Parsing in C

[Albert Holguin]

Did you try sprintf()? http://www.cplusplus.com/reference/clibrary/cstdio/sprintf/[^]

[Chris Losinger]

so, is it "n(nodename)" that signals the end of a 'node' and "d(whatever)" for the start? i always find state-machines the easiest way to do simple parsing. off the top of my head...

typedef struct item_t
{
char d[100];
char b[100];
char n[100];
} item;

const char *consume_between_parens(const char *p, char *out)
{
if (*p!='(') return NULL; // error

p++;
while (*p)
{
if (p==')') // done
{
p++;
break;
}
*out = *p; // copy
out++; p++; // next
}

return p;
}

void parse(const char *p, itemArray array ... some array thing)
{
enum {wantD, wantB, wantN} state = wantD;

item * curItem = null;

while (*p)
{
switch (state)
{
case wantD: // we need a 'd'
if (*p=='d')
{
curItem = malloc(sizeof(item));
p = consume_between_parens(p, curItem->d);
state = wantB;
}
else error
break;
case wantB: // we need a 'b'
if (*p=='b')
{
p = consume_between_parens(p, curItem->b);
state = wantN;
}
else error
case wantN: // we need an 'n'
if (*p=='n')
{
p = consume_between_parens(p, curItem->n);
addToArray(array, curItem);
state = wantD;
}
else error
break;
default:
error;
}
p++;
}
while

image processing toolkits | batch image processing

[David Crow]

Have you looked at strtok()?

"One man's wage rise is another man's price increase." - Harold Wilson

"Fireproof doesn't mean the fire will never come. It means when the fire comes that you will be able to withstand it." - Michael Simmons

"Show me a community that obeys the Ten Commandments and I'll show you a less crowded prison system." - Anonymous

[Software2007]

Fantastic! I did get it to work with very minor modifications.

typedef struct item_t
{
char d[8192];
char n[8192];
char i[8192];
} item;

char *consume_between_parens(char *p, char *out)
{
p++; //advance to where we think there is '('

if (*p!='(') return NULL; // error

int paren_count = 1;

p++;
while (*p)
{
if (*p=='(')
paren_count++;

   if (\*p==')') 
      paren\_count--;

   if(!paren\_count) //done
	   break;

   \*out = \*p; // copy
   out++; p++; // next

}

return p;
}

void parse(char *p)
{

item * curItem = NULL;

while (*p)
{

        if (\*p=='d')
        {
            curItem = (item\*)malloc(sizeof(item));
            p = consume\_between\_parens(p, curItem->d);
        }
        if (\*p=='n')
        {
            p = consume\_between\_parens(p, curItem->n);
        }
        if (\*p=='i')
        {
            p = consume\_between\_parens(p, curItem->i);
           // addToArray(array, curItem);
        }

 p++;

}

}

int _tmain(int argc, _TCHAR* argv[])
{
char *text ="1:7\22\\d(2011)n(0)i(711910)d(2010)n(1)i(711911)";

parse(text);

return 0;

}

[Software2007]

I just did. Thanks for mentioning it, this function actually makes things a lot easier, since it does most of the pointer work for you. I will give it a try. Thanks again

[Software2007]

I have one more question C- related:

char *getLine(char *p, char *out)
{
int Dees = 0;
char *px = p;
p++; //advance to where we think there is '('

if (*p!='(') return NULL; // error

int paren_count = 1;

while (*px)
{
if(*px=='d')
{
Dees++;
if(Dees >1 && *(px+1)=='(') //Done at 2nd 'd'
break;
}

   \*out = \*px; // copy
   out++; px++; // next

}
return px;
}

//main program
char *px="d(text)n(0)l(1)d(text2)n(1)l(2)...";
char line[8192];
char *py = getLine(px,line);

In the main program, after calling the function getLine(), I do get line correctly as in "d(text)n(0)l(1)" but filled with garbage after that as its trying to fill out 8192 characters. How do I make it not fill the rest of the unused space. Note, I don't know the size of "line" as it varies, but I need it big enough. I know strcpy could be useful here somewhere, just can't get it to work. I also tried in getLine() the following:

char field[4095];
strcpy(field,out);

But nothing gets copied, out shows null in the debugger. Thanks

[David Crow]

After the while() loop, you need to terminate out with a '\0' character.

"One man's wage rise is another man's price increase." - Harold Wilson

"Fireproof doesn't mean the fire will never come. It means when the fire comes that you will be able to withstand it." - Michael Simmons

"Show me a community that obeys the Ten Commandments and I'll show you a less crowded prison system." - Anonymous

[Software2007]

Fabulous!

while (*px != '\0')

did the trick. I have to admit i don't quite get it though, In the string "d(2011)n(0)i(711910)d(2010)n(1)i(711911)", I am telling to break out when it sees the next 'd(', so why do I need the terminating character? Could you clarify? Thanks much

[David Crow]

Software2007 wrote:

while (*px != '\0')

did the trick.

Not sure how, since these three staments are identical:

while (*px)
while (*px != 0)
while (*px != '\0')

What you should have done instead is add a '\0' character to out after the while() loop. That way, it'll be properly (i.e., null) terminated.

"One man's wage rise is another man's price increase." - Harold Wilson

"Fireproof doesn't mean the fire will never come. It means when the fire comes that you will be able to withstand it." - Michael Simmons

"Show me a community that obeys the Ten Commandments and I'll show you a less crowded prison system." - Anonymous

[Software2007]

Great, makes sense.