proc
{

. ..........

static TraceLog * theTrace;

}

and TraceLog  is another class defined as

class TraceLog
{
.............
public:
............
void logTrace(int severity, char* stringToBeLogged, ...)
.............

}

Thanks in advance....
Sanjay



"DISCLAIMER: This message is proprietary to Hughes Software Systems Limited
(HSS) and is intended solely for the use of the individual to whom it is
addressed. It may contain  privileged or confidential information and
should not be circulated or used for any purpose other than for what it is
intended. If you have received this message in error, please notify the
originator immediately. If you are not the intended recipient, you are
notified that you are strictly prohibited from using, copying, altering, or
disclosing the contents of this message. HSS accepts no responsibility for
loss or damage arising from the use of the information transmitted by this
email including damage from virus."

Re: pasting "." and "something" does not give a valid preprocessing token..

[LLeweLLyn Reese]

mskhan@hss.hns.com writes:

> I've just installed gcc 3.2 and I get the warning from the subject line when
> compiling code that worked just fine with 2.95.3. Can anyone tell me what the
>  warning means and how it can be rectified.
> 
> The offending code is a macro :
> 
> #define EMUL_TRACE(level, format, args...) proc::theTrace->logTrace((level,  ##format , ##args)

I think you do not  want the token pasting ## before format or args.

#define EMUL_TRACE(level, format, args...) proc::theTrace->logTrace((level,  format , args)

should do the trick.

The token-pasting operator is only for making tokens - 'foo##bar'
    becomes the single token 'foobar' and its result must always
    produce a single valid token (or the results are undefined). Your
    examples would have pasted a comma onto the begining of a
    multichar token, and , is only a valid token by itself.

> 
> 
> where "Proc" is a class defined as
> 
> class Proc
> {
> 
> . ..........
> 
> static TraceLog * theTrace;
> 
> }
> 
> and TraceLog  is another class defined as
> 
> class TraceLog
> {
> .............
> public:
> ............
> void logTrace(int severity, char* stringToBeLogged, ...)
> .............
> 
> }
> 
[snip]

Re: pasting "." and "something" does not give a valid preprocessing token..

[Marc Eberhard]

Dear Nathan!

On Tue, Aug 19, 2003 at 09:35:18AM +0100, Nathan Sidwell wrote:
> Marc Eberhard wrote:
> >  this-> ## v
> >  mynamespace:: ## v
> >
> >Is there a way to convince the preprocessor, that I know, what I'm doing
> >and to accept the result as a vaild token (what it is in C++)? This is
> >with gcc (GCC) 3.3.1 20030626 (Debian prerelease).

> they are not valid single tokens. The preprocessor is correct. Why do you 
> think you need a ## at all?

Ahhh, that's the solution. :-) I somehow thought, that I would have to use
## to combine the two and 2.95 happyly did the right thing with it. Only
when upgrading to 3.3 I suddenly ran into this problem. But you're perfectly
right... simply leaving the ## out solves the problem! Many thanks, I would
never have guessed, that the solution is that easy!!!

Thanks,
Marc
_______________________________________________________________________________

email: marc@greenie.net, marc.eberhard@alumni.tum.de, marc@affs.org.uk
email: m.a.eberhard@aston.ac.uk, web: http://www.aston.ac.uk/~eberhama/

Re: pasting "." and "something" does not give a valid preprocessing token..

[Nathan Sidwell]

Marc Eberhard wrote:
> Dear LLeweLLyn Reese!
> 
> You wrote:
> 
>>The token-pasting operator is only for making tokens - 'foo##bar'
>>     becomes the single token 'foobar' and its result must always
>>     produce a single valid token (or the results are undefined). Your
>>     examples would have pasted a comma onto the begining of a
>>     multichar token, and , is only a valid token by itself.
> 
> 
> Ran into the same problem with a construct like this:
> 
>   this-> ## v
> 
> If I change it to
> 
>   this->x ## v
> 
> it works, but that's obviously not what I want. Same problem when
> constructing namespace defines like:
> 
>   mynamespace:: ## v
> 
> Is there a way to convince the preprocessor, that I know, what I'm doing and
> to accept the result as a vaild token (what it is in C++)? This is with gcc
> (GCC) 3.3.1 20030626 (Debian prerelease).
they are not valid single tokens. The preprocessor is correct. Why do you think
you need a ## at all?

nathan

-- 
Nathan Sidwell    ::   http://www.codesourcery.com   ::     CodeSourcery LLC
          The voices in my head said this was stupid too
nathan@codesourcery.com    ::     http://www.planetfall.pwp.blueyonder.co.uk

Re: pasting "." and "something" does not give a valid preprocessing token..

[Marc Eberhard]

Dear LLeweLLyn Reese!

You wrote:
> The token-pasting operator is only for making tokens - 'foo##bar'
>      becomes the single token 'foobar' and its result must always
>      produce a single valid token (or the results are undefined). Your
>      examples would have pasted a comma onto the begining of a
>      multichar token, and , is only a valid token by itself.

Ran into the same problem with a construct like this:

  this-> ## v

If I change it to

  this->x ## v

it works, but that's obviously not what I want. Same problem when
constructing namespace defines like:

  mynamespace:: ## v

Is there a way to convince the preprocessor, that I know, what I'm doing and
to accept the result as a vaild token (what it is in C++)? This is with gcc
(GCC) 3.3.1 20030626 (Debian prerelease).

Thanks,
Marc
_______________________________________________________________________________

email: marc@greenie.net, marc.eberhard@alumni.tum.de, marc@affs.org.uk
email: m.a.eberhard@aston.ac.uk, web: http://www.aston.ac.uk/~eberhama/

Re: pasting "." and "something" does not give a valid preprocessing token..

[mskhan]

Hi Reese,
       Thanks a lot for your reply. I have done as you had suggested.
However it has now started giving Parse Error wherever the macro is being
called.

The description of the error is as follows::

UIAProc.cpp: In member   function `virtual void
UIAProc::UDPEventHandler(void*, int, const char*,  int)':
UIAProc.cpp:275: parse error   before `)' token

expecting o hear from you asap

thanks-
Sanjay





LLeweLLyn Reese <llewelly@lifesupport.shutdown.com> on 04/24/2003 12:07:55
PM

To:    Mohammed Sanjay Khan/HSS@HSS
cc:    gcc-help@gcc.gnu.org

Subject:    Re: pasting "." and "something" does not give a valid
       preprocessing token..


mskhan@hss.hns.com writes:

> I've just installed gcc 3.2 and I get the warning from the subject line
when
> compiling code that worked just fine with 2.95.3. Can anyone tell me what
the
>  warning means and how it can be rectified.
>
> The offending code is a macro :
>
> #define EMUL_TRACE(level, format, args...)
proc::theTrace->logTrace((level,  ##format , ##args)

I think you do not  want the token pasting ## before format or args.

#define EMUL_TRACE(level, format, args...) proc::theTrace->logTrace((level,
format , args)

should do the trick.

The token-pasting operator is only for making tokens - 'foo##bar'
    becomes the single token 'foobar' and its result must always
    produce a single valid token (or the results are undefined). Your
    examples would have pasted a comma onto the begining of a
    multichar token, and , is only a valid token by itself.

>
>
> where "Proc" is a class defined as
>
> class Proc
> {
>
> . ..........
>
> static TraceLog * theTrace;
>
> }
>
> and TraceLog  is another class defined as
>
> class TraceLog
> {
> .............
> public:
> ............
> void logTrace(int severity, char* stringToBeLogged, ...)
> {
> ......
> }
> .............
>
> }
>
 [snip]

Re: pasting "." and "something" does not give a valid preprocessing token

[Mihnea Balta]

Oh, sorry. I thought it wouldn't work without ##. Obviously, I was wrong, 
since i removed the ## and it works with no warnings. 

Thanks.

On Monday 10 March 2003 11:02, Momchil Velikov wrote:
> >>>>> "Mihnea" == Mihnea Balta <dark_lkml@mymail.ro> writes:
>
>     Mihnea>         conf.##__prop
>
> conf.__prop is not a valid proprocessor token.
>
> Just remove the token paste operator.
>
> ~velco
>
> ------------------------------------------------------------------------
> http://felicitari.mymail.ro/

Re: pasting "." and "something" does not give a valid preprocessing token

[Florian Weimer]

Mihnea Balta <dark_lkml@mymail.ro> writes:

> Sorry, I should have said "sv_port" instead of something in the subject line, 
> as that is the exact warning produced. It doesn't like my conf.##__prop 
> thing.

Maybe I'm missing something obvious, but why do you need the "##" in
this place at all?

Re: pasting "." and "something" does not give a valid preprocessing token

[Momchil Velikov]

>>>>> "Mihnea" == Mihnea Balta <dark_lkml@mymail.ro> writes:

    Mihnea>         conf.##__prop

conf.__prop is not a valid proprocessor token.

Just remove the token paste operator.

~velco

Re: pasting "." and "something" does not give a valid preprocessing token

[Mihnea Balta]

Sorry, I should have said "sv_port" instead of something in the subject line, 
as that is the exact warning produced. It doesn't like my conf.##__prop 
thing.

On Monday 10 March 2003 10:57, Mihnea Balta wrote:
> Hello,
>
> I've just installed gcc 3.2.2 and I get the warning from the subject line
> when compiling code that worked just fine with 2.95.3. The preprocessor
> output is what I expect it to be, but I'd like to know what the warning
> means. The offending code is:
>
> #define CONF_ADD_INT(__prop)                                               
> \ if(!strncmp(key, #__prop, strlen(#__prop))){                            \
> conf.##__prop = atoi(val);                                          \
> return 1;                                                           \ }
>
> statement that produces the warning:
> CONF_ADD_INT(sv_port);
> The preprocessor output is exactly what I expected:
>
> if(!strncmp(key, "sv_port", strlen("sv_port"))){ conf.sv_port = atoi(val);
> return 1; };
>
> Thanks.
>
> ------------------------------------------------------------------------
> http://felicitari.mymail.ro/

pasting "." and "something" does not give a valid preprocessing token

[Mihnea Balta]

Hello,

I've just installed gcc 3.2.2 and I get the warning from the subject line when 
compiling code that worked just fine with 2.95.3. The preprocessor output is 
what I expect it to be, but I'd like to know what the warning means.
The offending code is:

#define CONF_ADD_INT(__prop)                                                \
    if(!strncmp(key, #__prop, strlen(#__prop))){                            \
        conf.##__prop = atoi(val);                                          \
        return 1;                                                           \
    }

statement that produces the warning:
CONF_ADD_INT(sv_port);
The preprocessor output is exactly what I expected:

if(!strncmp(key, "sv_port", strlen("sv_port"))){ conf.sv_port = atoi(val); 
return 1; };

Thanks.