This is the mail archive of the cygwin-apps mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Maintainer ping?


On Sun, 3 Feb 2013 17:53:55 -0500, Christopher Faylor wrote:
> >On 2013-02-03 18:58, Christopher Faylor wrote:
> >>Given the problems that we seem to have with maintainers going absent
> >>would it make sense to send out a periodic ping every M days to all
> >>maintainers?  Then if we didn't receive an answer within N pings the
> >>package could be marked orphaned automatically.
> 
> As a refinement, we could also add something which detects when a person
> had been active in one of the cygwin lists and defers sending the ping.

Definitely; there's no sense in asking the obvious.

But I am unsure about how to best implement this, given a previous
package maintainer's comments last year[1].  In short, he maintained a
single, obscure package, which still worked (so no bug reports on
Cygwin or other deps), upstream was dead (so no RFUs), and nobody has
written about it to the list (so nothing to respond to).  When finally
pinged, he was still there, but how often would someone like that need
to be pinged?  I can just imagine the conversation:

>> Are you still with us?
> Yep.
[two months pass]
>> Are you still with us?
> Yep.
[two more months]
>> Are you still with us?
> Yes, I'm still here.
[and again]
>> Are you still with us?
> YES, I'm *still* here.
[and another time]
>> Are you still with us?
> YES, DIDN'T YOU HEAR ME THE FIRST TIME?!?
[once again]
>> Are you still with us?
...[silence]...
>> Are you still with us?
...
>> Hello?  Hello?  Anybody home?
...
>> I guess not.

So for someone like that, regular pinging may just be
counterproductive.  OTOH, with those who maintain a lot of packages, or
extremely important packages, there's much more cause for concern over
an extended absence, causing delays for the project as a whole when they
don't respond when needed.

So while a good idea, I would suggest using some common sense in its
implementation.


Yaakov

[1] http://cygwin.com/ml/cygwin-apps/2012-03/msg00167.html


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]