Policy RFC: When to touch other peoples packages (or: Fix stuff that needs fixing)

[Patrice Dumas]

On Mon, Oct 23, 2006 at 07:15:25PM +0200, Thorsten Leemhuis wrote:
> Hi all!

Overall I find the proposal very satisfying, I have some minor remarks.
One general remark, maybe it could be said somewhere that contacting by 
mail/irc could also be done in place or in addition to bugzilla. No need
to mention if it is too obvious.


> === Inactive Packager ===
> 
> The packager normally should keep track of his packages; that means:
> 
>   * looking at the [:Extras/PackageStatus: Package Status] pages in the
> wiki now and then

I am not sure about that one. I find the Package Status very usefull, but I
can't see in which case it would justify a change in a package without 
a report to the maintainer. Otherwise said, I don't think having an issue
listed on the Package Status enough to permit experienced users to modify
others packages without noticing them.

>   * fix EVR problems if they get mentioned in the problem reports on the
> fedora-extras-list

I would propose a refinement here: 
  * fix EVR problems if they get mentioned in the problem reports on the
  fedora-extras-list, except for EVR issues with devel which occurs outside
  of the pre-release mass-rebuilds.

The reasoning is that during testing periods those EVR issues with devel 
are not to be considered serious as to justify somebody modifying others
package without a notice.

>   * Important stuff should be fixed as quickly if possible -- waiting
> one day for the maintainer to show up and step in to fix a issue that
> got reported to him is considered enough

Sometimes it is even acceptable not to wait at all. For example one of
my packages screwed the build system and I think it was perfectly right
not to wait a minute to fix it.

>   * not that important problems should be dealt with quickly, too --
> waiting for 2-7 days (depending how bad the issue is) is considered enough
> 
>  * bugs need similar treatment like security problems:

I have a cosmetic remark here, maybe would be better like
   * bugs needing similar treatment like security problems:

>   * annoying, but not that harmful bugs -- waiting for 21 days is
> considered enough

That's not always true. I think that those case should better be agreed
upon in bugzilla discussion, and should only be acted upon if there is 
no disagreement, only a lack of responsivness.

>  * if you committed changes to another package wait some hours if
> possible (normally 24 or 48) before you actually build the updated
> package. That leaves some time for the maintainer to wake up ;-)

It may be obvious, but this is a good rule but only for issues that are
not serious.

> As experienced packagers count:
> 
>  * FESCo members

I don't want to be negative here, nor show disrepect to FESCo members,
but I don't think being FESCo members should qualify here. A FESCo member
could be there to represent something and not because he has great 
technical ability to fix packages. As a side note aren't the current 
FESCo members all sponsor? 
 
>  * Sponsors
> 
>  * Especially the QA and Security SIGs
> 
>  * the SIG that's handling the area in which the package or the fix is
> needed. But
> 
>   * the SIG has to exists for at least two months and needs to have
> shown activity in the past two months
> 
>   * the member is and experienced Extras packager and around for at
> least four months and member of the SIG for at least one month


I think that it would also be right if experienced fedora core packagers
could be considered as experienced Fedora packagers. Not all the core
packagers, since some don't seems to care that much about the packaging 
guidelines. What about a rule like 

* core packagers that have been through a number of reviews (for extra 
  or core packages) and are core packagers for more than 4 months.

Or something like that.

I think that the reverse could also be true, core packagers should seek 
more assistance in the extras community... But that's another debate.

--
Pat