Stephen Lau wrote:
> Garrett D'Amore wrote:
>> Stephen Lau wrote:
>>> As Shawn noticed... the open-sourcing of the e1000g driver yesterday
>>> contained proprietary source code headers in the diff.  This was a
>>> mistake by the engineer moving the code, and needs to be re-bridged.
>>> We can't have proprietary source code headers in public open source
>>> code, no matter how minor and trivial the code is.
>>>
>>> I've re-bridged the putback in onnv/onnv-gate, which means you will
>>> need to roll your repositories back to revision 3567:6d1c4dd92b63, or
>>> else your repositories will grow a new head.
>>>
>>> Thanks, and sorry for the inconvenience.
>>> cheers,
>>> steve
>>
>> Events like this are confusing to me.   Please bear with me.
>>
>> Maybe I'm just naive in my dealings with Hg, but is the need for this
>> kind of special attention a result of Hg itself, an artifact of the
>> current Hg bridge, or the result of rolling back a repository instead of
>> just "patching forward" (i.e. not treating the backout as 2nd update,
>> but instead actually "undoing" the original commit?)
>>
>> I'm not sure this is the same kind of Hg snafu in the past, but the
>> upshot is that I'm not sure whether I should be using Hg at this time,
>> or should I be sticking with SVN?  (I don't work on my ON clone every
>> day; the last couple of times this happened I just deleted my old
>> workspace and started with a fresh one.)
>>
>> Again, I apologize if I sound confused -- but I am.  I'd like to hear
>> some "official" word on what those of us on the outside _should_ be
>> using as the most up-to-date source for Solaris.
>
> Hi Garrett,
>     It's an artifact of rolling back a repository, and is not an
> artifact of Hg in any way.  The point of an SCM is to make sure no
> history is ever lost - correct?  The problem is, due to the legal
> aspect of this, we need to make the repository "forget" that it ever
> happened - so it's going counter to what an SCM should be doing.  We
> would have this exact same issue whether it was CVS, SVN, or even good
> old Teamware/SCCS.
>     The problem with "patching forward" is that the mistake is still
> in the history..
>     onnv/onnv-gate continues to be the most up-to-date source for
> Solaris.
>
> cheers,
> steve

Thanks.

Given that part of the point of the putback for e1000g was to open
source it, I would have expected the reviewers to carefully checked for
coverage before it was released.  (Both technical and legal reviewers.)

I would try to emphasize to the individuals involved that:

a) failure to catch this ahead of time has placed the company at risk
b) it cannot be undone (see Keith's message) -- please explain to
lawyers, etc. involved automatic mirroring
c) backing out like this has a distinctly negative impact on the entire
OpenSolaris community

This doesn't mean that it won't happen again, but it does reduce the
likelihood by making folks aware of the impact.

When I have open sourced stuff that I developed for another company I
have always had to jump thru hoops to get permission for everything --
but those hoops have prevented me from exposing anything without
explicit approval from the stake holders (cf. NetBSD's radeonfb, which I
had to get permission from both Tadpole/General Dynamics and ATI.  If
ATI changes their mind, too bad.  The source is already released, and
cannot ever be "taken back".)

-- 
Garrett D'Amore, Principal Software Engineer
Tadpole Computer / Computing Technologies Division,
General Dynamics C4 Systems
http://www.tadpolecomputer.com/
Phone: 951 325-2134  Fax: 951 325-2191

_______________________________________________
opensolaris-code mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/opensolaris-code

Reply via email to