a silent error

[RoS]

what does it say the standard for below program?
it has to run or not? or it is UB?
it is better know there is a error or the portability?
it compile and run with gcc?

-------------------------------
#include <stdio.h>

int main(void)
{char *p="123";

printf("jijdid\n", p);
return 0;
}
-------------------------------

 

[santosh]

what does it say the standard for below program?
it has to run or not? or it is UB?
it is better know there is a error or the portability?
it compile and run with gcc?

-------------------------------
#include <stdio.h>

int main(void)
{char *p="123";

printf("jijdid\n", p);
return 0;
}

It invokes undefined behaviour.

 

[RoS]

In data Mon, 03 Dec 2007 17:21:48 +0100, RoS scrisse:

--------------------

#include <stdio.h>

int main(void)
{char *p="123";

printf("jijdid\n", p);
return 0;
}
-------------------------------

for me not seem an error

 

[santosh]

In data Mon, 03 Dec 2007 17:21:48 +0100, RoS scrisse:

for me not seem an error

Which is one possible outcome of Undefined behaviour.

 

[Jean-Marc Bourguet]

It invokes undefined behaviour.

Where, why?

 

[santosh]

Where, why?

You are right sorry. Excess arguments after format specifiers are
processed are evaluated and ignored. So no undefined behaviour.

 

[RoS]

In data Mon, 03 Dec 2007 22:28:39 +0530, santosh scrisse:

Which is one possible outcome of Undefined behaviour.

in the sense i always wrong?

i post that because i did read something like above in my C code

i compile it using the current compiler in my pc but i
fear it could be not compile in some other one compiler

 

[Barry Schwarz]

In data Mon, 03 Dec 2007 17:21:48 +0100, RoS scrisse:

It is not an error from a language point of view. The standard
specifically allows it. However, there are some very popular
compilers which attempt to match the types of the subsequent arguments
against the specifications in the format string and will generate a
diagnostic if a mismatch is detected, as in this case. Furthermore,
when I see code like this I immediately assume the programmer forgot
something and wonder what else is missing from the function.


Remove del for email

 

[Harald van Dijk]

It is not an error from a language point of view. The standard
specifically allows it.

It almost does, but almost means it doesn't. It allows it in va_arg,
which may or may not be how printf is implemented. It would be a strange
implementation where this doesn't work, but it would be a valid one.

 

[Flash Gordon]

Harald van Dijk wrote, On 06/12/07 20:38:

It almost does, but almost means it doesn't. It allows it in va_arg,
which may or may not be how printf is implemented. It would be a strange
implementation where this doesn't work, but it would be a valid one.

Actually, the specification for fprintf explicitly specified that extra
parameters are evaluated but otherwise ignored. It is in the first
paragraph of the description in section 7.19.6.1 of n1256. So Barry was
correct.

 

[Harald van Dijk]

Harald van Dijk wrote, On 06/12/07 20:38:

Actually, the specification for fprintf explicitly specified that extra
parameters are evaluated but otherwise ignored. It is in the first
paragraph of the description in section 7.19.6.1 of n1256. So Barry was
correct.

Sorry, you and Barry are right, of course. I was thinking of something
else.