Thread: Link lists as datastructure

Hi,everybody!
The program reads in three types of input data using link lists.

I wrote several structures but i really don't know how to use link lists to implement it!
I'll appreciate if anybody give me any hint how to begin it!

struct store{
char * recordtype;
int itemqty;
float itemprice;
int markup;
};

this should be your declaration

Code:

/*FILE store.h*/

#ifndef _STORE_H_INC_
#define _STORE_H_INC_
typedef struct _store{ 
   char * recordtype; 
   int itemqty; 
   float itemprice; 
   int markup; 

   struct st_n {  /*linked list node*/
       _store *pnext,*pprev;
   }link;
}store;

struct st_list{
    struct st_h { /*linked list header...*/
         _store *pfirst, *plast;
    }head;
    int count;
};

#endif

As for the Linked list implementation
just look of the macroses in the attachment
i made...

Usage of some of the macroses:

Code:

/*LST_INIT:  init list*/
st_list some_list;
LST_INIT(&some_list.head);

Code:

/*LST_INS_TAIL:  inserts a node at the tail*/
st_list some_list;
store some_store;
LST_INS_TAIL(&some_list.head, &some_node,st_n);

Code:

/*LST_INS_HEAD:  inserts a node at the head*/
st_list some_list;
store some_store;
LST_INS_HEAD(&some_list.head, &some_node,st_n);

Code:

/*LST_REMOVE(head, elm, field):  removes a node from a list
WARNING!! this does not free the memory!
You should save the pointer 
and delete it after that*/

st_list some_list;
store *lp_some_store;/*some valid allocated pointer (malloc-ed)*/
store *save = lp_some_store; /*save the pointer*/
/*remove from list*/
LST_REMOVE(&some_list.head, lp_some_store, st_n)
/*free the mem*/
free(save);

LST_REMOVE(head, elm, field)

also if you allocate the nodes/stores
with malloc you should destroy lists like this:

Code:

void destroy(st_list *plist){
    store *pnode = plist->head.pfirst;
    if(pnode){
         do{
            LST_REMOVE(&plist->head, pnode, st_n)
            free(pnode);
            pnode = plist->head.pfirst;
         }while(pnode);
    }
    plist->head.pfisrt = plist->head.plast = NULL;
}

>>...This violates the implementation namespace ...
I do not violate nothing since this is pure C code...
there are no namespaces in C...
and this is C Programming forum...
(uhmms forgot to remove #if defined(__cplusplus) lines in "_list.h" again take a NOTE at the C .h extension
not <list>)


>>I'm also curious as to why you chose to implement this as a
>>series of macros.

because i think this is the best solution for
this type of linked list in pure C.
This is just an intermediate code

one should wrap those macroses (which are very general)
in functions like this:

Code:

typedef struct _strnode
{
    .....
}strnode;

int strlist_insert_node(strlist *plist, strnode *pnode){
  int ret=0;
  LST_INSERT(....)
  /* do other init stuff*/
  return ret;
}

Also this is very close to the idea
found in the standart Linux "list.h" implementation

mine (the "_list.h" ) is a variation of Linux "list.h"

>>future updates

future updates of WHAT..
be more specific

Aaw...!

Now i get it !

there something in my code that is not ISO style standart compliant!!!!

Well guess what ...
I have my own style...
and i like it that way with leading '_'
so i know the diff between:

#define MAGIC_1 0x1234
and :
#define _SOME_CONTROL_DEF

and since C/C++ syntaxis doesn't dissallow me to write :
#define _VALID_C_SYNTAX
i will use it like that.

i am sorry if you doesn't like the layout of the header file
everyone is permited change it without asking the author
Use it for comercial purposes if any of you want to do this
I don't care... This is (C)ommonly (U)sed (S)nippet in my
programs...
End of story

>>..if you doesn't like...
UPS!
pardon my english

if you *don't* like...

>>...Because void main is wrong...

Now this is totaly different thing :

void main it is not wrong
void main is just undefined behavior because
when the program returns the flow to the OS
it leaves the EAX unchanged ... and if the OS is expecting value from EAX something BAD could(or could not) happen

leading underscore WILL never cause such thing
(except for some fancy new compiler wich automaticly updates some software using '_' but can't think of any so this is "don't really care" scenario)
And it isn't wrong

>>...Then I hope you don't (W)ork (F)or a (N)uclear (P)ower (P)lant
The Nuclear Plant will never meltdown because of this...

AND in case of maintaining:


I prefer this :

Code:

#ifndef _NUCLEAR_CONTROL_V_1_1_1
#define  _NUCLEAR_CONTROL_V_1_1_1
....
#endif /*_NUCLEAR_CONTROL_V_1_1_1*/

#ifndef _NUCLEAR_CONTROL_V_1_1_2
#define  _NUCLEAR_CONTROL_V_1_1_2
....
#endif /*_NUCLEAR_CONTROL_V_1_1_2*/

than :

Code:

#ifndef NUCLEAR_CONTROL
#define  NUCLEAR_CONTROL
....
#endif /NUCLEAR_CONTROL*/

#ifndef _NUCLEAR_CONTROL
#define  _NUCLEAR_CONTROL
....
#endif /*_NUCLEAR_CONTROL*/