This past Monday I held a bug triage meeting here in the office. I
haven't gotten around to publishing the summary until now because of
busy-ness and procrastination.
However, we ended the bug triage with a meta-discussion about Bug
Triages and how to make them better. Before I summarize this past
triage, here are a couple of things that we plan to do.
* IRC bug triages
We really want more community involvement. Since our community is
so spread out, Erik suggested we try to use IRC to hold the
So, this coming Monday, I'll post a list of bugs that we'll be
covering and remind you that on Tuesday at 1700UTC I will be
holding a bug triage in #wikimedia-dev on IRC. Please join us,
but you may be asked to actually FIX some bugs if you have commit
For some that is simply too late or too early, don't worry, I plan
to alternate times so that the next week it will be done 6 hours
later at 2300UTC.
If you're like me and you can't instantly compute time zone
differences for different parts of the world, try this link:
* Narrower Triages
When I first started doing the triages, I said I would pick a
theme. However, I haven't kept that up. In addition to making
our triage process more public, this coming Tuesday I'll probably
be focusing on the “theme” of Resource Loader or API issues.
Maybe once a month there will be an escalation triage where issues
from the other triages that need input across domains will be
brought up and discussed amongst those who need to be consulted.
For any issues that need immediate escalation or action by the
Foundation, I'll notify Robla or CTWoo so that they can find the
right person to act.
Now for the notes from the meeting. See http://hexm.de/3x for the
etherpad we used.
The first three bugs just need someone to do them. While no one in
the triage meeting volunteered, anyone can quickly fix these. Consider
these possibilities if you want to hack on MediaWiki this weekend:
Cannot use inputbox with some page title
Section edit links shown to anon users even if they can't edit
New 404 Page
Now, the real triaging begins:
Pages loading without formatting too often
This problem has been discussed in triage before, so I brought it
up again to get a status update.
We've been working with some users on the bug to get more
information on this problem, but we are still having trouble
replicating the problem.
Since this seems to be tied to Resource Loader, it may be fixed
during the RL sprint that the Features Team has planned for next
Import bug on several pages
On this particular bug, we decided to turn the import of these
large pages into a shell bug. Other imports that timeout like
can also be turned into shell bugs.
Since the issue of timeouts like this is not really a new problem,
we're not going to spend a lot of time working on fixing the
However, if someone is interested in a project, this would be a
good one. A couple of ideas that were tossed around: someone
could improve performance or make imports like this use the job
Extension:TitleBlacklist doesn't seem to work on Commons
After a bit of discussion, Alolita said she would follow up with
Sam and NeilK have bashed their heads against the bug, but NeilK
just told me that they'll try again in a couple of weeks. In the
meantime, if anyone else wants to have a look, go for it!
Since I was in the office this week, I used the opportunity to pester
Tomasz about Offline and Mobile projects. This provided a good chance
to talk about the Collections extension that we got from Pediapress.
PDF Export issue
The Pediapress developer (Volker Haas?) does not seem to have much
time available for fixing bugs so we discussed finding a
python-savvy volunteer developer to work on the code. If anyone
wants to take a look, check out http://code.pediapress.com/ which
should provide all the code.
iPhone Native Crash with UTF-8
Tomasz uploaded a non-crashing version of this iPhone application
this week after our meeting.
We also discussed using Krinkle's TestSwarm instance (see
http://toolserver.org/~krinkle/testswarm/) to test the new mobile
extension. Tomasz said they already have the tests written so
integrating them shouldn't be too much work.
Next we talked about some policy issues that a couple of bugs raised
Enable e-mail EnotifWatchlist on all small wiki's
Because we want to measure the amount of email this will generate
before we actually turn it on, we want to create a null mailer
that will log via UDP so we can get counts without actually
sending the emails. Alolita asked Roan to work on this.
Translate $digitTransformTable to International Digits for Hindi wikis
Priyanka and I both looked at this bug but didn't see any
consensus. Based on Erik's feedback, we've asked them to keep the
discussion open or, alternatively, point to some official
government policy on the use of a particular digit set.
Next we discussed an old feature request:
$wgUseDynamicDates should work for other languages than English
I started out by asking if we should drop the feature (since it is
obviously controversial) or adapt it for other languages.
Brion pointed out that there are issues with the current
implementation and that adding it for other languages would be a
As a result, we decided the status quo was fine and I closed the
Since then, User:Jayvdb has started a discussion on Commons:
Periodical run of currently disabled special pages
We discussed several possible ways to fix this (including Amir's
suggesting from https://bugzilla.wikimedia.org/28710 “make
frequently updated special pages update themselves automatically”)
but in the end we asked Priyanka to see how we can get these
periodical updates running on a more frequently and updating the
page with information about when it is scheduled to be run NEXT
rather than just the last time it was run.
Apologies for any inconvenience and thank you to those who have already
completed the survey. We will keep the survey open for another couple of
weeks. But, we do hope you will consider responding to the email request
below (sent 2 weeks ago).
Dr. Jeffrey Carver
University of Alabama
(v) 205-348-9829 (f) 205-348-0219
From: Jeffrey Carver [mailto:email@example.com]
Sent: Monday, June 13, 2011 11:41 AM
Subject: Participation Requested: Survey about Open-Source Software
Drs. Jeffrey Carver, Rosanna Guadagno, Debra McCallum, and Mr. Amiangshu
Bosu, University of Alabama, and Dr. Lorin Hochstein, University of
Southern California, are conducting a survey of open-source software
developers. This survey seeks to understand how developers on distributed,
virtual teams, like open-source projects, interact with each other to
accomplish their tasks. You must be at least 19 years of age to complete the
survey. The survey should take approximately 15 minutes to complete.
If you are actively participating as a developer, please consider completing
Here is the link to the survey: http://goo.gl/HQnux
We apologize for inconvenience and if you receive multiple copies of this
email. This survey has been approved by The University of Alabama IRB board.
Dr. Jeffrey Carver
University of Alabama
(v) 205-348-9829 (f) 205-348-0219
While the topic of "how Mediawiki handles URLs" is on the table, let me point
out today's Slashdot piece, which notes that ICANN is about to open up the
*to everyone*, not just commercial registries.
Contemplate, if you will:
How will MW handle a FQDN with no dots in it, when that becomes legal?
Jay R. Ashworth Baylink jra(a)baylink.com
Designer The Things I Think RFC 2100
Ashworth & Associates http://baylink.pitas.com 2000 Land Rover DII
St Petersburg FL USA http://photo.imageinc.us +1 727 647 1274
In anticipation for proper HTTPS support on Wikimedia sites, we will
be enabling protocol relative URLs. This means that things like logos,
references to resources, and interwiki links will use links like:
instead of links like:
Doing this ensures that we don't split our squid/varnish caches.
We'll enable this for test.wikipedia.org some time before enabling it
globally, so that there will be at least a short time to test