[Koha-devel] The needs signoff queue is overloaded

[Jonathan Druart]

Hi devs,

The number of patches waiting for a signoff is around 220, which is
really high to make things fluid.
I know that everybody has the same issue and that it is part of my job
to make it down but there is not a lot activities from other QAers. I
am limited when signoffing patches as it makes me loose my QA token,
so I cannot help too much. This month there were only 2 different
QAers: Katrin and me so far.

One of the main issue is that this queue is processed as a LIFO queue,
which is, given the number of signoffers, the worst queue management
we can offer.

Please take the time to signoff a couple of patches per week, and send
love to the old patches :)

I have personally submitted some works to refactor and/or improve the
code. I have focused on the admin scripts:
  Bug 14828 - Move the item types related code to Koha::ItemTypes
  Bug 15293 - Move the C4::Branch related code to Koha::Libraries
  Bug 15379 - Move the authority types related code to Koha::Authority::Type[s]
  Bug 15084 - Move the currency related code to
Koha::Acquisition::Currenc[y|ies]
  Bug 13726 - Koha::Acquisition::Bookseller should use Koha::Object
Note that the last one depends on the Tax rewrite patch set, which is
(still) pending for QA.

The goal of these patches is to 1. propose a common way to write
simple scripts, 2. add tests where they were missing, 3. (really)
start the move to the Koha namespace.
The last is the most important as we can hope having a new Object
Oriented Koha API in the next months.
During the last years the project has suffered not to get funding for
these kind of refactoring/rewrite work.
Fortunately, I can now focus on that (thanks to Paul, Brendan and
Nick), but, contrary to my earlier belief, these kinds of patches
don't appeal to signoffers. What I can understand: it is not
interesting to test that a patch set does not introduce regressions!
:)

But the thing is that I am stuck in the refactoring process. I know
this problem very well and I know that if I continue without having
the previous patches pushed, I will have to rebase this work for
months/years.

I am also afraid not to manage to find signoffers for important bug
fixes. For instance
  Bug 13618 - Prevent XSS in the Staff Client and the OPAC
has been written 6 months ago and did not reach the QA queue yet.
In the past we have had to provide security release in a rush, because
of potential XSS attacks. This patch set propose a global solution, to
avoid them completely. I have already told about this patch several
times on 1. koha-devel, 2. the bug report and 3. the IRC channel.
It will be very hard to maintain it as *a lot* of files are impacted
and I am planning to abandon it soon if I have to rebase it several
times more.

There is a critical one, waiting in the queue for almost 2 months now:
  Bug 15030 - Certain values in serials' items are lost on next edit
Data lost possible...

Some of you complain about performance issues, you can note that an
omnibus report has been opened
  Bug 15342 - Performance 3.22 - Omnibus
and some patches provided, for discussion, on the dependent bug reports.

I know that Christmas is coming and that everybody is super busy, but
hopefully 2016 will see an underloaded queue :)

Cheers,
Jonathan