www-infrastructure-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From sebb <seb...@gmail.com>
Subject Re: Brooklyn not in http://people.apache.org/committers-by-project.html
Date Tue, 20 Jan 2015 23:23:02 GMT
On 20 January 2015 at 23:08, jan i <jani@apache.org> wrote:
> On 20 January 2015 at 23:48, sebb <sebbaz@gmail.com> wrote:
>
>> On 20 January 2015 at 21:12, Alan D. Cabrera <list@toolazydogs.com> wrote:
>> >
>> >> On Jan 20, 2015, at 1:00 PM, David Nalley <david@gnsa.us> wrote:
>> >>
>> >> On Tue, Jan 20, 2015 at 3:14 PM, Alan D. Cabrera <list@toolazydogs.com
>> <mailto:list@toolazydogs.com>> wrote:
>> >>>
>> >>>> On Jan 20, 2015, at 12:04 PM, David Nalley <david@gnsa.us>
wrote:
>> >>>>
>> >>>> On Tue, Jan 20, 2015 at 3:01 PM, Alan D. Cabrera <
>> list@toolazydogs.com> wrote:
>> >>>>>
>> >>>>>> On Jan 20, 2015, at 11:56 AM, David Nalley <david@gnsa.us>
wrote:
>> >>>>>>
>> >>>>>> On Tue, Jan 20, 2015 at 2:50 PM, Alan D. Cabrera <
>> list@toolazydogs.com <mailto:list@toolazydogs.com>> wrote:
>> >>>>>>>
>> >>>>>>>> On Jan 20, 2015, at 11:31 AM, jan i <jani@apache.org>
wrote:
>> >>>>>>>>
>> >>>>>>>> On 20 January 2015 at 20:22, Alan D. Cabrera <
>> list@toolazydogs.com <mailto:list@toolazydogs.com>> wrote:
>> >>>>>>>>
>> >>>>>>>>>
>> >>>>>>>>>> On Jan 20, 2015, at 11:19 AM, jan i <jani@apache.org>
wrote:
>> >>>>>>>>>>
>> >>>>>>>>>> On 20 January 2015 at 20:12, Alan D. Cabrera
<
>> list@toolazydogs.com>
>> >>>>>>>>> wrote:
>> >>>>>>>>>>
>> >>>>>>>>>>>
>> >>>>>>>>>>>> On Jan 20, 2015, at 11:08 AM, sebb
<sebbaz@gmail.com> wrote:
>> >>>>>>>>>>>>
>> >>>>>>>>>>>> On 20 January 2015 at 18:12, Alan
D. Cabrera <
>> adc@toolazydogs.com>
>> >>>>>>>>>>> wrote:
>> >>>>>>>>>>>>>
>> >>>>>>>>>>>>>> On Jan 20, 2015, at 10:08
AM, David Nalley <david@gnsa.us>
>> wrote:
>> >>>>>>>>>>>>>>
>> >>>>>>>>>>>>>> Is there a reason it needs
to be added?
>> >>>>>>>>>>>>>
>> >>>>>>>>>>>>> That seems like an odd question
and I would turn it around
>> and ask, is
>> >>>>>>>>>>> there a reason why it shouldn’t?
>> >>>>>>>>>>>>>
>> >>>>>>>>>>>>>> IIRC that page is derived
from the authorization file for
>> SVN -
>> >>>>>>>>>>>>
>> >>>>>>>>>>>> Yes
>> >>>>>>>>>>>>
>> >>>>>>>>>>>>>> Brooklyn doesn't use svn,
so no listing.
>> >>>>>>>>>>>>
>> >>>>>>>>>>>> It does not *need* an entry in asf-auth,
but one can be
>> provided.
>> >>>>>>>>>>>>
>> >>>>>>>>>>>>> Time to fix the tooling… :)
 Where’s the code that generates
>> those
>> >>>>>>>>>>> pages?
>> >>>>>>>>>>>>
>> >>>>>>>>>>>> The tooling is not broken.
>> >>>>>>>>>>>>
>> >>>>>>>>>>>> There is currently no readily accessible
data defining the
>> members of
>> >>>>>>>>>>>> the Brooklyn podling.
>> >>>>>>>>>>>>
>> >>>>>>>>>>>> Once a podling graduates, it will
have an LDAP group.
>> >>>>>>>>>>>
>> >>>>>>>>>>> Then what about all the other podlings
that are on this page?
>> >>>>>>>>>>>
>> >>>>>>>>>>
>> >>>>>>>>>> Documentation for podlings says you should
update that file, so
>> I did it
>> >>>>>>>>>> for corinthia even though we use git, and
it worked nicely.
>> >>>>>>>>>
>> >>>>>>>>> What file are you speaking of?
>> >>>>>>>>>
>> >>>>>>>>
>> >>>>>>>> this one
>> >>>>>>>>
>> https://svn.apache.org/repos/infra/infrastructure/trunk/subversion/authorization/asf-authorization-template
>> <
>> https://svn.apache.org/repos/infra/infrastructure/trunk/subversion/authorization/asf-authorization-template
>> >
>> >>>>>>>>
>> >>>>>>>> Search for "bookkeeper=breed"
>> >>>>>>>>
>> >>>>>>>> You need to add brooklyn after that line. Commit
the file and the
>> rest
>> >>>>>>>> happens automatically within 24 hours (people.a.o
is updated with
>> a cron
>> >>>>>>>> job).
>> >>>>>>>
>> >>>>>>> Is there a corresponding authorization file for git?
>> >>>>>>>
>> >>>>>>
>> >>>>>> No
>> >>>>>> Git authorization is much more coarse.
>> >>>>>> tl;dr - we parse the name of the repo before the first delimiter
and
>> >>>>>> look for a PMC in LDAP by that name and see if the committer
is a
>> >>>>>> member of that LDAP group.
>> >>>>>
>> >>>>> By PMC I think you mean project, correct?  But I’m not sure
if
>> podlings are in LDAP.
>> >>>>>
>> >>>>
>> >>>> No.
>> >>>> I meant PMC
>> >>>> Podlings are not projects in the top level sense, and have no entry
>> in LDAP.
>> >>>
>> >>> So for podlings it’s all incubator committers, as Jan said in another
>> email?
>> >>>
>> >>
>> >> Correct.
>> >>
>> >>> The podling committer membership and PPMC membership information seems
>> to be spread around if at all.  Does it make sense to create LDAP groups
>> for them to provide a canonical source?
>> >>>
>> >>
>> >> In my experience, podlings don't do a good job of keeping up with the
>> >> data that needs to be stored in so many locations.
>> >> (Their website, their status file, the svn auth file). Adding yet
>> >> another place to keep up with things seems the wrong direction to
>> >> head.
>> >
>> > I’m hearing a description of all the complicated things that occur
>> because we don’t put podling membership information in LDAP.
>>
>> What complicated things?
>>
>> > We can simplify that, that’s a tooling issue.
>> > there’s no requirement to have membership information in a website and
>> if there is it should be auto generated from LDAP anyway
>> > the status file should be auto generated from LDAP anyway
>> > the svn auth file should be pulling info from LDAP and does do that for
>> non-podlings
>>
>> Not every group in the SVN auth file is in LDAP
>>
>> >> Presumably if we added an LDAP group for the podling we'd also need to
>> >> add a PPMC group for the podling as well.
>>
>> No, I don't think that would be required.
>> Or a good thing, because PPMCs are not PMCs.
>>
>> >
>> > Yes, and that would be a good thing.
>>
>> It would be more work for Infra and the podling.
>> There is no distinction between PPMC and podling committers.
>> This only occurs once the podling graduates.
>>
>> >> I am also not sure that it gives a lot of advantages, and I know it
>> >> adds overhead, overhead that can currently only be dealt with by a PMC
>> >> Chair. With that said, what problem are we actually trying to solve?
>> >
>> > The problem that there is no source for PPMC membership at all and that
>> podling membership is implicitly managed in an SVN auth file.
>>
>> The source is the SVN auth file.
>>
> Actually not ! the SVN auth file contains all committers for a podling, but
> not who is PPMC.

I was under the impression that the PPMC consisted of all the
committers for the podling.

At graduation time, those committers who are no longer involved will
generally not be included in the initial PMC.

>
>> Frankly, I’m surprised that I’m getting pushback in putting podling group
>> information in LDAP.
>>
>> It would be more work overall.
>> The LDAP group would have to be created (and then deleted if the
>> podling does not graduate).
>>
>> And the group would still have to be maintained by someone.
>>
>> It's no harder to update the SVN auth file than to update LDAP.
>> Indeed I would say it is simpler. And it's obvious who is already in the
>> group.
>>
> I agree on that, but that would currently only give committers not PPMC of
> a podling.
>
> The only place PPMC is registred is in the podlings status file (xml).

That assumes there is a difference between committers and PPMC for a podling.
As I wrote above, that is not my understanding.

> rgds
> jan i
>
>
>
>>
>> >
>> > Regards,
>> > Alan
>> >
>> >
>> >
>>

Mime
View raw message