svn commit: r490569 - in head/graphics: mesa-dri mesa-dri/files mesa-libs

[Jan Beich]

Mathieu Arnold <mat at FreeBSD.org> writes:

> On Thu, Jan 17, 2019 at 03:34:37PM +0000, Jan Beich wrote:
>
>>   Approved by:	maintainer timeout (2 weeks after 2019Q1)
>
> portmgr hat on.
>
> Maintainer timeout means the maintainer never said anything.  In this
> case, the maintainers clearly told you NO repeatedly, see comments 3, 5,
> 9, and also by the maintainer setting explicitely
> "maintainer-feedback-".

In bug 233034 comment 1 said there was a dicision to wait for 2019Q1.
I've asked how to interpret this in comment 2, and comment 3 elaborated.
I've pinged the team in comment 6 but got no response for 2 weeks. Given
the silence, the decision to land in 2019Q1 and positive call for testing
I've landed the patch, assuming responsibility for any regressions.

Bug 230298 was obsoleted by bug 233034. I guess, landing it separately
led to a confusion. Even there the maintainer's feedback wasn't as
simple as NO: the alleged regressions turned out to be known issues,
wait until after EuroBSDCon became obsolete, maintainer-feedback-
was a misuse of the flag, lack of testing was covered by call for testing,
testing protocol never materialized, that bug was the place where the
coordination happened, etc.

Before opening the can of worms I wish you've waited for maintainer's
response if I did anything wrong when applying the timeout.

> When the maintainer says NO, it does not mean you have to wait 2 weeks
> to update the port, it means you absolutely will not, ever, commit the
> patch.

In that case, can you document maintainers can stall patches however long
they want and "not exclusive ownership" is actually false?

https://www.freebsd.org/doc/en/books/porters-handbook/makefile-maintainer.html

>
> Do not, ever, commit to the mesa ports without explicit maintainer
> approval.