passing uninitialized parameters

[Michael]

Hi,

once I read here that it is not 'a good idea'
to pass variables that are not initialized to
a function.

I have
void something ( double *vector );

....
int main ( void )
{
double vector[3];

something ( vector );
...
}

'something' puts its results in the array 'vector'.
Do I have to initialize vector before passing it?
I wonder because if i don't do do, 'splint' blames me of
being lazy ...

Thanks,

Michael

[Jens.Toerring]

Michael <mi****@gmx.net > wrote:

once I read here that it is not 'a good idea'
to pass variables that are not initialized to
a function. I have
void something ( double *vector ); ...
int main ( void )
{
double vector[3]; something ( vector );
...
} 'something' puts its results in the array 'vector'.
Do I have to initialize vector before passing it?

No, 'vector' is initialized, only the elements of vector aren't.
What arrives in the function is a well-initialized pointer to
the first element of the array you defined in main() and you can
use that to set the elements. I guess the warning was more about
something like

int main( void );
{
double *vector;
something( vector );
...
}

That wouldn't make any sense - if you use the (uninitialized) value
of this 'vector' in the function hell would break loose, and if you
assign a value to 'vector' in the function it won't be visible in
the caller, so passing the argument is useless.

Regards, Jens
--
\ Jens Thoms Toerring ___ Je***********@p hysik.fu-berlin.de
\______________ ____________ http://www.toerring.de

[Dan Pop]

In <cp**********@r zcomm2.rz.tu-bs.de> Michael <mi****@gmx.net > writes:

once I read here that it is not 'a good idea'
to pass variables that are not initialized to
a function.
Indeed, it's an attrocious idea, resulting in undefined behaviour most of
the time.
I have
void something ( double *vector );

...
int main ( void )
{
double vector[3];

something ( vector );
...
}

'something' puts its results in the array 'vector'.
This is not an example of passing an uninitialised variable to a
function. You're passing the address of vector[] to something(), not
its uninitialised contents.
Do I have to initialize vector before passing it?
No, as long as it is used only for output purposes (i.e. the function
completely ignores its current contents).
I wonder because if i don't do do, 'splint' blames me of
being lazy ...

splint doesn't know whether vector[] is passed to something() for input
or output purposes. If it's passed for input purposes, your code is
wrong, otherways it isn't.

Other languages remove this ambiguity, by allowing the programmer to
specify, in the function "prototype" , which arguments are input arguments
and which are output arguments. C doesn't, hence splint's dilemma.

Dan
--
Dan Pop
DESY Zeuthen, RZ group
Email: Da*****@ifh.de
Currently looking for a job in the European Union

[Flash Gordon]

On Fri, 10 Dec 2004 15:46:47 +0100
Michael <mi****@gmx.net > wrote:

Hi,

once I read here that it is not 'a good idea'
to pass variables that are not initialized to
a function.
You should pass either an initialised variable or something that is not
a variable.
I have
void something ( double *vector );

...
int main ( void )
{
double vector[3];

something ( vector );
...
}

'something' puts its results in the array 'vector'.
Do I have to initialize vector before passing it?
Here you are passing a *pointer* to the first element of vector. In
other words you are effectively passing the address of vector. This
obviously is not a variable, although it might not be constant if it is
an automatic object (which it is in your example). So, if the definition
of function"someth ing" is such that it is guaranteed to no read any
location that it has not written to then there is no need to initialise
vector. In fact, I would often consider it to be a bad thing since
instead of vector you might do

int main ( void )
{
double vector[1000];

something ( vector );
/* ... */
}

Obviously in this instance initialising vector would be rather
inefficient if the initialisation was not required.
I wonder because if i don't do do, 'splint' blames me of
being lazy ...

This is one of the occasions when you need to think about whether the
warning from splint is really a problem.

By asking the question rather than blindly shutting splint up you are
*definitely* doing the right thing.
--
Flash Gordon
Living in interesting times.
Although my email address says spam, it is real and I read it.

[CBFalconer]

Flash Gordon wrote:

.... snip ...
This is one of the occasions when you need to think about whether
the warning from splint is really a problem.

By asking the question rather than blindly shutting splint up you
are *definitely* doing the right thing.

The purpose of splint and lint is not to point out errors (although
it may), but to point out areas that require examination. It will
usually tell you why it is suspicious.

--
Chuck F (cb********@yah oo.com) (cb********@wor ldnet.att.net)
Available for consulting/temporary embedded and systems.
<http://cbfalconer.home .att.net> USE worldnet address!

[Michael]

Hi,

thank you all for these answers. Thats what I want
to know.

Bye, and have a nice December

Michael