On 7 January 2015 at 06:13, Paolo Bonzini <pbonz...@redhat.com> wrote:
>
>
> On 25/11/2014 15:17, Peter Maydell wrote:
>> Questions for review:
>>  * can we do the git cherry-pick thing I mention above?
>
> I'm afraid that would double the size of the repository (in terms of
> number of commits).
>
> One possibility is this:
>
>     git checkout 158142c2
>     git cherry-pick this-series~5      # patch 1
>     git cherry-pick this-series~2      # patch 4
>     git cherry-pick this-series~1      # patch 5
>     git merge origin/master
>     git checkout -p -- this-series~1   # patch 5
>     git commit --amend -C HEAD
>     git cherry-pick this-series        # patch 6
>
> Then the reverts are implicit in the merge commit.
>
> However, this will probably still cause problems with bisection, since
> the merge base between 2.2 and 2.3 will be in prehistory.

I meant more in the sense of how we generated the eventual
relicensing-patch to apply to current master, rather than
trying to make master's actual history have weird things in it.
Anyway, it sounds like the thing I did is the best we can do.

>>  * should we squash the revert and reimplement patches together?
>>    (avoids bisection break but makes the revert-and-reimplement less clear)
>
> I think we can do that, provided the reimplemented functions are in
> different places than the originals, so that the revert-and-reimplement
> is still clear from the diff.

OK. I'll make sure I do that with the next version.

>>  * are people happy with my attempt to clarify the licensing status of
>>    the source files in patch 6, and my choice of GPLv2+ for future
>>    contributions to them?
>
> I think SoftFloat-2a is more appropriate but I don't really care.

The thing is that after all these relicensings we end up with a
file with a mix of licenses in it. So for somebody actually
using the file the controlling license is GPLv2+. (In particular
all the RedHat contributions are GPLv2+, not SoftFloat-2a...)
So it seemed simplest to say 'GPLv2+ for future changes'; but
it wouldn't make any major difference to pick softfloat-2a I guess
(and that is the license covering the bulk of the code so it
does make more sense in some ways).

Thanks for looking at the series.
-- PMM

Reply via email to