On Wed, Nov 14, 2012 at 2:12 PM, jan iversen <jancasacon...@gmail.com> wrote:
> On 14 November 2012 20:01, Rob Weir <robw...@apache.org> wrote:
>
>> On Wed, Nov 14, 2012 at 1:57 PM, jan iversen <jancasacon...@gmail.com>
>> wrote:
>> > +1 in general to your ideas, it would be VERY nice to have an easy way,
>> and
>> > the more we all do to make it easy the more developers will work for both
>> > projects. I do however have one question.
>> >
>> > Regarding the mark #AOOCONTRIBUTION. It an AOO committer take the code
>> and
>> > integrate it, would that not be a clear violation of the ICLA paragraph
>> 7.
>> > As I read it, taking code requires a lot of extra red tape, compared to
>> if
>> > someone actively sends the code and asks a committer to integrate it ?
>> >
>> > I might be wrong, but from past experience with apache, taking source
>> that
>> > has not clearly been sent with the purpose of integration, can lead to
>> > problems. Remember it is not easy to proof who actually set the flag,
>> > whereas a mail sent is a clear indication.
>> >
>>
>> I agree that this would only work if we know that the patch author set
>> the flag.  But this can be done via normal means.  If you recall, I
>> didn't ask for your fingerprints or a DNA sample before integrating
>> your patches ;-)  Unless shown otherwise I hope we can assume that no
>> one is committing fraud, like editing someone else's commit to add a
>> tag to it.
>>
> No you did not, but as I wrote...I sent the patches on a public mail to
> you, so there are no doubt about my intentions. Just for the sake of
> discussion (I am not implying anybody would do the following), assume I
> issue a patch for AOO (or LO) and do not set the flag, LO (or AOO) wants
> the patch so an administrator "assumes" I forgot the set the flag and helps
> me.
>

So your point is that emails are harder to intentionally or
accidentally change.  Version control and BZ are editable (even
Subversion memos can be changed) but we have these systems set up to
echo changes to email lists (issues@ and commits@) so there is a full
audit trail.   So I think on the Apache side we can handle this.  But
I'm not sure whether LibreOffice also echos all BZ changes, etc., to a
list.

> If I may extend your idea a little, when the flag is sent, AOO/LO sends a
> confirmation e-mail to the developer and asks if it correct to integrate
> (correctly formulated this mail can even give the developer more
> motivation) ?
>

Any solution that relies on custom development is a bit harder.

>
>>
>> -Rob
>>
>> > Jan I
>> >
>> >
>> >
>> > On 14 November 2012 19:28, Rob Weir <robw...@apache.org> wrote:
>> >
>> >> I've heard some discussion and interest in this topic off-list.  There
>> >> has been some practical experience, but nothing that we've written
>> >> down or promoted.  I'd be interested in seeing if we can come up with
>> >> some solid best practices.
>> >>
>> >> The problem:  Many (most?) open source contributors are not opposed to
>> >> AOO or LO.  They are just interested in helping out.  If they produce
>> >> a patch, or documentation, fix a bug or add a translation, they want
>> >> to maximize the public good that comes from that work.  License
>> >> differences are confusing and frustrating and bring them no joy.  They
>> >> want a set of clear instructions for how they can  do the most good
>> >> with the least process overhead.
>> >>
>> >> Naturally, I'm looking at this from the AOO side.  But most of these
>> >> issues are symmetrical.  So for sake of argument, suppose I identify
>> >> myself primarily as a LibreOffice developer/translator/technical
>> >> author, and I want to make my work available more broadly.  What
>> >> should I do?  As I see it, the issues are threefold:  communications,
>> >> technical integration and license.
>> >>
>> >> On the communications side, how do I let AOO know that I've done work
>> >> that I want to contribute to them?  Sending a note to dev@ or posting
>> >> a patch in AOO's BZ would work, of course.  But both require extra
>> >> work for the contributor.  Are there any lighter weight ways of doing
>> >> this?  For example, could we suggest a tag that could be used in git
>> >> or Bugzilla, for the contributor to indicate their intent that the
>> >> contribution be made available to AOO as well?   Something like
>> >> #AOOCONTRIBUTION ?  That would make it easy for us to search for such
>> >> items.
>> >>
>> >> Technical integration -- Due to divergence between the projects, not
>> >> every LO patch can be applied to AOO automatically.  Some will, but
>> >> many will require adaptation.  Certainly the contributor could
>> >> integrate and build their patch for both products.  That would be
>> >> idea.  But it is asking a lot.  Would we accept less?  Or maybe we
>> >> sugest areas where technical integration would be easier and require
>> >> no extra work?  Otherwise, integration would require extra work on our
>> >> end.  But this is not fatal.  In fact it could lead to a set of "easy
>> >> tasks" for new developers.
>> >>
>> >> License -- the differences here are well-known, but are easily solved.
>> >>  A contributor merely needs to state that they are making their patch
>> >> available to AOO under ALv2.  There are various ways to record this
>> >> fact publicly.  One is to make the statement in the source system (git
>> >> or BZ).  But that is extra work.  Another way might be submit an iCLA
>> >> to Apache.  Another way might be to publicly record an intention on
>> >> our dev@ list, along the lines of, "All of my (future/past)
>> >> LibreOffice contributions should be considered also contributions
>> >> under the Apache License 2.0 to the Apache OpenOffice project".
>> >>
>> >> Another other ideas?
>> >>
>> >> -Rob
>> >>
>>

Reply via email to