Hi Rudy,
> Hi Anthony,
>
> I agreed its difficult, but it doesn't mean the problem is
> not existing as it
> is consistently appearing
> everytime we try to do any reverse engineering java code for
> the past one year.
I can understand your frustrations but everything you are saying points to
problems with your classpath environments - remember that Rose is actually a
Java compiler - it compiles the Java into UML. This means that you have to
set it up with the correct and consistent environment like other IDEs.
Also like other IDEs, we have an internal as well as external way of setting
this up. Differences or omissions in these values are the most common cause
of problems - as it is for any other Java environment.
>
> Large number of files = ~50 - ~200 java files.
> The size for each java file averages about 100-2000 lines of
> code where all
> kinds of java features
> are used (e.g. inner classes, non-public classes, abstract classes,
> inheritance, interface, etc.)
> Our classpath expand towards 5 different directories.
These are not uncommon project stats - I have had quite a few that meet or
exceed these specs.
>
> The problem is not consistent; that is, I can saved the Rose
> model file and
> give it to the support
> (as I have done before) when the error occurred. But when
> the model is
> reloaded with a different
> Rose session, Rose do not produce the same error. I think Rose is not
> initializing some variable
> properly at the start of reverse engineering.
I think this could be your classpath variables that are not being
initialised. Are some folks using the CLASSPATH environment variable and
others using Rose? Do some folks have virtual path map variables set and
others not? Are you setting up your environment and not saving the model
afterwards? All these can lead to the behaviour you mention.
>
> I have only one suggestion:
> if rational has a good-size java project, I suggest try using
> Rose to reverse
> engineer the source
> multiple times.
We regularly reverse engineer the JDK and J2EE libraries from both source
and class files with no problems - these are big and complex enough!!
>
> Ironically, the same source code reverse engineered properly
> using TogetherSoft
> white-board
> edition (a freebie) :-)
>
Without commenting too much on competitor's products, none of the other
tools that I know of actually check what they are reverse engineering, so in
terms of UML diagrams, you get what you pay for - as they say, GIGO (garbage
in, garbage out) - at least we tell you it's garbage before we draw it :-).
The particular tool you mention will 'correct' the error by automatically
creating the class for you if it is needed - unfortunately the last project
I was on had a couple of programmers who were either dyslexic or too fast
for their keyboards and there were a lot of 'corrected' classes that no one
knew about until integration time :-)
When we reverse engineered in Rose, we found many problems that were causing
bogus builds depending on what order your directories were in the classpath.
Regards,
Dave.
--
Davyd Norris - Consultant
Regional Services Organisation (Asia Pacific)
Rational Software
Sydney, Australia.
Phone: +61 2 9419 0100
Fax : +61 2 9419 0160
HDesk: +61 2 9419 0111
--
>
************************************************************************
* Rose Forum is a public venue for ideas and discussions.
* For technical support, visit http://www.rational.com/support
*
* Admin.Subscription Requests: [EMAIL PROTECTED]
* Archive of messages:
http://www.rational.com/products/rose/usergroups/rose_forum.jtmpl
* Other Requests: [EMAIL PROTECTED]
*
* To unsubscribe from the list, please send email
*
* To: [EMAIL PROTECTED]
* Subject:<BLANK>
* Body: unsubscribe rose_forum
*
*************************************************************************
