Charles Oliver Nutter wrote:
Wayne Meissner wrote:
That correct, with a small wrinkle. Where there is an ABI compatible
library, which can be used as a complete replacement, and the binary
is dynamically linked, then it could be argued that the code is
intended to link with the ABI compatible library, and not the GPLed
variant.
Given that libeditline is BSD licenced, and provides the same api as
libreadline (indeed, on MacOS, it _is_ liibreadline), then the GPL
would not apply - at least as far as using readline via FFI.
Yeah, Conrad and I discussed that aspect of it too. And this is purely
my supposition, but I don't think it would be hard to argue in court
that *any* library dynamically linking to another could be claimed to
expect only that library's ABI, and that since said ABI could
potentially be reimplemented, GPL would again not apply. Or at least,
that sounds logical in my head.
It also seems difficult to believe that companies shipping libreadline
(like Apple) haven't already been over this, or aren't at least
comfortable that the "linking clause" is ambiguous enough that they
don't need to worry about it. Last I checked, OS X was not GPL, and it
both ships and links to libreadline.
Before going any further, take a look at RMS arguing about this here
(specifically about the ABI for readline, and so on):
http://clisp.cvs.sourceforge.net/*checkout*/clisp/clisp/doc/Why-CLISP-is-under-GPL
Cheers
--
Ola Bini (http://olabini.com)
Ioke creator (http://ioke.org)
JRuby Core Developer (http://jruby.org)
Developer, ThoughtWorks Studios (http://studios.thoughtworks.com)
Practical JRuby on Rails (http://apress.com/book/view/9781590598818)
"Yields falsehood when quined" yields falsehood when quined.
---------------------------------------------------------------------
To unsubscribe from this list, please visit:
http://xircles.codehaus.org/manage_email
