Programming Forums > Application Development > C++

Sane · Jan 19th, 2008, 3:59 PM

Quote:

Originally Posted by grumpy

I am not a believer in spoon-feeding anyone except small babies or people who are somehow incapacitated.

But not if you're the cause of their incapacitation, right?

grumpy · Jan 19th, 2008, 4:06 PM

While I have the capability to do so, I have never incapacitated anyone's ability to learn, and hopefully will never have to.

Cache · Jan 19th, 2008, 5:28 PM

Channel 9 did an interview with Ale Contenti on exception handling. People here might find it interesting.

Understanding Exceptions and When/How to Handle Them:
http://channel9.msdn.com/Showpost.aspx?postid=343189

I was surprised to learn about the amount of bookkeeping that goes on under the covers when using exceptions.

peaceofpi · Jan 20th, 2008, 3:44 PM

Quote:

Originally Posted by grumpy

No, it is not. It does, however, rely on the reader being willing to put some effort into understanding rather than being provided with a set of simple recipes they can apply without thought.

I am not a believer in spoon-feeding anyone except small babies or people who are somehow incapacitated.

I never said I wanted to be spoonfed, and I'll tell you now I still don't want it. Honestly, as someone who knows as little about exceptions as I do, the article went far over my head pretty quickly. I suppose that's my problem, but it certainly doesn't mean I'm part of the herd of people here with low post counts and high expectations.

Jabo · Jan 20th, 2008, 11:05 PM

From what I know about error handling (which is pretty slim) it could be more robust. Let me use an example. If you're capturing keyboard input in your program, you poll for whatever keystrokes and if found, you mark it as handled, otherwise it goes on to the OS to be handled. Error handling, from my recollection, doesn't have this function. Why?

The error handling mechanism automatically handles every error, even if it doesn't? That seems broken.

peaceofpi · Jan 21st, 2008, 9:57 AM

How's this for a first attempt?

 c++ Syntax (Toggle Plain Text) 
bool isprime(int n)
{
    try
    {
        if (n == 0) throw 0;
        if (n < 0) throw -1;
        if (n % 2 == 0) return false;
        if (n == 1 || n == 2) return true;
        for (int i = 3; i < n/2; i++)
            if (n % i == 0)
                return false;
    }
    catch(int i)
    {
        // handle 0 or negative numbers here
    }
    return true;
}
bool isprime(int n)
{
    try
    {
        if (n == 0) throw 0;
        if (n < 0) throw -1;
        if (n % 2 == 0) return false;
        if (n == 1 || n == 2) return true;
        for (int i = 3; i < n/2; i++)
            if (n % i == 0)
                return false;
    }
    catch(int i)
    {
        // handle 0 or negative numbers here
    }
    return true;
}

grumpy · Jan 21st, 2008, 2:47 PM

That code would be thrown out in any serious code review, and the author sacked on the spot.

There is no practical need to throw/catch an exception if a function that that detects an error can recover from it. It would be simpler to just recover on the spot ..... and return false. The code is simpler, easier to understand, and probably (depending on how exception handling is implemented by your compiler) more efficient.

0,1, and negative values are, by definition, non-prime (yes, your algorithm is incorrect as it returns true for 1) so the error recovery is trivial. I won't even address that fact that, efficiency wise, the loop is .... bleh!

Sane · Jan 21st, 2008, 3:08 PM

Yikes! Your function says 1 is a prime number, and 2 is composite!

peaceofpi · Jan 22nd, 2008, 1:40 PM

Factors of 2: 1, 2, ?

As for grumpy

Quote:

Originally Posted by grumpy

That code would be thrown out in any serious code review

Quote:

Originally Posted by me

How's this for a first attempt?

I got my answer, I guess I should have known better than to ask further questions.

Regardless,

 c++ Syntax (Toggle Plain Text) 
bool isprime(int n)
{
    if (n <= 1) return false;
    if (n == 2) return true; // stays till someone comes up with another factor of 2
    for (int i = 3; i <= n/2; i++)
        if (n % i == 0)
            return false;
    return true;
}
bool isprime(int n)
{
    if (n <= 1) return false;
    if (n == 2) return true; // stays till someone comes up with another factor of 2
    for (int i = 3; i <= n/2; i++)
        if (n % i == 0)
            return false;
    return true;
}

Sane · Jan 22nd, 2008, 3:19 PM

I can't tell if you were agreeing or disagreeing with me, but either way, your new code handles n=2 properly now. Your old code returned false when n % 2 == 0. That was incorrect, since 2 is a prime number.

Anyways, you have it now.

Jan 19th, 2008, 4:06 PM	#12
grumpy Programming Guru Join Date: Jun 2005 Location: Adelaide, South Australia Posts: 1,198 Rep Power: 5	Re: Exception handling: point is...? While I have the capability to do so, I have never incapacitated anyone's ability to learn, and hopefully will never have to.

Jan 19th, 2008, 5:28 PM	#13
Cache Hobbyist Join Date: Sep 2005 Posts: 259 Rep Power: 3	Re: Exception handling: point is...? Channel 9 did an interview with Ale Contenti on exception handling. People here might find it interesting. Understanding Exceptions and When/How to Handle Them: http://channel9.msdn.com/Showpost.aspx?postid=343189 I was surprised to learn about the amount of bookkeeping that goes on under the covers when using exceptions.

Jan 20th, 2008, 11:05 PM	#15
Jabo Not a user? Join Date: Sep 2007 Posts: 245 Rep Power: 1	Re: Exception handling: point is...? From what I know about error handling (which is pretty slim) it could be more robust. Let me use an example. If you're capturing keyboard input in your program, you poll for whatever keystrokes and if found, you mark it as handled, otherwise it goes on to the OS to be handled. Error handling, from my recollection, doesn't have this function. Why? The error handling mechanism automatically handles every error, even if it doesn't? That seems broken.

Jan 21st, 2008, 9:57 AM	#16
peaceofpi hi: for(;;) goto hi; Join Date: Jun 2006 Posts: 86 Rep Power: 3	Re: Exception handling: point is...? How's this for a first attempt? c++ Syntax (Toggle Plain Text) bool isprime(int n) { try { if (n == 0) throw 0; if (n < 0) throw -1; if (n % 2 == 0) return false; if (n == 1 \|\| n == 2) return true; for (int i = 3; i < n/2; i++) if (n % i == 0) return false; } catch(int i) { // handle 0 or negative numbers here } return true; } bool isprime(int n) { try { if (n == 0) throw 0; if (n < 0) throw -1; if (n % 2 == 0) return false; if (n == 1 \|\| n == 2) return true; for (int i = 3; i < n/2; i++) if (n % i == 0) return false; } catch(int i) { // handle 0 or negative numbers here } return true; } __________________ How do you play Religious Roulette? Stand around in a circle and blaspheme till someone gets struck by lightning.

Jan 21st, 2008, 2:47 PM	#17
grumpy Programming Guru Join Date: Jun 2005 Location: Adelaide, South Australia Posts: 1,198 Rep Power: 5	Re: Exception handling: point is...? That code would be thrown out in any serious code review, and the author sacked on the spot. There is no practical need to throw/catch an exception if a function that that detects an error can recover from it. It would be simpler to just recover on the spot ..... and return false. The code is simpler, easier to understand, and probably (depending on how exception handling is implemented by your compiler) more efficient. 0,1, and negative values are, by definition, non-prime (yes, your algorithm is incorrect as it returns true for 1) so the error recovery is trivial. I won't even address that fact that, efficiency wise, the loop is .... bleh!

Jan 21st, 2008, 3:08 PM	#18
Sane Programming Guru Join Date: Apr 2005 Posts: 1,799 Rep Power: 5	Re: Exception handling: point is...? Yikes! Your function says 1 is a prime number, and 2 is composite!

Jan 22nd, 2008, 3:19 PM	#20
Sane Programming Guru Join Date: Apr 2005 Posts: 1,799 Rep Power: 5	Re: Exception handling: point is...? I can't tell if you were agreeing or disagreeing with me, but either way, your new code handles `n=2` properly now. Your old code returned false when `n % 2 == 0`. That was incorrect, since 2 is a prime number. Anyways, you have it now.

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Exception Handling and Events	InfoGeek	C#	3	Jan 9th, 2008 1:10 AM
exception handling class	rwm	C++	5	Apr 10th, 2007 5:17 AM
exception handling	l2u	C++	2	Mar 3rd, 2007 5:47 PM
Exception Handling with Threads	Harakim	Software Design and Algorithms	13	Aug 8th, 2006 6:00 AM
exception handling	chrome_knob	Java	2	May 19th, 2005 5:45 PM

Programming Forums > Application Development > C++

Exception handling: point is...?