Grant said:
Let's say I use a GPL'd python module (e.g. something installed
in site-packages) in an application.
Let's also say I use py2exe to package and distribute said
application.
Is what I'm distributing a "derived work" of the GPL'd python?
Or is py2exe's packaging of the module's .pyc file and my
application code's .pyc files a "mere aggregation" so that I
only have to provide source code for the GPL'ed module and not
for my application code?
IOW, do I have to GPL my application code and distribute source
code for it?
Here's my personal spin on it - I think that if your distributed binary
will contain GPL'ed code, that simplifies it - it makes your distributed
binary subject to the terms of the GPL.
Yet again, personal interpretation abounds, but in the spirit of the
GPL, if *what you actually distribute* for people to run *contains* the
GPL'ed stuff and uses it, then it's surely a derivative work, is it not?
If you wanted to avoid the issue with the library, it'd be worth
reimplementing it or offering it separately.
Overall, if you're going to release modules and insist GPLness, they
should really be released under the LGPL by convention shouldn't they?
Maybe you should contact the developer about that.
I really wish people would consider the LGPL when they're releasing
libraries etc. - OSS developers need to realise that it's encumbent on
them not to erode or threaten the legal status of the GPL, and that
releasing software under the GPL in this manner not only throws things
into uncertainty, but could (and IANAL so feel free to enlighten me
here) ultimately end up with the GPL being compromised in court in
future IP disputes, couldn't it?
I'd almost prefer to boycott software that uses the GPL frivolously or
inappropriately, because I fear it could threaten the GPL's existence
ultimately.
</high horse>
Anyway, have you come to a conclusion on this one? (or have I missed a post)