I was struggling with a recursive algorithm to build the CLASSPATH because
of inherit. (Ought I follow what is on this depend, or what is on the
project? Do I follow 'runtime' despite inherit, or in conjunction with,
etc.) I persuaded myself of a number of permutations, none of which were
correct -- until I finally arrived where I am now. I think the answer to the
first is 'both', and (especially now I separate out/set the bootclasspath)
the results appear to bear that out.

http://lsd.student.utwente.nl/gump/index.html
http://lsd.student.utwente.nl/gump/todos.html

Most of these look 'correct', i.e. issues that the teams would resolve if we
could/would nag them. That said, despite 163 successes, there are 228 not
attempted due to pre-requisite failures (on LSD). Maybe those would uncover
combinations of dependencies that would break the logic, I can't tell.

Projects | Successes | Failures | Prereqs | No Works | Packages
512       |  163          | 25         | 228      | 56              | 40

I see some odd things (almost every run) but then almost every run is
different (not just my changes) so I've not been able to put my finger on
anything. I tried to compare to gump.covalent.com/log, but with different
times of run, I'm sometimes comparing apples and oranges. I need to find a
way to determine if there are any other subtle differences that I'm missing.

I feel that results on LSD differ from results on dotnot.org, especially
w.r.t XALAN. Xalan causes a lot of things not to build. (See:
http://lsd.student.utwente.nl/gump/xml-xalan/xml-xalan2.html)

What are folks thoughts on how I verify that  the Gump logic is now complete
and correct? What are folks thoughts on how I verify that the environment
Gump runs in/produces is now complete and correct?

Thanks in advance for any help.

regards,

Adam
--
Experience Sybase Technology...
http://www.try.sybase.com


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to