[pypy-dev] speed.pypy.org and mercurial

Paolo Giarrusso p.giarrusso at gmail.com
Sat Dec 18 10:11:14 CET 2010

On Fri, Dec 17, 2010 at 16:06,  <exarkun at twistedmatrix.com> wrote:
> On 02:48 pm, tobami at googlemail.com wrote:
>>@Jean-Paul:  I thought it would only be a problem if the license was
> Yes, that's right.  But I got sad anyway.
>>But in any case, what is the conclusion, if any? Can the Mercurial API
>>be used from LGPL code or not?
> This seems unanswerable.  If you believe some people, yes.  If you
> believe other people, no.  The only way to resolve the conflict is to do
> it, and either get sued by the copyright holder or not get sued by the
> copyright holder (you could argue that the outcome of the lawsuit must
> also be considered, but by the time you've gotten sued you probably
> don't care anymore).

Well, one point is that the matter is complex enough to make you
suable. Moreover, the point that you're technically not linking seems
extremely technical and specific to Python, and even to CPython - a
JIT compiler links together the code, even more than it happens for C
libraries. Therefore, explaining it to a judge would take an extremely
expensive lawyer.

However, would the Mercurial developers sue you (through some
FSF-sponsorship or the like)? You can ask them if they're bothered by
your license. If they're not, they might still change their mind, but
at least they'll (likely) ask you to change license before suing you
(so they save the money for the lawyer).

Best regards
Paolo Giarrusso - Ph.D. Student

More information about the Pypy-dev mailing list