malloc and functions

[raphfrk]

I am having an issue with malloc and gcc. Is there something wrong
with my code or is this a compiler bug ?

I am running this program:

#include <stdio.h>
#include <stdlib.h>

typedef struct pxl {
double lon, lat;
double x,y,z;
double px,py;
} pixel;
struct chn {
int index;
struct nde *node;
};

struct nde {
pixel position;
struct chn *forward;
struct chn *reverse;
int *chain_num;
int size;
};

typedef struct chn chain;
typedef struct nde node;

node *createnode( );

main()
{

node *startn;

startn = createnode( );

startn->size = 2;
startn->forward = malloc( 2 * sizeof( chain * ) );
startn->reverse = malloc( 2 * sizeof( chain * ) );

printf("sf %p\nsr %p\n\n",
startn->forward,
startn->reverse
);

printf("sf %p\nsr %p\n\n",
startn->forward,
startn->reverse
);

}

node *createnode( )
{
node *node;
node = malloc( sizeof(node) );
return node;
}

I get the following output.

sf 0x660168
sr 0x660178

sf 0xa383731
sr 0x660178

The startn->forward point changes between the two printfs despite
there being no code between them.

[Richard]

raphfrk <ra*****@netsca pe.netwrites:

>
node *createnode( )
{
node *node;
node = malloc( sizeof(node) );

Just a quick glance, the above "confused me".

node? Try "node * pnode" ...

No idea if it will fix your problem.

[raphfrk]

On Jul 15, 4:09 pm, Richard <rgr...@gmail.c omwrote:

raphfrk <raph...@netsca pe.netwrites:

node *createnode( )
{
node *node;
node = malloc( sizeof(node) );

Just a quick glance, the above "confused me".

node? Try "node * pnode" ...

No idea if it will fix your problem.

Yeah, that fixed it. I have been trying to figure it out for ages
(well about 3 hours). gcc must be misinterpreting what I wanted. It
didn't give a warning though.

Thanks alot.

[Robert Gamble]

On Jul 15, 10:57 am, raphfrk <raph...@netsca pe.netwrote:

I am having an issue with malloc and gcc. Is there something wrong
with my code or is this a compiler bug ?

I am running this program:

#include <stdio.h>
#include <stdlib.h>

typedef struct pxl {
double lon, lat;
double x,y,z;
double px,py;
} pixel;

struct chn {
int index;
struct nde *node;
};

struct nde {
pixel position;
struct chn *forward;
struct chn *reverse;
int *chain_num;
int size;
};

typedef struct chn chain;
typedef struct nde node;

node *createnode( );

main()

'int main (void)' is better.

{

node *startn;

startn = createnode( );

startn->size = 2;
startn->forward = malloc( 2 * sizeof( chain * ) );
startn->reverse = malloc( 2 * sizeof( chain * ) );

printf("sf %p\nsr %p\n\n",
startn->forward,
startn->reverse
);

printf("sf %p\nsr %p\n\n",
startn->forward,
startn->reverse
);

The %p conversion specifier expects a void * argument, you need to
cast your arguments appropriately.

}

node *createnode( )
{
node *node;

Yuck. Using the same name for a typedef and a variable (esp. one that
is a pointer to that type) is ugly and has the potential to cause
great confusion, see below.

node = malloc( sizeof(node) );

Hint: Does sizeof apply to the typedef name or the variable name?
It's probably not what you expect.
You should also check the return value of malloc for failure.

return node;
}

I get the following output.

sf 0x660168
sr 0x660178

sf 0xa383731
sr 0x660178

The startn->forward point changes between the two printfs despite
there being no code between them.

Fix the issues mentioned above and let us know if you still have a
problem.

Robert Gamble

[rzed]

raphfrk <ra*****@netsca pe.netwrote in
news:11******** **************@ d55g2000hsg.goo glegroups.com:

On Jul 15, 4:09 pm, Richard <rgr...@gmail.c omwrote:

>raphfrk <raph...@netsca pe.netwrites:

node *createnode( )
{
node *node;
node = malloc( sizeof(node) );

Just a quick glance, the above "confused me".

node? Try "node * pnode" ...

No idea if it will fix your problem.

Yeah, that fixed it. I have been trying to figure it out for
ages (well about 3 hours). gcc must be misinterpreting what I
wanted. It didn't give a warning though.

Thanks alot.

It may have fixed it, but it doesn't explain it. Your call to
createnode came before the first printf. So why does the confusion
matter to the printf?

By the way, I repeated the printf more times, and each time after
the first, it prints the same values as the second printf. That
is, it doesn't change after that first time. What could possibly
be happening under the hood to get that bizarre result?

--
rzed

[Richard]

raphfrk <ra*****@netsca pe.netwrites:

On Jul 15, 4:09 pm, Richard <rgr...@gmail.c omwrote:

>raphfrk <raph...@netsca pe.netwrites:

node *createnode( )
{
node *node;
node = malloc( sizeof(node) );

Just a quick glance, the above "confused me".

node? Try "node * pnode" ...

No idea if it will fix your problem.

Yeah, that fixed it. I have been trying to figure it out for ages
(well about 3 hours). gcc must be misinterpreting what I wanted. It
didn't give a warning though.

Rubbish in rubbish out :-; Personally, even if it ever was "ok" I would
never have a variable with the same name as a type. A variable should
reflect the object to which it refers. In this case a pointer to a node
- hence a "pnode".

>
Thanks alot.

No probs. Glad to help. Good luck.

Ask in a gcc newsgroup here about the warnings - seems strange:

gnu.gcc.help

[Kenny McCormack]

In article <87************ @gmail.com>, Richard <rg****@gmail.c omwrote:
....

>Rubbish in rubbish out :-; Personally, even if it ever was "ok" I would
never have a variable with the same name as a type.

This is common with structs. You frequently see code like: struct stat stat;
Of course: int int
doesn't work as well.

>A variable should reflect the object to which it refers. In this case a
pointer to a node - hence a "pnode".

[Robert Gamble]

On Jul 15, 11:48 am, gaze...@xmissio n.xmission.com (Kenny McCormack)
wrote:

In article <87ejj9isnb.... @gmail.com>, Richard <rgr...@gmail.c omwrote:

...

Rubbish in rubbish out :-; Personally, even if it ever was "ok" I would
never have a variable with the same name as a type.

This is common with structs. You frequently see code like: struct stat stat;

This is true but the issue is much worse with typedefs since they
share the same namespace as ordinary identifiers and thus have greater
potential for confusion.

Of course: int int
doesn't work as well.

Well that's because int is a keyword.

typedef int INT;
{
INT INT;
....
}

This "works" but like the example presented by the OP is horrible
form.

Robert Gamble

[Richard]

ga*****@xmissio n.xmission.com (Kenny McCormack) writes:

In article <87************ @gmail.com>, Richard <rg****@gmail.c omwrote:
...

>>Rubbish in rubbish out :-; Personally, even if it ever was "ok" I would
never have a variable with the same name as a type.

This is common with structs. You frequently see code like: struct stat stat;
Of course: int int
doesn't work as well.

I simply don't like it. More than likely it will confuse a debugger too
never mind someone maintaining your code.

[Thad Smith]

rzed wrote:

raphfrk <ra*****@netsca pe.netwrote in
news:11******** **************@ d55g2000hsg.goo glegroups.com:

>>On Jul 15, 4:09 pm, Richard <rgr...@gmail.c omwrote:

>>>raphfrk <raph...@netsca pe.netwrites:

node *createnode( )
{
node *node;
node = malloc( sizeof(node) );

Just a quick glance, the above "confused me".

node? Try "node * pnode" ...

No idea if it will fix your problem.

Yeah, that fixed it. I have been trying to figure it out for
ages (well about 3 hours). gcc must be misinterpreting what I
wanted.

That's /one/ way of saying that you don't know the language.

It didn't give a warning though.

Try turning up the warnings. If you still doesn't warn, try using lint.
You are allowed to redeclare identifiers in an inner scope. It has good
uses, but pitfalls as well. You found one of the pitfalls.

It may have fixed it, but it doesn't explain it. Your call to
createnode came before the first printf. So why does the confusion
matter to the printf?

The error did not allocate sufficient space for struct nde.
Initializing that struct in the following code placed values into
unallocated space. printf probably used that space.

By the way, I repeated the printf more times, and each time after
the first, it prints the same values as the second printf. That
is, it doesn't change after that first time. What could possibly
be happening under the hood to get that bizarre result?

Writing beyond the allocated bounds of an object often has surprising
results.

--
Thad