float to string

[icarus31]

Hi,

I have an issue to convert a float to a string with the Coldfire MCF52259. Here what I am trying to do:

char str[10];

float fval = 1201 / 10;

sprintf (str, "%3.1f\n", fval);

The debugger shows that fval is 120.100. This is what I expect to see. But "str" is not set correctly. If I do the same thing with an integer, there is no problem:

char str[10];

int val = 1201

sprintf (str, "%d\n", val);

What I am doing wrong?

Thanks in advance

[vier_kuifjes]

This thread may give you a hint for a solution:

https://community.freescale.com/thread/46585

[scifi]

---

icarus31 wrote:

float fval = 1201 / 10;

... 

The debugger shows that fval is 120.100.

---

That's odd. This code should initialize fval with 120.0. I wonder if you edited the code before posting it. Maybe there are other significant differences?

---

icarus31 wrote:

sprintf (str, "%3.1f\n", fval);

---

Perhaps you meant "%5.1f"? Or simply "%.1f"?

---

icarus31 wrote:

But "str" is not set correctly.

---

 What's the contents of str?

[icarus31]

SciFi,

I should say "float fval = (float) 1201/10;"  I did not copy paste the code because I was working on another computer.

The "%3.1f" was basically make sure it will be 3 digits before the dot and 1 digit after it. So, it should print 120.1. The "%5.1f" or "%.1f", for that particular case will have the same result as what I wrote.

The "str" variable was set to :

str[0] = 'f'

str[0] = '\x00'

str[1] = '\x91'

str[2] = '\x9C'

str[3] = '\x00'

str[4] = '\x00'

str[5] = '\x00'

str[6] = '\x00'

str[7] = '\x00'

str[8] = '\x00'

[scifi]

---

icarus31 wrote:

The "%3.1f" was basically make sure it will be 3 digits before the dot and 1 digit after it. 

---

So I figured. But that's not what it does. Look up format specifier reference.

---

icarus31 wrote:

The "str" variable was set to :

str[0] = 'f'

str[0] = '\x00'

str[1] = '\x91'

str[2] = '\x9C'

str[3] = '\x00'

str[4] = '\x00'

str[5] = '\x00'

str[6] = '\x00'

str[7] = '\x00'

str[8] = '\x00' 

---

 This appears to support BugMan's suggestion.

[TomE]

For those browsing this quickly and wanting to know why it printed what it did:

scifi wrote:

> icarus31 wrote:

> > The "str" variable was set to :

> >

> > str[0] = 'f'

> > str[0] = '\x00'

> >

> > "%3.1f"

> This appears to support BugMan's suggestion.

That was because the limited library scanned the print string, got to the "f", wasn't built for floats and thus converted "%f" to "f" in the buffer.

Tom

[Nouchi]

Hello,

If I can give an advice, if you can, you'd better switch to CW7.2.2, because there's big improvement between CW7.1 and CW7.2.

If you keep CW7.1, if I remember, you have to rebuild libraries to enable printf floating point support.

[icarus31]

Sorry for the delay. I did not try anything yet, but I will once I come back to that project.

I will keep you posted

Thanks all!

[Nouchi]

Hi,

If you use CW7.2, by librarian item, I meant Edit->"YourProject" Settings->Linker->Librarian.

You will find the same in CW10 : Project->Properties->C/C++ Build->Settings->tools settings->Librarian

Regards,

Emmanuel

[icarus31]

This is interresting, my version of CW 7.1.2 does not have the "Librarian" section under "Linker". I have "ColdFire Processor" and "Global Optimizations"

[Nouchi]

Hello,

Your syntax seems to be right, did you check linker options?

You have to select int_FP option for print formatter in Librarian item, it could be the problem.

regards,

Emmanuel

[icarus31]

BugMan (Emmanuel),

Does the "Librarian item" is something is set in CodeWarrior? Is it in "Edit -> Int Flash Debug Settings" menu?