First of all, thank you so much for asking.

I'll answer to the best of my ability because I work for SAS, who
bought that intellectual property from rPath, so I know something of
where things stand.  This is my understanding, but I'm not authorized
to speak on SAS's behalf, so it isn't a promise made by SAS.

With that disclaimer out of the way...

On Mon, Mar 04, 2013 at 09:54:17AM +0000, Uri Simchoni wrote:
> Is there an open-source project for Conary and rAPA these days? A place to 
> report bugs etc?

For Conary, the process of setting that up is under way.  The first
two steps happened already: publishing updates to Conary out to
bitbucket (in the same location it was previously), and relicensing
Conary under the ASL to make it compatible with more licenses.  I
have been part of the process to bring back public access to issue
tracking for Conary.  It has always been the intent, but there is
a bit more process involved when SAS runs it than when rPath ran it.
(I expect there also to be a site that answers this question so that
I can just point to it instead of writing emails.)

rPath had announced end-of-life for rAPA before SAS acquired some
of rPath's assets. I'm not aware that SAS has any specific interest
in resurrecting work on rAPA.  Basically, since it was released
under an open source license and had an announced end-of-life, I
don't think that it would be forking if another community were to
work on it.  I don't expect SAS would be part of that community.
I can ask to confirm my understanding if you want.

> I have a bug report and couple of fixes to Conary and rAPA - it looks like 
> rAPA leaks memory when running UpdateTroves plugin.
> 
> rAPA bug is circular reference with destructor - fix uses a weakref, perhaps 
> not the cleanest one but is designed to be small. A cleaner fix might be to 
> refactor the code to remove the circular reference altogether.

Unless you want to take on the refactoring, I think that the weakref
is the solution you'll want to use. :)

> Conary bugs are memory and refcount leaks in C code.

Thank you!

One of the things at rPath that made it hard to accept contributions
to Conary was that it required a copyright assignment to rPath, which
not everyone could sign.  The good side of that, though, was that it
did not take any discussion for SAS to re-license Conary under the
ASL, which has rather broad compatibility with other licenses.

At SAS, I expect it to take a contributor agreement but not a copyright
assignment.  The open question, as far as I know so far, is which form
of contributor agreement.

Finally, I think you forgot to post the patches, unless I have somehow
accidentally enabled a "strip attachments from posts" feature in
setting up this list in mailman...

_______________________________________________
Foresight-devel mailing list
[email protected]
https://lists.foresightlinux.org/mailman/listinfo/foresight-devel

Reply via email to