«

Hi All

Here is a program which basically tokenizes a string based on space
seperation.
But it does not run bcoz i am directly trying to manuplate the string.
I cannot use standard string functions due to performance issues

can anybody figure out whats gone wrong with this



#include<stdio.h>
#include<string.h>
#include<malloc.h>


#define SP ' '


char **tokenize(char *);


char ** tokenize(char *ch)// to tokenize the string based on space
{


int count = 0;
char *mrk = ch;


while(*ch != '\0')
{
if(*ch == SP)


count++;


ch++;


}


count = count + 2;


ch = mrk;


char **c = (char **)malloc(count * sizeof(ch));


int count2;
char *fst,*lst;


fst = lst = ch;


for (count2 = 0; count2 < count - 1; count2++)
{


lst = strchr(lst,SP);


*lst = '\0'; //the DDD crashes at this point


lst++;


c[count2] = fst;


fst = lst;
}


c[count - 1] = NULL;


return c;



}


int main()
{

char *s = "hello how do u do";


char **t = tokenize(s);


printf("the str is %s\n",*t[0]);//similarly for s[1] & s[2]



}

shyam wrote:
> Hi All
>
> Here is a program which basically tokenizes a string based on space
> seperation.
> But it does not run bcoz i am directly trying to manuplate the string.
> I cannot use standard string functions due to performance issues
>
> can anybody figure out whats gone wrong with this
>
>
>
> #include<stdio.h>
> #include<string.h>
> #include<malloc.h>
Non-standard header. malloc() is declared in:
#include <stdlib.h>
>
>
> #define SP ' '
>
>
> char **tokenize(char *);
>
>
> char ** tokenize(char *ch)// to tokenize the string based on space
> {
>
>
> int count = 0;
> char *mrk = ch;
>
>
> while(*ch != '\0')
> {
> if(*ch == SP)
>
>
> count++;
>
>
> ch++;
>
>
> }
>
>
> count = count + 2;
>
>
> ch = mrk;
>
>
> char **c = (char **)malloc(count * sizeof(ch));
There's no need to cast the return value of malloc(), doing so can mask
errors. Better:
char **c = malloc(count * sizeof *c);
>
>
> int count2;
> char *fst,*lst;
>
>
> fst = lst = ch;
>
>
> for (count2 = 0; count2 < count - 1; count2++)
> {
>
>
> lst = strchr(lst,SP);
>
>
> *lst = '\0'; //the DDD crashes at this point
Right. `lst' points somewhere into the array of chars passed to the
function. In this case (see main() below) it is a string literal which
is not guaranteed to be modifiable.
>
>
> lst++;
>
>
> c[count2] = fst;
>
>
> fst = lst;
> }
>
>
> c[count - 1] = NULL;
>
>
> return c;
>
>
>
> }
>
>
> int main()
> {
>
> char *s = "hello how do u do"
Try:
char s[] = "hello how do u do";

[BTW, you had a missing semicolon above. Cut and paste *real* code in
the future.]
;
>
>
> char **t = tokenize(s);
>
>
> printf("the str is %s\n",*t[0]);//similarly for s[1] & s[2]
>
>
>
> }
>
Next time:
1) Post *real* code.
2) Get rid of all those absolutely useless blank lines.

Also, be aware that there may be (and most likely are) further bugs in
your code; there's just no way I'm going to go through it with a fine
toothed comb when presented like this.

HTH,
--ag

--
Artie Gold -- Austin, Texas
http://goldsays.blogspot.com (new post 8/5)
http://www.cafepress.com/goldsays
"If you have nothing to hide, you're not trying!"

On 28 Nov 2005 18:02:46 -0800, "shyam" <> wrote:

>Hi All
>
>Here is a program which basically tokenizes a string based on space
>seperation.
>But it does not run bcoz i am directly trying to manuplate the string.
>I cannot use standard string functions due to performance issues
>
>can anybody figure out whats gone wrong with this
>
>
>
>#include<stdio.h>
>#include<string.h>
>#include<malloc.h>

Non-standard header. malloc and family are in stdio.h.

>
>
>#define SP ' '
>
>
>char **tokenize(char *);
>
>
>char ** tokenize(char *ch)// to tokenize the string based on space

Why are you triple spacing between every line? Does your monitor
really show 300 lines or is there some reason you don't like to see as
much of your code as possible? However, you do get points for
indenting nicely.

>{
> int count = 0;
> char *mrk = ch;
> while(*ch != '\0')
> {
> if(*ch == SP)
> count++;
> ch++;
> }
> count = count + 2;
> ch = mrk;
> char **c = (char **)malloc(count * sizeof(ch));

Many pre-C99 systems do not tolerate definitions after executable
code.

Don't cast the return from malloc. It seldom helps and cause the
compiler to suppress a warning you really want to see.

> int count2;
> char *fst,*lst;
> fst = lst = ch;
> for (count2 = 0; count2 < count - 1; count2++)
> {
> lst = strchr(lst,SP);
> *lst = '\0'; //the DDD crashes at this point

If you look in main, you will see that the argument you passed to this
function is a pointer to a string literal. This statement attempts to
change the contents of that literal. This is undefined behavior. An
immediate crash is one of the best types of undefined behavior.

> lst++;
> c[count2] = fst;
> fst = lst;
> }
> c[count - 1] = NULL;
> return c;
>}
>
>int main()
>{
> char *s = "hello how do u do";
> char **t = tokenize(s);
> printf("the str is %s\n",*t[0]);//similarly for s[1] & s[2]

The %s requires a char* for its argument. t is a char**; t[0] is a
char*; *t[0] is a char. I think you wanted t[0].

Don't you think your should print t[1], t[2], until you get to the
t[n] which is NULL.

>}


<<Remove the del for email>>

On 2005-11-29, Barry Schwarz <> wrote:
> Non-standard header. malloc and family are in stdio.h.

they're in what, now?

Jordan Abel wrote:

> On 2005-11-29, Barry Schwarz <> wrote:
> > Non-standard header. malloc and family are in stdio.h.
>
> they're in what, now?

The standard header, stdio.h, as he said. What's confusing you?



Brian

Default User:
> Jordan Abel:
>
>> On 2005-11-29, Barry Schwarz <> wrote:
>> > Non-standard header. malloc and family are in stdio.h.
>>
>> they're in what, now?
>
> The standard header, stdio.h, as he said. What's confusing you?

Only if merely the first three letters are significant.

Jirka

Default User wrote:
> Jordan Abel wrote:
>>On 2005-11-29, Barry Schwarz <> wrote:

>>>Non-standard header. malloc and family are in stdio.h.

>>they're in what, now?

> The standard header, stdio.h, as he said. What's confusing you?


Are you sure about that, Brian? What's <stdlib.h> for, then?

Default User wrote:
> Jordan Abel wrote:
>
>> On 2005-11-29, Barry Schwarz <> wrote:
>>> Non-standard header. malloc and family are in stdio.h.
>> they're in what, now?
>
> The standard header, stdio.h, as he said. What's confusing you?

The fact that malloc and friends are declared in stdlib.h and not
stdio.h possibly?
--
Flash Gordon
Living in interesting times.
Although my email address says spam, it is real and I read it.

On 2005-11-29, Default User <> wrote:
> Jordan Abel wrote:
>
>> On 2005-11-29, Barry Schwarz <> wrote:
>> > Non-standard header. malloc and family are in stdio.h.
>>
>> they're in what, now?
>
> The standard header, stdio.h, as he said. What's confusing you?

the fact that thy're actually in stdlib.h, and, on my system at least,
including stdio.h does not cause declarations from stdlib.h to become
available.

Jordan Abel wrote:

> On 2005-11-29, Default User <> wrote:
> > Jordan Abel wrote:

> the fact that thy're actually in stdlib.h, and, on my system at least,
> including stdio.h does not cause declarations from stdlib.h to become
> available.

Damn it. Sorry.


Brian