Mailing-List: contact dev-help@commons.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Commons Developers List" <dev@commons.apache.org>
Subject: Re: [math] Exception Design
To: Commons Developers List <dev@commons.apache.org>
References: <56784178.9060000@gmail.com>
 <e4ae22d24e65538ee9d86fde7d2fd742@scarlet.be>
From: Ole Ersoy <ole.ersoy@gmail.com>
Message-ID: <5678D51F.8000401@gmail.com>
Date: Mon, 21 Dec 2015 22:44:16 -0600
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101
 Thunderbird/38.4.0
MIME-Version: 1.0
In-Reply-To: <e4ae22d24e65538ee9d86fde7d2fd742@scarlet.be>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit


On 12/21/2015 06:44 PM, Gilles wrote:
> On Mon, 21 Dec 2015 12:14:16 -0600, Ole Ersoy wrote:
>> Hi,
>>
>> I was considering jumping into the JDKRandomGenerator exception
>> discussion, but I did not want to hijack it.
>>
>> Not sure if any of you have had a chance to looks at this:
>> https://github.com/firefly-math/firefly-math-exceptions/
>>
>> https://github.com/firefly-math/firefly-math-exceptions/blob/master/src/main/java/com/fireflysemantics/math/exception/MathException.java
>
> I had a rapid look; unfortunately not in sufficient details to
> grasp the major departures from the existing framework.
> Could you display one or two examples comparing CM and firefly?
In addition to what I summarized below one detail that I think is important is that the ExceptionTypes enum allows for more exact targeting of the exception root cause.  For instance right now I have the following arithmetic exception types:

     /**
      * MATH ARITHMETIC EXCEPTIONS
      */
     MAE("MAE"),
     MAE__INTEGER_OVERFLOW("MAE__INTEGER_OVERFLOW"),
     MAE__LONG_OVERFLOW("MAE__LONG_OVERFLOW"),
     MAE__OVERFLOW_IN_ADDITION("MAE__OVERFLOW_IN_ADDITION"),
     MAE__OVERFLOW_IN_SUBTRACTION("MAE__OVERFLOW_IN_SUBTRACTION"),
     MAE__GCD_OVERFLOW_32_BITS("MAE__GCD_OVERFLOW_32_BITS"),
     MAE__GCD_OVERFLOW_64_BITS("MAE__GCD_OVERFLOW_64_BITS"),
     MAE__LCM_OVERFLOW_32_BITS("MAE__LCM_OVERFLOW_32_BITS"),
     MAE__LCM_OVERFLOW_64_BITS("MAE__LCM_OVERFLOW_64_BITS"),
     MAE__DIVISION_BY_ZERO("MAE__DIVISION_BY_ZERO"),

So by looking at the exception type we know exactly what the issue is.  With this approach CM will always only have 1 exception.  If more types are needed then just add another line to the ExceptionTypes Enum.  The new type is used to look up the message template in the I18N resource bundle.


>
> It looks neat.
Thanks :)
> But I did not see how localization is handled.

I did leave localization out.  I think localization was a hard requirement in earlier versions of CM, but I'm hoping that there is some flexibility on this and that future versions can defer to a utility that uses the ExceptionTypes Enum instance as the key to look up the internationalized template string.

>
>> I think it satisfies everyone's requirements with:
>> - A single MathException (No hierarchy)
>
> That would not satisfy everyone. :-!
>
>> - The ExceptionTypes Enum contains all the exception types
>> - The ExceptionTypes Enum 'key' maps to the corresponding message 1 to 1
>> - The ExceptionFactory (Utility) throws exceptions, if necessary,
>> that have always have a single unique root cause, such as NPEs
>
> I was wondering whether the "factory" idea could indeed satisfy
> everyone.
>
> Rather than throwing the non-standard "MathException", the factory would
> generate one of the standard exceptions, constructed with the internal
> "MathException" as its cause:
I think it's good that CM throws CM specific exceptions.  This way when I write the handler I can know that the exception is CM specific without having to unwrap it.

>
> public class ExceptionFactory {
>
>   public static void throwIllegalArgumentException(MathException e) {
>     throw new IllegalArgumentException(e);
>   }
>
>   public static void throwNullPointerException(MathException e) {
>     throw new NullPointerException(e);
>   }
>
>   // And so on...
> }
>
> So, CM code would be
>
> public class Algo {
>
>   public void evaluate(double value) {
>     // Check precondition.
>     final double min = computeMin();
>     if (value < min) {
>       final MathException e = new MathException(NUMBER_TOO_SMALL).put(CONSTRAINT, min).put(VALUE, value);
>       ExceptionFactory.throwIllegalArgumentException(e);
>     }
>
>     // Precondition OK.
>   }
> }
Another thing that I hinted to is that the the factory builds in the precondition check in the throw method.  So that the line:

     if (value < min) {

can be nixed.

>
>
>
> Then, in an application's code:
>
> public void appMethod() {
>   // ...
>
>   // Use CM.
>   try {
>     Algo a = new Algo();
>     a.evaluate(2);
>   } catch (IllegalArgumentException iae) {
>     final Throwable cause = iae.getCause();
>     if (cause instanceof MathException) {
>       final MathException e = (MathException) cause;
>
>       // Rethrow an application-specific exception that will make more sense
>       // to my customers.
>       throw new InvalidDataInputException(e.get(CONSTRAINT), e.get(VALUE));
>     }
>   }
> }
>
> This is all untested.
> Did I miss something?

I think you got it all...But the handler will be shorter if the exception is not wrapped.  The pattern I'm used to is that libraries wrap the exceptions of other libraries in order to offer a standardized facade to the user.  For example Spring wraps Hibernate exceptions, since Spring is a layer on top of Hibernate and other data access providers.

Ole

>
> Gilles
>
>
>> - The context captures the exception parameters keyed by an the
>> 'ExceptionKeys' enum.  Each module can introduce more keys as
>> necessary.
>> - The toString() method can be used as the initial exception message
>>
>> The way developers should deal with this exception is:
>> 1) Catch
>> 2) Get the type (Enum)
>> 3) Get the  parameters (Context)
>> 4) Get the method that threw it
>> 5) Get the class threw it
>> 6) Rethrow the above in an application specific exception, log it, or
>> display it.  Construct a localized message using the enum type to look
>> up the exception template if needed.
>>
>> WDYT?
>>
>> Cheers,
>> - Ole
>>
>> P.S. Here's the entire test demo (Pasted below for convenience):
>>
>> https://github.com/firefly-math/firefly-math-exceptions/blob/master/src/test/java/com/fireflysemantics/math/exceptions/MathExceptionTest.java
>>
>> /**
>>   *  Licensed under the Apache License, Version 2.0 (the "License");
>>   *  you may not use this file except in compliance with the License.
>>   *  You may obtain a copy of the License at
>>   *
>>   *  http://www.apache.org/licenses/LICENSE-2.0
>>   *
>>   *  Unless required by applicable law or agreed to in writing, software
>>   *  distributed under the License is distributed on an "AS IS" BASIS,
>>   *  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
>>   *  See the License for the specific language governing permissions and
>>   *  limitations under the License.
>>   */
>>
>> package com.fireflysemantics.math.exceptions;
>>
>> import static com.fireflysemantics.math.exception.ExceptionKeys.CONSTRAINT;
>> import static com.fireflysemantics.math.exception.ExceptionKeys.VALUE;
>> import static
>> com.fireflysemantics.math.exception.ExceptionTypes.NUMBER_TOO_SMALL;
>> import static org.junit.Assert.assertEquals;
>> import static org.junit.Assert.assertTrue;
>>
>> import org.junit.Test;
>>
>> import com.fireflysemantics.math.exception.ExceptionFactory;
>> import com.fireflysemantics.math.exception.MathException;
>>
>> public class MathExceptionTest {
>>
>>     @Test(expected = MathException.class)
>>     public void verifyThrows() {
>>         throw new MathException(NUMBER_TOO_SMALL);
>>     }
>>
>>     @Test
>>     public void verifyCode() {
>>         try {
>>             throw new MathException(NUMBER_TOO_SMALL);
>>         } catch (MathException e) {
>>             assertEquals(e.getType(), NUMBER_TOO_SMALL);
>>         }
>>     }
>>
>>     @Test
>>     public void verifyContext() {
>>         try {
>>             throw new MathException(NUMBER_TOO_SMALL).put(CONSTRAINT,
>> 2).put(VALUE, 1);
>>         } catch (MathException e) {
>>             assertEquals(e.get(CONSTRAINT), 2);
>>             assertEquals(e.get(VALUE), 1);
>>         }
>>     }
>>
>>     @Test
>>     public void verifyToString() {
>>         try {
>>             throw new MathException(NUMBER_TOO_SMALL).put(CONSTRAINT,
>> 2).put(VALUE, 1);
>>         } catch (MathException e) {
>> assertTrue(e.toString().contains(NUMBER_TOO_SMALL.toString()));
>>             assertTrue(e.toString().contains("1"));
>>             assertTrue(e.toString().contains("2"));
>>             assertTrue(e.toString().contains(CONSTRAINT));
>>             assertTrue(e.toString().contains(VALUE));
>>         }
>>     }
>>
>>     @Test
>>     public void verifyFactory() {
>>         try {
>>             ExceptionFactory.throwNumberToSmallException(1, 2, "foo");
>>         } catch (MathException e) {
>>             assertTrue(e.getType() == NUMBER_TOO_SMALL);
>>             assertEquals(e.get(CONSTRAINT), new Integer(2));
>>             assertEquals(e.get(VALUE), new Integer(1));
>>             assertEquals(e.get("foo"), new Integer(1));
>>         }
>>     }
>> }
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> For additional commands, e-mail: dev-help@commons.apache.org
>
>


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org