double to int weirdness

[nsa.usa]

Hi,
Converting a double to int by doing h=(int)(j)in the below code
doesn't give the expected result. I'm trying to create 32 random
numbers from 1 to 36. So as you see in the output, I get f.ex. in the
first line j=10.000000 but converted to int is equals h=1103175862
???? what gives? please help, I have a headache now, I read the faq it
doesn't help :-(
Cheers and thanks!,
Tobias

/* SESSIONID_LENGTH = 32 */
int i,t,h;
double j,k,l;

for(t=0; t<SESSIONID_LENGTH; ++t) {
k = rand();
l = RAND_MAX;
j = ( k / l ) * 36;
j = abs(j)+1;
h = (int)(j);
#ifdef DEBUG
printf("t=%d j=%f k=%d h=%d",t,j,k,h);
printf("\n");
#endif
}

Outputs:
t=0 j=10.000000 k=-2113929216 h=1103175862
t=1 j=33.000000 k=-1807745024 h=1105006335
t=2 j=8.000000 k=1358954496 h=1102651077
t=3 j=5.000000 k=-1560281088 h=1102130931
t=4 j=23.000000 k=1132462080 h=1104436416
t=5 j=36.000000 k=-444596224 h=1105187835
t=6 j=15.000000 k=746586112 h=1103685495
t=7 j=26.000000 k=1702887424 h=1104606873
t=8 j=18.000000 k=1308622848 h=1104051639
t=9 j=17.000000 k=1602224128 h=1104022361
t=10 j=15.000000 k=-452984832 h=1103710345
t=11 j=1.000000 k=1610612736 h=1097455182
t=12 j=2.000000 k=0 h=1100513581
t=13 j=13.000000 k=1593835520 h=1103471617
t=14 j=36.000000 k=209715200 h=1105179076
t=15 j=29.000000 k=1006632960 h=1104771106
t=16 j=16.000000 k=-1795162112 h=1103811038
t=17 j=30.000000 k=692060160 h=1104803125
t=18 j=8.000000 k=2030043136 h=1102713577
t=19 j=22.000000 k=-29360128 h=1104351131
t=20 j=10.000000 k=1526726656 h=1103155560
t=21 j=25.000000 k=910163968 h=1104551646
t=22 j=7.000000 k=-469762048 h=1102595806
t=23 j=22.000000 k=1937768448 h=1104341221
t=24 j=22.000000 k=-1891631104 h=1104326173
t=25 j=9.000000 k=-1543503872 h=1102952931
t=26 j=27.000000 k=1761607680 h=1104646387
t=27 j=19.000000 k=1140850688 h=1104153565
t=28 j=17.000000 k=-612368384 h=1103948406
t=29 j=35.000000 k=-1153433600 h=1105106370
t=30 j=19.000000 k=1933574144 h=1104187859
t=31 j=26.000000 k=-1363148800 h=1104610710

[regis]

ns*****@gmail.com wrote:

Hi,
Converting a double to int by doing h=(int)(j)in the below code
doesn't give the expected result. I'm trying to create 32 random
numbers from 1 to 36. So as you see in the output, I get f.ex. in the
first line j=10.000000 but converted to int is equals h=1103175862
???? what gives? please help, I have a headache now, I read the faq it
doesn't help :-(
Cheers and thanks!,
Tobias

/* SESSIONID_LENGTH = 32 */
int i,t,h;
double j,k,l;

for(t=0; t<SESSIONID_LENGTH; ++t) {
k = rand();
l = RAND_MAX;
j = ( k / l ) * 36;
j = abs(j)+1;
h = (int)(j);

#ifdef DEBUG
printf("t=%d j=%f k=%d h=%d",t,j,k,h);
printf("\n");
#endif
}

k and h are double, so either cast them to int in printf:

printf("t=%d j=%f k=%d h=%d",t,j,(int)k,(int)h);

or change the last two %d to %f:

printf("t=%d j=%f k=%f h=%f",t,j,k,h);

[jacob navia]

ns*****@gmail.com wrote:

Hi,
Converting a double to int by doing h=(int)(j)in the below code
doesn't give the expected result. I'm trying to create 32 random
numbers from 1 to 36. So as you see in the output, I get f.ex. in the
first line j=10.000000 but converted to int is equals h=1103175862
???? what gives? please help, I have a headache now, I read the faq it
doesn't help :-(
Cheers and thanks!,
Tobias

/* SESSIONID_LENGTH = 32 */
int i,t,h;
double j,k,l;

for(t=0; t<SESSIONID_LENGTH; ++t) {
k = rand();
l = RAND_MAX;
j = ( k / l ) * 36;
j = abs(j)+1;
h = (int)(j);
#ifdef DEBUG
printf("t=%d j=%f k=%d h=%d",t,j,k,h);
printf("\n");
#endif
}

When I compile under lcc-win32 I obtain:
d:\lcc\mc68\test>lc trand.c
Warning trand.c: 18 printf argument mismatch for format d. Expected int
got double
0 errors, 1 warning

Change the printf format and recompile.

jacob

[Tim Prince]

ns*****@gmail.com wrote:

Hi,
Converting a double to int by doing h=(int)(j)in the below code
doesn't give the expected result. I'm trying to create 32 random
numbers from 1 to 36. So as you see in the output, I get f.ex. in the
first line j=10.000000 but converted to int is equals h=1103175862
???? what gives? please help, I have a headache now, I read the faq it
doesn't help :-(
Cheers and thanks!,
Tobias

/* SESSIONID_LENGTH = 32 */
int i,t,h;
double j,k,l;

for(t=0; t<SESSIONID_LENGTH; ++t) {
k = rand();
l = RAND_MAX;
j = ( k / l ) * 36;
j = abs(j)+1;
h = (int)(j);
#ifdef DEBUG
printf("t=%d j=%f k=%d h=%d",t,j,k,h);
printf("\n");
#endif
}

Outputs:
t=0 j=10.000000 k=-2113929216 h=1103175862
t=1 j=33.000000 k=-1807745024 h=1105006335
t=2 j=8.000000 k=1358954496 h=1102651077
t=3 j=5.000000 k=-1560281088 h=1102130931
t=4 j=23.000000 k=1132462080 h=1104436416
t=5 j=36.000000 k=-444596224 h=1105187835
t=6 j=15.000000 k=746586112 h=1103685495
t=7 j=26.000000 k=1702887424 h=1104606873
t=8 j=18.000000 k=1308622848 h=1104051639
t=9 j=17.000000 k=1602224128 h=1104022361
t=10 j=15.000000 k=-452984832 h=1103710345
t=11 j=1.000000 k=1610612736 h=1097455182
t=12 j=2.000000 k=0 h=1100513581
t=13 j=13.000000 k=1593835520 h=1103471617
t=14 j=36.000000 k=209715200 h=1105179076
t=15 j=29.000000 k=1006632960 h=1104771106
t=16 j=16.000000 k=-1795162112 h=1103811038
t=17 j=30.000000 k=692060160 h=1104803125
t=18 j=8.000000 k=2030043136 h=1102713577
t=19 j=22.000000 k=-29360128 h=1104351131
t=20 j=10.000000 k=1526726656 h=1103155560
t=21 j=25.000000 k=910163968 h=1104551646
t=22 j=7.000000 k=-469762048 h=1102595806
t=23 j=22.000000 k=1937768448 h=1104341221
t=24 j=22.000000 k=-1891631104 h=1104326173
t=25 j=9.000000 k=-1543503872 h=1102952931
t=26 j=27.000000 k=1761607680 h=1104646387
t=27 j=19.000000 k=1140850688 h=1104153565
t=28 j=17.000000 k=-612368384 h=1103948406
t=29 j=35.000000 k=-1153433600 h=1105106370
t=30 j=19.000000 k=1933574144 h=1104187859
t=31 j=26.000000 k=-1363148800 h=1104610710

Who is to guess what you are doing, if you don't post compilable C code?
For example, if you were writing C, you would not only make a function
of this, and include standard headers, but you would use printf formats
which agree with the data types. Any C compiler would have at least
optional diagnostics, which you should read. If you used printf() in a
normal way, you could not get the results you show. There must be
several relevant topics in the FAQ, but you don't reveal which you
considered.

[William Hughes]

On Jul 28, 5:45 pm, regis <[e-mail]:

[e-mail] wrote:

Hi,
Converting a double to int by doing h=(int)(j)in the below code
doesn't give the expected result. I'm trying to create 32 random
numbers from 1 to 36. So as you see in the output, I get f.ex. in the
first line j=10.000000 but converted to int is equals h=1103175862
???? what gives? please help, I have a headache now, I read the faq it
doesn't help :-(
Cheers and thanks!,
Tobias

/* SESSIONID_LENGTH = 32 */
int i,t,h;
double j,k,l;

for(t=0; t<SESSIONID_LENGTH; ++t) {
k = rand();
l = RAND_MAX;
j = ( k / l ) * 36;
j = abs(j)+1;
h = (int)(j);

#ifdef DEBUG
printf("t=%d j=%f k=%d h=%d",t,j,k,h);
printf("\n");
#endif
}

k and h are double, so either cast them to int in printf:

printf("t=%d j=%f k=%d h=%d",t,j,(int)k,(int)h);

or change the last two %d to %f:

printf("t=%d j=%f k=%f h=%f",t,j,k,h);

note k is a double, but h is an int. The original format is wrong
for k, but right for h. However, it looks like
both k and h are not being correctly output.
Since the format specifier for k is wrong, it makes
sense that k is not printing correctly, but why is
h not printing correctly.

In general there is no way to know. Once you get
one format specifier wrong you invoke undefined behaviour and
anything can happen.

At a guess the arguments are put in a stack byte by byte
the number of bytes to put is determined by the type
of the argument and the number of bytes to use is determined by the
format
specifier. In this case k is a double, so 8 bytes
are put on the array, however the format specifier is int
so only four bytes are used when k is printed. When h is printed
you get the next four bytes of k, rather than the four bytes
of h. Of course this is only a guess, something else entirely
may be happening, and the 8 and 4 bytes, while common, are hardly
universal. Still, we have one plausible scenario

Try fixing the format specifier and recompiling.
(Worked for me)

Your compiler may be able to warn you about this type of problem.
With gcc 3.2.2 using gcc - Wall gives the warning

int format, double arg (arg 4)

- William Hughes

[regis]

William Hughes wrote:

On Jul 28, 5:45 pm, regis <[e-mail]:

>>[e-mail] wrote:

>>>Converting a double to int by doing h=(int)(j)in the below code
doesn't give the expected result. I'm trying to create 32 random
numbers from 1 to 36. So as you see in the output, I get f.ex. in the
first line j=10.000000 but converted to int is equals h=1103175862
???? what gives? please help, I have a headache now, I read the faq it
doesn't help :-(
Cheers and thanks!,
Tobias

>> /* SESSIONID_LENGTH = 32 */
int i,t,h;
double j,k,l;

>> for(t=0; t<SESSIONID_LENGTH; ++t) {
k = rand();
l = RAND_MAX;
j = ( k / l ) * 36;
j = abs(j)+1;
h = (int)(j);

>> #ifdef DEBUG
printf("t=%d j=%f k=%d h=%d",t,j,k,h);
printf("\n");
#endif
}

k and h are double, so either cast them to int in printf:
printf("t=%d j=%f k=%d h=%d",t,j,(int)k,(int)h);
or change the last two %d to %f:
printf("t=%d j=%f k=%f h=%f",t,j,k,h);

note k is a double, but h is an int. The original format is wrong
for k, but right for h.

oups!

[Martin Ambuhl]

ns*****@gmail.com wrote:

Hi,
Converting a double to int by doing h=(int)(j)in the below code
doesn't give the expected result. I'm trying to create 32 random
numbers from 1 to 36. So as you see in the output, I get f.ex. in the
first line j=10.000000 but converted to int is equals h=1103175862
???? what gives? please help, I have a headache now, I read the faq it
doesn't help :-(

#include <stdio.h /* mha: added, needed for printf */
#include <stdlib.h /* mha: added, needed for rand() */
#define SESSIONID_LENGTH 32 /* mha: needed to make compilable code */
#define DEBUG /* mha: added to active printf's */

/* mha: added main, needed to have a program at all */
int main(void)
{
int t, h; /* mha: removed unused variable i */
double j, k, l;

for (t = 0; t < SESSIONID_LENGTH; ++t) {
k = rand();
l = RAND_MAX;
j = (k / l) * 36;
j = abs(j) + 1;
h = j; /* mha: Removed unneeded cast.
Unnecessary casts should not appear
in your code. */
#if 0
/* mha: the above 5 lines are better replaced with */
h = 1 + 36. * rand() / (RAND_MAX + 1.);
#endif
#ifdef DEBUG
/* mha: fixed format "%d" for k, which is a double */
printf("t=%d j=%g k=%g h=%d", t, j, k, h);
printf("\n");
#endif
}
return 0; /* mha: added. Needed for C89, a good
idea for C99 */
}

Cheers and thanks!,
Tobias

/* SESSIONID_LENGTH = 32 */
int i,t,h;
double j,k,l;

for(t=0; t<SESSIONID_LENGTH; ++t) {
k = rand();
l = RAND_MAX;
j = ( k / l ) * 36;
j = abs(j)+1;
h = (int)(j);
#ifdef DEBUG
printf("t=%d j=%f k=%d h=%d",t,j,k,h);
printf("\n");
#endif
}

Outputs:
t=0 j=10.000000 k=-2113929216 h=1103175862
t=1 j=33.000000 k=-1807745024 h=1105006335
t=2 j=8.000000 k=1358954496 h=1102651077
t=3 j=5.000000 k=-1560281088 h=1102130931
t=4 j=23.000000 k=1132462080 h=1104436416
t=5 j=36.000000 k=-444596224 h=1105187835
t=6 j=15.000000 k=746586112 h=1103685495
t=7 j=26.000000 k=1702887424 h=1104606873
t=8 j=18.000000 k=1308622848 h=1104051639
t=9 j=17.000000 k=1602224128 h=1104022361
t=10 j=15.000000 k=-452984832 h=1103710345
t=11 j=1.000000 k=1610612736 h=1097455182
t=12 j=2.000000 k=0 h=1100513581
t=13 j=13.000000 k=1593835520 h=1103471617
t=14 j=36.000000 k=209715200 h=1105179076
t=15 j=29.000000 k=1006632960 h=1104771106
t=16 j=16.000000 k=-1795162112 h=1103811038
t=17 j=30.000000 k=692060160 h=1104803125
t=18 j=8.000000 k=2030043136 h=1102713577
t=19 j=22.000000 k=-29360128 h=1104351131
t=20 j=10.000000 k=1526726656 h=1103155560
t=21 j=25.000000 k=910163968 h=1104551646
t=22 j=7.000000 k=-469762048 h=1102595806
t=23 j=22.000000 k=1937768448 h=1104341221
t=24 j=22.000000 k=-1891631104 h=1104326173
t=25 j=9.000000 k=-1543503872 h=1102952931
t=26 j=27.000000 k=1761607680 h=1104646387
t=27 j=19.000000 k=1140850688 h=1104153565
t=28 j=17.000000 k=-612368384 h=1103948406
t=29 j=35.000000 k=-1153433600 h=1105106370
t=30 j=19.000000 k=1933574144 h=1104187859
t=31 j=26.000000 k=-1363148800 h=1104610710

[Martin Ambuhl]

regis wrote:

ns*****@gmail.com wrote:

>Hi,
Converting a double to int by doing h=(int)(j)in the below code
doesn't give the expected result. I'm trying to create 32 random
numbers from 1 to 36. So as you see in the output, I get f.ex. in the
first line j=10.000000 but converted to int is equals h=1103175862
???? what gives? please help, I have a headache now, I read the faq it
doesn't help :-(
Cheers and thanks!,
Tobias

/* SESSIONID_LENGTH = 32 */
int i,t,h;
double j,k,l;

for(t=0; t<SESSIONID_LENGTH; ++t) {
k = rand();
l = RAND_MAX;
j = ( k / l ) * 36;
j = abs(j)+1;
h = (int)(j);

#ifdef DEBUG
printf("t=%d j=%f k=%d h=%d",t,j,k,h);
printf("\n");
#endif
}

k and h are double, so either cast them to int in printf:

h is not a double. Using the wrong specifier for k caused printf not to
be able to find h where it was promised to be:

>
printf("t=%d j=%f k=%d h=%d",t,j,(int)k,(int)h);

^^^
useless cast

>
or change the last two %d to %f:

printf("t=%d j=%f k=%f h=%f",t,j,k,h);

^^
incorrect specifier

[nsa.usa]

On Jul 28, 11:45 pm, regis <[e-mail]:

k and h are double, so either cast them to int in printf:

printf("t=%d j=%f k=%d h=%d",t,j,(int)k,(int)h);

or change the last two %d to %f:

printf("t=%d j=%f k=%f h=%f",t,j,k,h);

Hey thanks! h is an int so I still don't understand why i have to cast
it?? but if I cast both k and h then it works, if I cast only h then
it does not work (somehow k interferes with h?).

Anyway I got it now so thanks.
Tobias

[regis]

ns*****@gmail.com wrote:

On Jul 28, 11:45 pm, regis <[e-mail]:

>k and h are double, so either cast them to int in printf:

printf("t=%d j=%f k=%d h=%d",t,j,(int)k,(int)h);

or change the last two %d to %f:

printf("t=%d j=%f k=%f h=%f",t,j,k,h);

Hey thanks! h is an int so I still don't understand why i have to cast
it??

you don't have to.
my mistake, I saw h as a double.

but if I cast both k and h then it works, if I cast only h then
it does not work (somehow k interferes with h?).

yes, because printf is a variadic function.
If you want to understand why k interferes with h,
see Willian Hughes's post.
He answers your question, but replied to me...