my get_line function

[qubit67]

this is my get_line function:

#define BUFFER_SIZE 1

char *get_line(FILE *file) 
{
	FILE *fp = file;
	
	char *buffer;
	int ch,  i=0;
	unsigned int buffer_size = BUFFER_SIZE, new_buffer_size;

	assert(buffer_size > 0);

	buffer = (char *) malloc (buffer_size * sizeof (char));
	
	while ((ch = getc(fp)) != EOF && ch != '\n')
	{
		if (buffer_size == i) 
		{
			new_buffer_size = buffer_size * 2;
			assert(new_buffer_size >= buffer_size);
			buffer = (char *) realloc (buffer, new_buffer_size * sizeof (char));
			buffer_size = new_buffer_size;
		}
		buffer[i] = ch;
		i++;
	}
	if (ch == '\n')			 
	{
		buffer[i++] = '\0';
		return buffer; 		/*finish it off with a null byte*/ 
	}
	if ((ch == EOF) && (i > 1))
	{
		buffer [i++] = '\0';
		return buffer;
	}
	else
	{
		return NULL;
	}
}

my questions are; does the i++ in the buffer need to be done or does the while loop update i when EOF or '\n' is encountered, and if the first input is the EOF, will NULL be returned as in the last statement?

[qubit67]

im sorry, what do you mean?

[robwhit]

sorry I just deleted it. I read it wrong. my fault.

[whiteflags]

> my questions are; does the i++ in the buffer need to be done

No. At this point in the code i will not be used again and shouldn't need to be updated at all. As long as i is the biggest index in the array then you should place a '\0' correctly.

buffer = (char *) realloc (buffer, new_buffer_size * sizeof (char));

...
 
else
{
   return NULL;
}

These are some leaky and troublesome areas. You haven't thought about the need to free buffer if EOF is the first thing you see, or if realloc breaks so I recommend that you think about it now. Search the forum to see previous attempts at this idea.

Have you considered reading more than one character from the start?

[robwhit]

does the i++ in the buffer need to be done

you mean in this?

buffer [i++] = '\0';

no, it doesn't do anything since it's the post-increment operator and the i value isn't used any more.

does the while loop update i when EOF or '\n' is encountered

no. i is in the body, and if the loop condition fails, the loop body doesn't run.

if the first input is the EOF, will NULL be returned as in the last statement?

yes.

[qubit67]

thanks for you input guys, will post the changed code when im finished

[ssharish2005]

Why dont u use just fgets to read a file?

ssharish2005

[twomers]

With very little difficulty you could do something that the C++ version does and allow variation in the terminating character:

char *get_line( FILE *f, char ch='\n' ) {
  /* ... */
}

and change all the '\n' 's to ch.

[qubit67]

Why dont u use just fgets to read a file?

ssharish2005

im not sure how to write a dynamic wrapper for it.

[brewbuck]

Why dont u use just fgets to read a file?

fgets() takes a maximum length. The whole point of this get_line() function is to have no restriction on the length.

[ssharish2005]

fgets() takes a maximum length. The whole point of this get_line() function is to have no restriction on the length.

My bad never read the whole post

ssharish2005

[qubit67]

so, I use getc and realloc with a dynamic buffer so the string has no size limitations. so far I have:

char *get_line(FILE *file) 
{
	FILE *fp = file;
	
	char *buffer;
	int ch,  i=0;
	unsigned int buffer_size = BUFFER_SIZE, new_buffer_size;

	assert(buffer_size > 0);

	buffer = (char *) safe_malloc (buffer_size * sizeof (char));
	
	while ((ch = getc(fp)) != EOF && ch != '\n')
	{
		if (buffer_size == i) 
		{
			new_buffer_size = buffer_size * 2;
			assert(new_buffer_size >= buffer_size);
			buffer = (char *) safe_realloc (buffer, new_buffer_size * sizeof (char));
			buffer_size = new_buffer_size;
		}
		buffer[i] = ch;
		i++;
	}
	if (ch == '\n')			 
	{
		buffer[i] = '\0';
		return buffer; 		/*finish it off with a null byte*/ 
	}
	if ((ch == EOF) && (i >= 1))
	{
		buffer[i] = '\0';
		return buffer;
	}
	else
	{
                                free (buffer);
		return NULL;
	}
}

There requirements of the function are :

get_line reads a single line of text from its argument file pointer. If the EOF is encountered
immediately without reading any characters from the file, the NULL pointer should be
returned. Otherwise a pointer to a dynamically allocated string should be returned. The
string must be large enough to fit an entire line of input. Lines are ended by the newline
character, or when the EOF is encountered. Ending newline characters should be read in by
the function, but they should NOT be stored in the string which is returned.

My questions are:

1. does my function fullful the requirements?

2. is the guard of the while loop appropriate, or should I move the conditions : ch == EOF and ch == '\n' to the body of the loop with a break to end the loop then assign a null?

3. In actuality, is this function correct, could it be modified to be better than it is?

[matsp]

I would combine the end-criteria of EOF and newline - as you are doing exactly the same thing in two if-statements. This is generally not a good idea (because you'll be forgetting to change one side when you add something in the other, and then get strange behaviour sometimes).

And your comment, which is a bit superfluous, about finishing off with a NUL-byte is on the wrong line.

--
Mats

[Salem]

Also here
http://cboard.cprogramming.com/showthread.php?t=93518
One of them closed.