log10: SING Error -> Try/Catch??

Thread View: borland.public.cppbuilder.nativeapi

7 messages

[T] Back to Threads [V] Tree View [G] Group Articles

7 total messages Started by Dan Wed, 18 Apr 2007 14:55

#2164

Author: Dan
Date: Wed, 18 Apr 2007 14:55

49 lines
2044 bytes

log10 functions where the value passed to the function is negative 
results in a "LOG10: DOMAIN error" if the value is 0 then it's a "LOG10: 
SING error"  For example the following code will create a DOMAIN error.


#include <math.h>

double DoLogFunction(const double &val)
{
   return log10(val);
}

void main()
{
   DoLogFunction(0);  //SING ERROR
   DoLogFunction(-1); //DOMAIN ERROR
}


Try/Catch around the 'DoLogFunction()' function call or even around the 
log10 call doesn't work.

Of course I can test val to see if it's negative and non-zero.  The 
problem is, I think I have done that in the following code:

double DoLogFunction(const double &val)
{
   double result = 0;
   if (val>0)
   {
      try
      {
         result = log10(val);
      }
      catch (...)
      {
        ;//do nothing, result is already 0
      }
   }
   return result;
}

However, occasionally I still get the error which results in a Windows 
dialog box appearing indicating the error.

What is the BEST way to do this type of trapping so that I don't get the 
error?

thanks in advance

Re: log10: SING Error -> Try/Catch??

#2166

Author: Thomas Maeder [TeamB]
Date: Wed, 18 Apr 2007 19:22

39 lines
1950 bytes

Dan <Dan@do.not.spam> writes:

> log10 functions where the value passed to the function is negative
> results in a "LOG10: DOMAIN error" if the value is 0 then it's a
> "LOG10: SING error" For example the following code will create a
> DOMAIN error.
>
>
> #include <math.h>
>
> double DoLogFunction(const double &val)
> {
>    return log10(val);
> }
>
> void main()

The return type of main() has to be int.

> {
>    DoLogFunction(0);  //SING ERROR
>    DoLogFunction(-1); //DOMAIN ERROR
> }
>
>
> Try/Catch around the 'DoLogFunction()' function call or even around
> the log10 call doesn't work.

Because what the program does is causing undefined behavior, not
throwing exceptions.

> What is the BEST way to do this type of trapping so that I don't get
> the error?

As for most problems, there isn't a uniform best solution.

Have you considered providing your own version of _matherr()?

Re: log10: SING Error -> Try/Catch??

#2187

Author: Dan
Date: Tue, 24 Apr 2007 13:56

13 lines
1496 bytes

> The return type of main() has to be int.
Ok; I just used this as an example; of course I didn't strictly copy and 
paste code, but instead just tried to make an example (I work in the 
embedded world more often, and void main() is acceptable there. My 
apologies.

> Because what the program does is causing undefined behavior, not
> throwing exceptions.
Is there a way to catch these so that they don't go all the up to the 
top of the application?

> Have you considered providing your own version of _matherr()?
How?

Re: log10: SING Error -> Try/Catch??

#2191

Author: Thomas Maeder [TeamB]
Date: Tue, 24 Apr 2007 16:12

28 lines
2069 bytes

Dan <Dan@do.not.spam> writes:

>> The return type of main() has to be int.
>
> Ok; I just used this as an example; of course I didn't strictly copy
> and paste code, but instead just tried to make an example (I work in
> the embedded world more often, and void main() is acceptable
> there. My apologies.

void main() isn't acceptable anywhere. At no time in history did C or
C++ accept any return type from main() other than int.

All this isn't extremely important, but whenever somebody posts code
to a newsgroup, somebody else might use it as a (good) example; I
therefore tend to be pedantic. My apologies :-)

>> Because what the program does is causing undefined behavior, not
>> throwing exceptions.
>
> Is there a way to catch these so that they don't go all the up to
> the top of the application?
>
>> Have you considered providing your own version of _matherr()?
>
> How?

RTFM. Or Google. Or both.

Re: log10: SING Error -> Try/Catch??

#2192

Author: Alan Bellingham
Date: Tue, 24 Apr 2007 16:20

18 lines
1610 bytes

maeder@glue.ch (Thomas Maeder [TeamB]) wrote:

>void main() isn't acceptable anywhere. At no time in history did C or
>C++ accept any return type from main() other than int.

Though in a freestanding implementation, it may not be required that
main() actually exist!

>All this isn't extremely important, but whenever somebody posts code
>to a newsgroup, somebody else might use it as a (good) example; I
>therefore tend to be pedantic. My apologies :-)

Bearing in mind the sheer volume of Herbert-Schildt-inspired code out
there, it sometimes seems like a losing battle.

Alan Bellingham
-- 
ACCU Conference 2008: 2-5 April 2008 - Oxford (probably), UK

Re: log10: SING Error -> Try/Catch??

#2193

Author: Bob Gonder
Date: Tue, 24 Apr 2007 16:44

9 lines
1169 bytes

Dan wrote:

>> Have you considered providing your own version of _matherr()?

>How?

http://msdn2.microsoft.com/en-us/library/aa298453(VS.60).aspx

There is also a section in the BDS help on it.

Re: log10: SING Error -> Try/Catch??

#2195

Author: Bruce Salzman
Date: Tue, 24 Apr 2007 20:43

14 lines
1404 bytes

"Alan Bellingham" <alan@lspace.org> wrote in message 
news:lebs231ac8kcr1nvp2utompgbbt9djq1qh@4ax.com...
>
> Bearing in mind the sheer volume of Herbert-Schildt-inspired code 
> out
> there, it sometimes seems like a losing battle.
>

The legacy of a former prog-rock keyboard player.
:^)

-- 
Bruce

Thread Navigation

This is a paginated view of messages in the thread with full content displayed inline.

Messages are displayed in chronological order, with the original post highlighted in green.

Use pagination controls to navigate through all messages in large threads.

Back to All Threads