Eclipse Community Forums: Java Development Tools (JDT) » Determining specific java class where a specific bug originated from

Home » Language IDEs » Java Development Tools (JDT) » Determining specific java class where a specific bug originated from

Show: Today's Messages :: Show Polls :: Message Navigator

Determining specific java class where a specific bug originated from [message #256871]

Sat, 11 October 2008 20:16

babich25@hotmail.com

Messages: 16
Registered: July 2009

Junior Member

Specific bugs within the bugzilla are distributed over several java
classes and across multiple package hierarchy. I am conducting research on
prediction models, and so far have not been able to determine for each bug
as to where exactly the bug originated. I would like to find out if there
is any way to determine where the specific bug did originate first -
specific java class or package. Any suggestions appriciated.

Report message to a moderator

Re: Determining specific java class where a specific bug originated from [message #256875 is a reply to message #256871]

Sun, 12 October 2008 01:00

Walter Harley

Messages: 847
Registered: July 2009

Senior Member

<babich25@hotmail.com> wrote in message
news:8f4104c014cd4451d67e303897557f33$1@www.eclipse.org...
> Specific bugs within the bugzilla are distributed over several java
> classes and across multiple package hierarchy. I am conducting research on
> prediction models, and so far have not been able to determine for each bug
> as to where exactly the bug originated. I would like to find out if there
> is any way to determine where the specific bug did originate first -
> specific java class or package. Any suggestions appriciated.

I question your assumption.

Bugs are sometimes associated with a specific java class or package;
sometimes not. Sometimes a "bug" is a difference of opinion. Often a bug
is caused by a disagreement between the assumptions two different pieces of
code - and often in such a case when it is fixed, it is not fixed in the
"guilty" code but simply in the code where it will be easier to fix. Even
in a seemingly trivial case like a missing null check, the bug may be in the
code that produced the null, rather than in the code that failed to check
it. Bugs due to synchronization often are due to cycles of assumptions
distributed throughout the code, such that every individual assumption seems
good but taken as a whole they do not work.

If you prefer to ignore that and proceed anyway, I would suggest that you
start from CVS, rather than from Bugzilla: search CVS commit records for
references to specific bugs. Usually if the primary purpose of a code
change is to fix a bug, it will be indicated in the commit record.

Report message to a moderator

Previous Topic:	Building a modified version of org.eclipse.jdt.core
Next Topic:	Unable to open Eclipse Help when offline

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

]

Current Time: Tue Oct 01 10:33:57 GMT 2024

.:: Contact :: Home ::.

Breadcrumbs

Sign up to our Newsletter