Subject: Re: Class Form Checker bug fix...
From: Paul Freeman (pfreeman@ccs.neu.edu)
Date: Wed Dec 11 2002 - 17:19:26 EST
Yes it should. Your second set of stats look correct I will copy them
and update the SETUP file. Thanks for letting me know.
Paul
Dale Vaillancourt wrote:
> Paul,
>
> Does this bugfix affect the statistics that you get with the default
> settings in AnyDynamic.java? Your SETUP.txt in the command line
> package states that you obtained the following statistics:
>
> Statistics:
> Type Count %Total
> Total Checks 1168
> Violations 458 39.21
> Valid Checks 710 60.78
>
> But when I built and ran the project with default AnyDynamic.java, I
> obtained the following:
>
> Statistics:
> Type Count %Total
> Total Checks 1168
> Violations 95 8.13
> Valid Checks 1073 91.86
>
> Thanks!
> Dale
>
>
> On Wednesday, December 11, 2002, at 03:49 PM, Paul Freeman wrote:
>
>> Hi Class -
>>
>> If any of you are still working on part2a of the final exam, there
>> was a bug introduced in the Class Form checker a while ago that is
>> now fixed. As we had discussed in class quite a while ago, a change
>> was made to include a check of the target class as a valid class and
>> not to include what was originally a check of the "declaring type" of
>> the method being checked. This solution had been proposed to fix a
>> problem with flagging inherited method calls as violations. Well it
>> introduced a bug where subclass target objects were now flagged as
>> violations. The correct implementation is to check both. This fix
>> has been made and is available in the versions of both the:
>> LoDChecker Eclipse Plugin
>> command-line_aplib
>>
>> to be used with the final.
>>
>> Some of the students had pointed out this bug. Sorry this has been
>> fixed so late. Professor Lieberherr will be notified and I am sure
>> this will not affect your grades.
>>
>> Paul
>>
>>
> --
> Dale Vaillancourt
> dalev@ccs.neu.edu
>
>
This archive was generated by hypermail 2b28 : Wed Dec 11 2002 - 17:23:07 EST