java-nio-charset-enhanced -- Milestone 4 is released

Martin Buchholz martinrb at
Wed Apr 15 01:01:25 UTC 2009

(Sorry for my continuing extreme communications lag)

On Mon, Apr 6, 2009 at 20:45, Mark Reinhold <mr at> wrote:
>> Date: Mon, 06 Apr 2009 17:27:41 -0700
>> From: Martin Buchholz <martinrb at>
>> On Mon, Mar 30, 2009 at 21:51, Mark Reinhold <mr at> wrote:
>>>> Date: Sun, 29 Mar 2009 11:27:26 -0700
>>>> From: Martin Buchholz <martinrb at>
>>>> The problem is more human.  One would like to give credit for good
>>>> ideas or good analysis, but the only official way to give credit in a
>>>> commit message is via a simple
>>>> Contributed-by: email-address
>>>> which raises legal doubts even when there is no copyrighted material.
>>> Exactly what sort of legal doubts do you have in mind here?
>>> We already require the contributor of any nontrivial change to submit an
>>> SCA prior to that change being integrated.  Is that not sufficient?
>> I am worried about:
>> - contributions that do not require a copyright assignment,
>>   e.g. an excellent bug report
> Bug reports submitted to, to, or in
> e-mail to an list are covered by appropriate Terms of
> Use (ToU) statements.  Are they not sufficient?

The problem is not that *I* have any legal doubts about whether
a contribution is legally acceptable, but that I worry that *others*
will have doubts if I put anyone's name in a Contributed-by: when that
name cannot be found in the SCA signers list.

The phrase "Contributed-by:" *sounds* like it has legal meaning.

I understand why the Guide imposes draconian restrictions on
the format of a changeset comment, but I'd like to see some
loosening of those restrictions.  I would like the ability to add free form
text to the commit messages, even with the risk of teams adopting
local conventions about their use.

But if we're going to allow just one loosening,
it would be to allow *multiple* *free-form*



>> - contributions from a Google employee, because of repeated institutional
>>   amnesia that Google has a blanket SCA, which is not reflected in
> I'm not sure why that is, but I'll look into it.
>>>> I guess one can abuse the Summary: field to squeeze in thank-yous,
>>>> but it's pretty obvious that you are circumventing the process.
>>> Perhaps -- but if we need to fix the process, then we can do that.
>> It seems that for changes for which I have an external contributor,
>> there is a large variety in the nature of the contribution;
>> sometimes an off-hand remark will cause an avalanche of work on my part.
>> The nature of the collaboration cannot
>> be expressed in the commit comment.  Perhaps we could have
>> a Thank-you: with free-form content in addition to a Contributed-by:
>> with the latter reflecting a legal copyright author to keep the lawyers happy.
> You seem to think that the Contributed-by line exists in order to keep
> Sun's lawyers happy.  It doesn't.  Sun's lawyers actually know nothing
> about it.
> We defined the Contributed-by line in order to give credit to non-author
> contributors in a visible and permanent way.  We could equally well have
> called it the "Thank-you" line, but we didn't.
> As a changeset author you're free to list anyone whom you think worthy of
> credit in the Contributed-by line.  Is the problem here that you assumed
> otherwise?  If so then we just need to improve the documentation.
>> Speaking of which -- hopefully jcheck will soon be open-sourced?!
> Yes.
> - Mark

More information about the core-libs-dev mailing list