Re: Declining package maintenance requests? (Was Re: Large number of packages to be orphaned on Feb 26)

[Jason L Tibbitts III]

> "FAL" == Fabio Alessandro Locati  writes:

FAL> If a person is not able to make a click in 7 days (maybe vacation
FAL> periods could be excluded from the count), why should he be able to
FAL> do so in the following 21 days?

I think that a better question is:

If a maintainer is not able to deal with such things in a reasonable
time, why are they the only maintainer listed for a package?

We used to have a vacation page, but that's kind of weird privacy page.
I always try to let people I trust know when I'm going to be away.
Certainly if you are going to be away so long that you can't address a
comaintainership request in a reasonable time then that's exactly the
situation where you need comaintainers.

We really need to not be in the situation where a package has only one
maintainer.  And we really need to make sure that everyone knows that
provenpackagers are going to be touching your packages and people are
going to try and get involved.  They just are.  If you can't handle that
(and maybe having to do an occasional merge or revert) then you'd better
have something at the top of your spec explaining the situation.

 - J<
--
devel mailing list
devel@lists.fedoraproject.org
http://lists.fedoraproject.org/admin/lists/devel@lists.fedoraproject.org

Re: Declining package maintenance requests? (Was Re: Large number of packages to be orphaned on Feb 26)

[Fabio Alessandro Locati]

On Sun, Feb 21, 2016 at 01:58:55PM +0100, Ralf Corsepius wrote:
> On 02/21/2016 11:44 AM, Fabio Alessandro Locati wrote:
> 
> >Now, I think that it makes sense to have the POC/Package Admins able to
> >approve and/or decline requests. I think the following 3 improvements
> >would greatly improve the ACL request experience:
> >
> >- auto-accept the ACL requests after 7 days of no POC/PA responded
> 
> I am against any automated approvals and dislike your proposal. Also, I do
> not see how automated ACLs would change anything about the situation in
> question.

True,
Sorry, I should have clarified it better: this would solve the problem I
see way more present of unresponsive maintainers preventing new
maintainers to maintain packages.
 
> Besides this, 7 days would be much too short. 4 weeks would be much more
> appropriate.

If a person is not able to make a click in 7 days (maybe vacation
periods could be excluded from the count), why should he be able to do
so in the following 21 days? Also, Fedora already have ACL cases where
the POC/PA have 7 days to respond, otherwise a positive response is
assumed.

> >- introduce an (optional) text box to allow to bundle a message with the
> >   ACL request
> Such a message should be mandatory, because I've received ACLs from people
> I've never heard before and from "inexperienced and overly eager
> new-comers".

I have said optional because a lot of times the asking-maintainer and
the POC already knows each others. Obviusly a small message could be ok
in those cases as well.

> >- allow (and force) a formal message in the case of ACL denial
> This would mean to allow people to implement ACL-request spam, with the load
> ending on those peoples' shoulders who are actually doing the Fedora work.

I don't see how this could happend.
Maybe I explained my thought in a bad way.
I'll try to explain with an example:
- user A asks ACL for package X
- X POC decides to decline A request.
In my proposal, X must justify the decline.

Best,
Fale

-- 
Fabio Alessandro Locati

PGP Fingerprint: B960 BE9D E7A8 FA12 273A  98BB 6D6A 29D6 709A 7851
https://keybase.io/fale


signature.asc
Description: PGP signature
--
devel mailing list
devel@lists.fedoraproject.org
http://lists.fedoraproject.org/admin/lists/devel@lists.fedoraproject.org

Re: Declining package maintenance requests? (Was Re: Large number of packages to be orphaned on Feb 26)

[Kevin Kofler]

Jonathan Underwood wrote:
> My personal opinion is that (with a few possible exceptions, perhaps
> critical path, or a subset of that), co-maintainership requests
> shouldn't ever be turned down. That's really not in the spirit of
> Fedora. Which makes me wonder why, for most packages, we even need
> approval from the point of contact when requesting commit access on
> packages.

Huh? This just doesn't make sense. Either we have ACLs or we don't. If you 
are going to just blanket-allow ACL requests, then why not just drop the 
whole ACL scheme entirely (i.e., s/provenpackager/packager/ as in the good 
old cvsextras days)?

Kevin Kofler
--
devel mailing list
devel@lists.fedoraproject.org
http://lists.fedoraproject.org/admin/lists/devel@lists.fedoraproject.org

Re: Declining package maintenance requests? (Was Re: Large number of packages to be orphaned on Feb 26)

[Ralf Corsepius]

On 02/21/2016 11:44 AM, Fabio Alessandro Locati wrote:


Now, I think that it makes sense to have the POC/Package Admins able to
approve and/or decline requests. I think the following 3 improvements
would greatly improve the ACL request experience:

- auto-accept the ACL requests after 7 days of no POC/PA responded


I am against any automated approvals and dislike your proposal. Also, I 
do not see how automated ACLs would change anything about the situation 
in question.


Besides this, 7 days would be much too short. 4 weeks would be much more 
appropriate.




- introduce an (optional) text box to allow to bundle a message with the
   ACL request
Such a message should be mandatory, because I've received ACLs from 
people I've never heard before and from "inexperienced and overly eager 
new-comers".



- allow (and force) a formal message in the case of ACL denial
This would mean to allow people to implement ACL-request spam, with the 
load ending on those peoples' shoulders who are actually doing the 
Fedora work.


Ralf
--
devel mailing list
devel@lists.fedoraproject.org
http://lists.fedoraproject.org/admin/lists/devel@lists.fedoraproject.org

Re: Declining package maintenance requests? (Was Re: Large number of packages to be orphaned on Feb 26)

[Jonathan Underwood]

On 21 February 2016 at 10:44, Fabio Alessandro Locati
 wrote:
> On Sun, Feb 21, 2016 at 10:10:53AM +, Jonathan Underwood wrote:
>> This jumped out at me as particularly worrisome:
>>
>> On 19 February 2016 at 19:15, Fabio Alessandro Locati
>>  wrote:
>> > If Christopher will not respond, I can take care of:
>> [snip]
>> > Also I can help with the i3 packages as I volounteered few weeks ago and
>> > have been shot down by Christopher becase "no more help is needed".
>> >
>>
>> My personal opinion is that (with a few possible exceptions, perhaps
>> critical path, or a subset of that), co-maintainership requests
>> shouldn't ever be turned down. That's really not in the spirit of
>> Fedora. Which makes me wonder why, for most packages, we even need
>> approval from the point of contact when requesting commit access on
>> packages.
>>
>> Perhaps we need to have a clear policy on this?
>
> In my experience, only in two cases my requests for ACL have been
> denied:
>
> - one is the case we are speaking about, and you can read it from [0]
> - one is a case of a cold-sent ACL request that have been denied.
>   Getting to talk 2 minutes with the POC, introducing myself and saying
>   "hi" made him approve the ACL with no further questions asked.
>
> Now, I think that it makes sense to have the POC/Package Admins able to
> approve and/or decline requests. I think the following 3 improvements
> would greatly improve the ACL request experience:
>
> - auto-accept the ACL requests after 7 days of no POC/PA responded
> - introduce an (optional) text box to allow to bundle a message with the
>   ACL request
> - allow (and force) a formal message in the case of ACL denial


I think these are three excellent proposals, thanks!
--
devel mailing list
devel@lists.fedoraproject.org
http://lists.fedoraproject.org/admin/lists/devel@lists.fedoraproject.org

Re: Declining package maintenance requests? (Was Re: Large number of packages to be orphaned on Feb 26)

[Fabio Alessandro Locati]

On Sun, Feb 21, 2016 at 10:10:53AM +, Jonathan Underwood wrote:
> This jumped out at me as particularly worrisome:
> 
> On 19 February 2016 at 19:15, Fabio Alessandro Locati
>  wrote:
> > If Christopher will not respond, I can take care of:
> [snip]
> > Also I can help with the i3 packages as I volounteered few weeks ago and
> > have been shot down by Christopher becase "no more help is needed".
> >
> 
> My personal opinion is that (with a few possible exceptions, perhaps
> critical path, or a subset of that), co-maintainership requests
> shouldn't ever be turned down. That's really not in the spirit of
> Fedora. Which makes me wonder why, for most packages, we even need
> approval from the point of contact when requesting commit access on
> packages.
> 
> Perhaps we need to have a clear policy on this?

In my experience, only in two cases my requests for ACL have been
denied:

- one is the case we are speaking about, and you can read it from [0]
- one is a case of a cold-sent ACL request that have been denied.
  Getting to talk 2 minutes with the POC, introducing myself and saying
  "hi" made him approve the ACL with no further questions asked.

Now, I think that it makes sense to have the POC/Package Admins able to
approve and/or decline requests. I think the following 3 improvements
would greatly improve the ACL request experience:

- auto-accept the ACL requests after 7 days of no POC/PA responded
- introduce an (optional) text box to allow to bundle a message with the
  ACL request
- allow (and force) a formal message in the case of ACL denial

Best,
Fale

[0] https://bugzilla.redhat.com/show_bug.cgi?id=1276844
-- 
Fabio Alessandro Locati

PGP Fingerprint: B960 BE9D E7A8 FA12 273A  98BB 6D6A 29D6 709A 7851
https://keybase.io/fale


signature.asc
Description: PGP signature
--
devel mailing list
devel@lists.fedoraproject.org
http://lists.fedoraproject.org/admin/lists/devel@lists.fedoraproject.org

Declining package maintenance requests? (Was Re: Large number of packages to be orphaned on Feb 26)

[Jonathan Underwood]

This jumped out at me as particularly worrisome:

On 19 February 2016 at 19:15, Fabio Alessandro Locati
 wrote:
> If Christopher will not respond, I can take care of:
[snip]
> Also I can help with the i3 packages as I volounteered few weeks ago and
> have been shot down by Christopher becase "no more help is needed".
>

My personal opinion is that (with a few possible exceptions, perhaps
critical path, or a subset of that), co-maintainership requests
shouldn't ever be turned down. That's really not in the spirit of
Fedora. Which makes me wonder why, for most packages, we even need
approval from the point of contact when requesting commit access on
packages.

Perhaps we need to have a clear policy on this?

Cheers,
Jonathan
--
devel mailing list
devel@lists.fedoraproject.org
http://lists.fedoraproject.org/admin/lists/devel@lists.fedoraproject.org