No answer to easy bug policy

Michael Schwendt mschwendt at gmail.com
Tue Jul 15 10:46:26 UTC 2008


On Mon, 14 Jul 2008 18:20:19 +0200, Matej Cepl wrote:

> On 2008-07-14, 15:48 GMT, Jesse Keating wrote:
> > Whilst I do like that you're trying to solve this problem, 
> > I think one big hole here is who decides that the "fix" is 
> > either easy, or right?
> 
> A bug with EasyFix keyword, and if you don't agree just change 
> the Keyword (yes, there should be some ACLs so that it cannot be 
> turned on again, but there are hopefully not that many people who 
> would be arguing with)?

There's a problem with this. It's not the only one, but just one of many
which make the entire issue not easy to solve.

The maintainers say they can't cope with the huge amount of "bugzilla
spam". Probably they don't see the EasyFix keyword, because they don't
look at the ticket at all. And an EasyFix ticket is not more important
than other tickets.

There is the NonResponsiveMaintainers policy already, which covers a very
similar process: trying to get a response from a maintainer prior to
further action.

More bugzilla mails make it worse. Do we need more bureaucracy to
determine which maintainers can't handle all their bugzilla mail? How many
maintainers are affected by huge amounts of bz mail? How many packages
are affected?

Whatever email filtering technology the maintainers use, they either don't
get to see the additional bugzilla messages (e.g. when sorting by thread)
or they skip them deliberately.

Private "ping" attacks don't make it better either, if such methods are
used to force maintainers to re-prioritise their work. Either there is
time to look at a ticket and bug, or there isn't. Why is it that direct
mail is considered more important than bugzilla mail? Bugzilla explicitly
asks reporters not to reply privately.

And what do we do? We establish processes where everyone, who is stubborn
enough, may hunt down maintainers via private mail or on IRC/IM to respond
to EasyFix bugs. A process that must be repeated everytime a new ticket is
not responded to soon enough. Instead of using bugzilla's severity flags,
private mail is the way to draw maintainer's attention? That sounds wrong
to me.

It would be helpful if registered Fedora Package Collection contributors
[with pkg cvs commit access] could set a special flag in bugzilla,
indicating that they are committed to preparing a test update. Then
package maintainers would ACK/NAK that request, knowing it is a fellow
Fedora packager who made it and who is willing to fix the bug. And we do
want to improve the communication inside the project, don't we? This isn't
about hundreds of such requests already, is it?

Do the maintainers see through all their commit mails and notifications
from koji? Or is that another source of too much mail traffic?


Adding a lot of bureaucracy for other contributors before they could apply
an EasyFix makes the process more tedious than necessary. It is another
attempt at solving the problem at the wrong end. While a maintainer seems
to be free to ignore bugs in shipped products (which is especially bad wrt
regressions), you expect volunteers to go through a tedious process, which
involves waiting several weeks plus waiting even longer to get an answer
from the maintainer. This means that the volunteer has to stay up-to-date
in that matter for more time than necessary. In a month, upstream might
publish a new minor release with further bug-fixes. And suddenly the pkg
maintainer has time again and wants to ship an upgrade instead of just a
patched pkg. That leads to wasted effort especially on the volunteer's
side.

I'd rather prefer a policy for package maintainers to respond to specially
marked tickets from fellow fedora contributors in a timely manner. And
if that results in tickets which are still not answered, timeout periods
can be applied and give contributors the opportunity to prepare a
test update (and only a test update!).




More information about the fedora-devel-list mailing list