This is the mail archive of the
mailing list for the Mauve project.
Re: legal question about writing mauve test cases
- From: Dalibor Topic <robilad at kaffe dot org>
- To: Greg Orlowski <greg dot orlowski at gmail dot com>
- Cc: mauve-discuss at sources dot redhat dot com
- Date: Sun, 13 Nov 2005 14:19:36 +0100
- Subject: Re: legal question about writing mauve test cases
- References: <email@example.com>
Greg Orlowski wrote:
Is it legal to use the reflection api's w/ sun's jdks to generate
lists of method signatures and then compare them to gnu-classpath's
method signatures as a reference-point from which to develop test
cases for mauve?
I assume that you're interested in seing which classes/methods/fields
are missing/bad in GNU Classpath, in order to figure out which areas to
go after, you may be interested in looking at Japitools comparison
results at http://www.kaffe.org/~stuart/japi/
Does using the reflection capabilities of java with sun's libraries
constitute reading source code (legally)?
No, afaict. Otoh, if you are using or writing, say, an automated test
generator for libraries, the cleanest way to do it is to generate your
tests against GNU Classpath sourcecode, and then to run the tests
against a non-free VM, to see if the assumptions codified in GNU
Classpath also exist in the non-free VM.
On a side note, CnC from http://www.cc.gatech.edu/cnc/welcome.html might
be interesting for such a task.
I'd like to contribute to classpath, but I'll be using sun's tools
extensively at my job, and I'm worried that I'll be tainted by doing
Afaik, using Sun's tools does not taint you, since you don't get to see
actual implementation details by just using java/javac. Looking at the
sources or having a NDA/contract with Sun not to disclose their
implementation details otoh, would prevent you from contributing
respective souce code to GNU Classpath. In that case, there are still
other areas one can contribute to.