this simple code snippet

char a[] ="1";
char *end;
double val;

val = strtod(a, &end);

printf("%f %d\n", val, atoi(a));

prints out
6217.000000 1

does strtod not work or am I doing something wrong?

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 4/16/21 11:20 PM, vavr...@gmail.com wrote:

this simple code snippet

char a[] ="1";
char *end;
double val;

val = strtod(a, &end);

printf("%f %d\n", val, atoi(a));

prints out
6217.000000 1

does strtod not work or am I doing something wrong?

Sitting at my Linux box, this code works as expected:

#include <stdio.h>
#include <stdlib.h>

void main()
{
char a[] ="1";
char *end;
double val;

val = strtod(a, &end);

printf("%f %d\n", val, atoi(a));
}

But I find if I omit <stdlib.h> output get a little wonky (and I have
attendant compiler warnings). Are you including that?

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

including the stdlib.h did help with the code but am still having issues with doubles in my main project even with including the stdlib. I have no real need for floating point but I am using an open source library.



On Saturday, April 17, 2021 at 10:42:31 AM UTC-4, schmidtd wrote:

On 4/16/21 11:20 PM, vavr...@gmail.com wrote:

this simple code snippet

char a[] ="1";
char *end;
double val;

val = strtod(a, &end);

printf("%f %d\n", val, atoi(a));

prints out
6217.000000 1

does strtod not work or am I doing something wrong?

Sitting at my Linux box, this code works as expected:

#include <stdio.h>
#include <stdlib.h>

void main()
{
char a[] ="1";
char *end;
double val;

val = strtod(a, &end);

printf("%f %d\n", val, atoi(a));
}
But I find if I omit <stdlib.h> output get a little wonky (and I have attendant compiler warnings). Are you including that?

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Use #pragma lint 7 (or -1) to catch missing functions and prototypes.

-------
ProLine: kelvin@pro-kegs

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Saturday, April 17, 2021 at 6:20:10 PM UTC-7, Kelvin Sherlock wrote:

Use #pragma lint 7 (or -1) to catch missing functions and prototypes.

I didn't really appreciate the value of function prototypes in C until I was trying to call malloc() on a IIgs. Without the prototype it assumed the function returned a 16-bit int instead of a 32-bit pointer, and things didn't go well when I tried to
dereference the result.

Since "double" is wider than 16 bits, every function that uses it as an argument or result needs a prototype. Turning up the warnings is a very good idea.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

I usually have lint -1 for compiles. I was running into the issue in a larger project where I did have lint -1 so I wrote quick little test program to test it out where I forgot the lint. While the little test program works the large project does not.
The following even does not work correct

int x=1;
double d =x;

d winds up being some large number. Does anyone need to know if you need to startup SANE to use doubles?



On Sunday, April 18, 2021 at 10:55:18 AM UTC-4, thef...@gmail.com wrote:

On Saturday, April 17, 2021 at 6:20:10 PM UTC-7, Kelvin Sherlock wrote:

Use #pragma lint 7 (or -1) to catch missing functions and prototypes.

I didn't really appreciate the value of function prototypes in C until I was trying to call malloc() on a IIgs. Without the prototype it assumed the function returned a 16-bit int instead of a 32-bit pointer, and things didn't go well when I tried to

dereference the result.

Since "double" is wider than 16 bits, every function that uses it as an argument or result needs a prototype. Turning up the warnings is a very good idea.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Sunday, April 18, 2021 at 11:32:54 AM UTC-7, vavr...@gmail.com wrote:

[...] The following even does not work correct

int x=1;
double d =x;

d winds up being some large number. Does anyone need to know if you need to startup SANE to use doubles?

Are you confident 'd' is being displayed correctly? i.e. does "d = 1.0" work, but "d = x" does not? Does it work correctly if you cast it to float before passing it to printf()?

Your original example assumed that strtod() was failing... are you sure the problem isn't with printf()?

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

runtime int-to-double conversion uses SANE (FI2X then FX2D).

If SANE wasn't loaded I would expect it to crash due to stack imbalance
(since SANE isn't there to clean up the stack).

In <c673a52a-20ed-4592-880b-68cba62f9302n@googlegroups.com>
"vavr...@gmail.com" <vavruska@gmail.com> writes:

I usually have lint -1 for compiles. I was running into the issue in a larger >project where I did have lint -1 so I wrote quick little test program to >test it out where I forgot the lint. While the little test program works
the large project does not. The following even does not work correct

int x=1;
double d =x;

d winds up being some large number. Does anyone need to know if you need
to startup SANE to use doubles?

On Sunday, April 18, 2021 at 10:55:18 AM UTC-4, thef...@gmail.com wrote:

On Saturday, April 17, 2021 at 6:20:10 PM UTC-7, Kelvin Sherlock wrote:

Use #pragma lint 7 (or -1) to catch missing functions and prototypes.

I didn't really appreciate the value of function prototypes in C until

I was trying to call malloc() on a IIgs. Without the prototype it assumed
the function returned a 16-bit int instead of a 32-bit pointer, and things >didn't go well when I tried to dereference the result.

Since "double" is wider than 16 bits, every function that uses it as an >argument or result needs a prototype. Turning up the warnings is a very good >idea.

-------
ProLine: kelvin@pro-kegs

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)