How to correct the code segment???????? ?????

[biplab]

for(i=0;i<256;i ++)
{
hist[i]=0;
}
for(i=0;i<heigh t;i++)
{
for(j=0;j<width ;j++)
{
k=(long int)rgb1[i][j];
hist[k]=hist[k]+1;
}
}

for(i=0;i<256;i ++)
{
if(hist[i]!=0)
tmp++;
else continue;
}

sum_hist=(int** )malloc(tmp*siz eof(int*));
for(i=0;i<tmp;i ++)
sum_hist[i]=(int*)malloc(2 *sizeof(int));

for(i=0;i<256;i ++)
{
if(hist[i]!=0)
{
//printf("%d %d\n",i,hist[i]);
sum_hist[i][0]=(int)i;
sum_hist[i][1]=(int)hist[i];
printf("%d %d\n",sum_hist[i][0],sum_hist[i][1]);

}
else continue;
}
for(i=0;i<tmp;i ++)
printf("%d\n",s um_hist[i][1]);

In the above code segment sum_hist is a double pointer of type long
int...the problem is that when the values of sum_hist is printed withn
the is block..it is giving correct result..but when outside the for
loop..the value is printed......wr ong values are printed.....how to
correct the code...

[Richard]

Keith Thompson <ks***@mib.orgw rites:

Richard<rg****@ gmail.comwrites :

>Andrew Poelstra <ap*******@supe rnova.homewrite s:

[...]

>>Here is your code, corrected for the (mainly stylistic) issues I
mentioned else thread. It is abundantly clear to me now what the
problem is. I hope it is also clear to you:

/* Begin uncompilable segment */

memset(hist , 0, 256);

Horrible. Code fails review.

Did you have something in mind other than the use of the magic number
256?

Nope.

>

>>for(i = 0; i < height; i++)
for(j = 0; j < width; j++)
{
k = rgb1[i][j];
if(hist[k] == 0)
++count;
++hist[k];
}

sum_hist = malloc(count * sizeof *sum_hist);
for(i = 0; i < count; i++)
sum_hist[i] = malloc(2 * sizeof *sum_hist[i]);

Silly to use sizeof on a variable element. Overly clever and liable to
mislead. It also assumes optimization will spot the obvious.

It's simply an instance of the usual clc idiom for calling malloc:

I know what it is. And what form of it it is.

ptr = malloc(count * sizeof *ptr);
with "sum_hist[i]" replacing "ptr". It won't mislead anyone who's
familiar with the idiom. As for assuming that "optimizati on will spot
the obvious", I'm afraid I have no idea what you mean; would you care
to explain? (If you're referring to not evaluating the operand of
sizeof, that's not an optimization, it's a language requirement.)

[snip]

--

[Barry Schwarz]

On Sat, 13 Sep 2008 14:45:49 GMT, Andrew Poelstra
<ap*******@supe rnova.homewrote :

>On 2008-09-13, biplab <ge*******@gmai l.comwrote:

>for(i=0;i<256; i++)
{
hist[i]=0;
}

You need to fix your formatting. 2-space indents are
recommended for posting on Usenet. In any case, be
consistent, and use whitespace around operators.

Also, this could be replaced with:
memset(hist, 0, 256);

Only if sizeof hist[i] is 1 which is unlikely for anything other than
a char.

>
Or potentially:
memset(hist, 0, sizeof hist);

Why was the mistake prone recommendation above a definite and this
which looks correct a potential?
--
Remove del for email

[Barry Schwarz]

On Sat, 13 Sep 2008 07:17:08 -0700 (PDT), biplab <ge*******@gmai l.com>
wrote:

>for(i=0;i<256; i++)
{
hist[i]=0;
}
for(i=0;i<heigh t;i++)
{
for(j=0;j<width ;j++)
{
k=(long int)rgb1[i][j];
hist[k]=hist[k]+1;
}
}

for(i=0;i<256;i ++)
{
if(hist[i]!=0)
tmp++;
else continue;
}

sum_hist=(int** )malloc(tmp*siz eof(int*));

Since you claim below that sum_hist is a long**, this may not allocate
the correct amount of space if sizeof(int*) < sizeof(long*).

In any event, this code requires a diagnostic because there is no
implicit converstion between int** and long**.

> for(i=0;i<tmp;i ++)
sum_hist[i]=(int*)malloc(2 *sizeof(int));

This code is even more likely to fail since sizeof(long) is more
frequently greater than sizeof(int). The same diagnostic is required
regarding the implicit pointer conversion.

>
for(i=0;i<256;i ++)
{
if(hist[i]!=0)
{
//printf("%d %d\n",i,hist[i]);
sum_hist[i][0]=(int)i;
sum_hist[i][1]=(int)hist[i];
printf("%d %d\n",sum_hist[i][0],sum_hist[i][1]);

%d requires an int. You are passing a long. This invokes undefined
behavior.

>
}
else continue;
}
for(i=0;i<tmp;i ++)
printf("%d\n",s um_hist[i][1]);

In the above code segment sum_hist is a double pointer of type long
int...the problem is that when the values of sum_hist is printed withn
the is block..it is giving correct result..but when outside the for
loop..the value is printed......wr ong values are printed.....how to
correct the code...

Post a compilable example that demonstrated the undesired behavior.

--
Remove del for email

[Andrew Poelstra]

On 2008-09-14, Barry Schwarz <sc******@dqel. comwrote:

On Sat, 13 Sep 2008 14:45:49 GMT, Andrew Poelstra
<ap*******@sup ernova.homewrot e:

>>On 2008-09-13, biplab <ge*******@gmai l.comwrote:

>>for(i=0;i<256 ;i++)
{
hist[i]=0;
}

You need to fix your formatting. 2-space indents are
recommended for posting on Usenet. In any case, be
consistent, and use whitespace around operators.

Also, this could be replaced with:
memset(hist, 0, 256);

Only if sizeof hist[i] is 1 which is unlikely for anything other than
a char.

Of course. My bad.

>>
Or potentially:
memset(hist, 0, sizeof hist);

Why was the mistake prone recommendation above a definite and this
which looks correct a potential?

The original code did not declare any variables (hence my
mistake above) and I was unsure if hist was an array or a
pointer.

--
Andrew Poelstra ap*******@wpsof tware.com
To email me, use the above email addresss with .com set to .net