www-infrastructure-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alan D. Cabrera" <l...@toolazydogs.com>
Subject Re: Brooklyn not in http://people.apache.org/committers-by-project.html
Date Thu, 22 Jan 2015 18:15:51 GMT

> On Jan 22, 2015, at 10:05 AM, sebb <sebbaz@gmail.com> wrote:
> 
> On 22 January 2015 at 17:54, Alan D. Cabrera <adc@toolazydogs.com> wrote:
>> 
>>> On Jan 22, 2015, at 9:36 AM, jan i <jani@apache.org> wrote:
>>> 
>>> On 22 January 2015 at 18:04, Alan D. Cabrera <adc@toolazydogs.com <mailto:adc@toolazydogs.com>>
wrote:
>>> 
>>>> 
>>>>> On Jan 22, 2015, at 8:14 AM, sebb <sebbaz@gmail.com> wrote:
>>>>> 
>>>>> On 22 January 2015 at 15:52, Alan D. Cabrera <adc@toolazydogs.com>
>>>> wrote:
>>>>>> 
>>>>>>> On Jan 22, 2015, at 7:34 AM, sebb <sebbaz@gmail.com> wrote:
>>>>>>> 
>>>>>>> AIUI PPMCs have no legal status within the ASF.
>>>>>>> 
>>>>>>> If LDAP committee entries are created for them, then the tooling
needs
>>>>>>> to be adjusted to cater for this.
>>>>>>> There is other data that really needs to go into LDAP as well.
>>>>>>> AIUI changing LDAP structure is not trivial so ideally all the
changes
>>>>>>> need to be done at once.
>>>>>> 
>>>>>> I am espousing for the creation of a new OU, i.e. Podlings.
>>>>> 
>>>>> It would still require changes to LDAP and Infra processes.
>>>> 
>>>> Changes in LDAP, of course, but these are new orthogonal changes.  Changes
>>>> to Infra processes, yes, but we are cleaning technical debt here and so
>>>> that’s to be expected.  The change is not burdensome.
>>>> 
>>>>>>> However if my suggestion of adding -ppmc entries to asf-auth
is
>>>>>>> acceptable to Infra, then this can be implemented very quickly.
>>>>>> 
>>>>>> What is this asf-auth file? Is it the subversion authorization file?
>>>>> 
>>>>> Yes.
>>>>> 
>>>>> As noted else-thread that is how Corinthia committers are documented.
>>>> 
>>>> Pointing to a prior use of a bad practice is not itself a justification of
>>>> the bad practice.
>>>> 
>>> 
>>> Why not make a step by step approach, looking at the mail is is obvious
>>> that a LDAP change right now misses:
>>> - backing from infra
>>> - backing from the IPMC chair (who would need at least for a while to do
>>> the job)
>>> - a discussion if a REST API is so secure that Infra will allow it (to be
>>> honest I would not)
>>> - the tools that use the REST API.
>>> 
>>> It is a lot more than just collecting the data.
>>> 
>>> as Sebb suggested we can expand the svn-auth file, it is not perfect but it
>>> collects the information PPMC and committer in one place. Once that is done
>>> we could have a longer discussion on
>>> - which information should in general be appended to LDAP (and do we want
>>> LDAP to be our central repository for committer data)
>>> - if yes, can we allow a REST API that updates LDAP, and how to make it
>>> secure
>> 
>> 
>> If you want to put PPMC and committer information in a Subversion authorization file,
I can’t stop you.  What worries me, other than it is a bad practice, is that we’re simply
adding more technical debt that needs to be undone; you guys are already using it as an excuse
to not move to LDAP.
> 
> It is not *adding* to the technical debt.
> 
> The asf-auth file is the correct place for defining SVN groups that
> are not in LDAP.
> The podlings that do use SVN *already* have the committer groups here.
> 
> It so happens that the ppmc groups won't have any SVN folders that need auth.
> Likewise the podlings that use Git don't *need* committer entries.

If you don’t see how that is not a poor practice, then there’s nothing else I can say.
 Let’s just drop it and agree to disagree.

>> But please, I am totally baffled by the pushback on putting this information in LDAP.
 This is not rocket science and there’s not a lot of refactoring that needs to be done to
get us in a good place.
> 
> I'm equally baffled why you seem unwilling to try this simple improvement.

I apologize for my tone.


Regards,
Alan



Mime
View raw message