t https://lists.wikimedia.org/mailman/listinfo/qa would have
to be updated to point to wikitech-l and a final mail to announce the
switch.
Antoine Musso
___
QA mailing list
QA@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/qa
Hello,
//I supposedly announced it last week but had an issue with my mailing
list setup.//
Last week and this Monday, I have overhauled the way jenkins job builder
should be used. When previously we relayed on a fork in
integration/jenkins-job-builder.git , it is now available directly under
Hello,
I have released a new version of Quibble last week - version 0.0.39.
It enables the MediaWiki REST API entry point (/rest.php), a feature
required for an API testing tool.
Thank you Clara Andrew-Wani for the patch.
The CI job for mediawiki/tools/api-testing has been updated, other jobs
Hello,
I will eventually rebuild all the WMCS instances we use for Jenkins to:
* use instances with less RAM (32G down to 24G)
* migrate from Jessie to Stretch
* rename them from "slave" to "agent"
It is currently pending available of docker-ce Debian package for Stretch.
The task:
Hello,
I have upgraded tox to 3.10.0 for most of the CI Docker containers and
CI jobs. The previous version was 2.9.1.
The changelog for tox is:
* https://tox.readthedocs.io/en/latest/changelog.html
All containers/jobs have been updated except the special cases:
* operations-dnslint:
ing/feedback you may provide!
--
Antoine Musso
quibble --run=phpunit,selenium --resolve-requires --fail-on-extra-requires
___
QA mailing list
QA@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/qa
On 04/04/2019 16:52, Antoine Musso wrote:
Hello,
I will upgrade Zuul on Monday April 8th around 7am UTC. That is merely a
maintenance upgrade with patches for long awaited fixes and will no more
block upgrading Gerrit to 2.16.
The task is https://phabricator.wikimedia.org/T208426 and list all
is querying a change would sometime yield an
unrelated patchset which caused CI to always fail :/
https://phabricator.wikimedia.org/T198968
Thank you Paladox for all the bug you caught and the preliminary testing!
--
Antoine Musso
___
QA mailing list
Hello,
I have cut a new version of Quibble: 0.0.30 , it is not yet deployed on
CI but I plan to rebuild all containers next week with that version.
I wrote a summary for the last few months of changes:
https://phabricator.wikimedia.org/phame/post/view/155/
It can notably be made to clone
Hello,
I recently had to investigate a CI failure that was quite mysterious.
PHPUnit would suddenly abort with: exit status -11
That is a segfault which involves retrieving a core dump and investigate
it with a debugger.
I wrote about my journey in a blog post which would be of interest
there. The build fails for now though :(
--
Antoine Musso
___
QA mailing list
QA@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/qa
Hello,
I have updated our Jenkins Job Builder fork in
integration/jenkins-job-builder.git . The main reason is to fix the XML
configuration for the build timeout plugin.
You will want to update your local fork which should then point to
commit a06d173e98ccea125be9f2e3b783cebeaba19767 . And
Hello,
On CI, Chromium has not been update to 0.72 yet. It crashes when using
--headless or --remote-debugging-port, at least with the package
provided by Debian Stretch ( 72.0.3626.96-1~deb9u1 ).
I have thus been holding the upgrade.
Debian testing/unstable has a newer version (
On 13/04/2018 10:48, Kunal Mehta wrote:
> I LOVE IT
>
> But really, I was failing to reproduce a test failure[1] locally and I
> gave quibble a try and it failed! After a bit of playing around I was
> able to do a full git bisect, pinpointing the problem.
>
> My main feature request would be to
Hello,
I would like to introduce Quibble, a small script intended to easily run
all MediaWiki tests with just a few keystrokes. It will ultimately be
used to power up all the CI jobs having to deal with MediaWiki.
During the Vienna hackathon (2017), we had a lot of discussions
regarding Docker
h:
- given the build curl the build artifacts for a file like report.txt
- aggregate the different files / prettify
- write a new comment to the Gerrit change
I am not sure how doable it is though :(
--
Antoine Musso
___
QA mailing list
QA@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/qa
On 02/02/2018 09:51, Kunal Mehta wrote:
>
> Ideally the "stayed the same" could be split to a third state, so we can
> properly congratulate people when coverage goes up. And somehow to
> expose the actual % change in Gerrit.
In Jenkins that is the UNSTABLE state. Typically when a previous build
On 15/12/2017 17:11, Antoine Musso wrote:
> Hello,
>
> Most of the npm test jobs are now running on Docker. On Monday I will
> upgrade the Docker container to use Chromium 63 instead of 62.
>
> Non npm jobs (eg Mediawiki QUnit tests) are still on Jessie with
> Chromium 57.
stretch.
--
Antoine Musso
___
QA mailing list
QA@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/qa
er in #wikimedia-releng to revert [1] to put everything back
> to the way it was.
>
> [1] https://gerrit.wikimedia.org/r/#/c/387500/
Well done Kunal!
I also started migrating jobs running 'tox' toward Docker containers. So
far they all pass just fine!
--
Antoine Musso
_
Hello,
On deployment-prep and integration, I have removed Salt entirely in
favor of Cumin.
It is mostly similar but way better to keep the inventory, make sure you
reach all the instances. Cumin comes with an elaborate query scheme.
One requires sudo access to be able to use it since commands
On 04/09/2017 22:08, Antoine Musso wrote:
Hello,
CI has kept Ubuntu Trusty instance solely for Zend PHP 5.5. Last week I
managed to build a rough package for Debian Jessie and got them
provisioned on the instances.
Hence I have started migrating jobs from Trusty to Jessie, testing them
On 05/09/2017 07:21, Legoktm wrote:
On 09/04/2017 01:08 PM, Antoine Musso wrote:
CI has kept Ubuntu Trusty instance solely for Zend PHP 5.5. Last week I
managed to build a rough package for Debian Jessie and got them
provisioned on the instances.
Hence I have started migrating jobs from
Hello,
CI has kept Ubuntu Trusty instance solely for Zend PHP 5.5. Last week I
managed to build a rough package for Debian Jessie and got them
provisioned on the instances.
Hence I have started migrating jobs from Trusty to Jessie, testing them
on all affected repositories as I am doing
:
* recreate the tox virtualenv, it should be sufficient to just delete
the .tox directory.
* reinstall: pip install --user .
--
Antoine Musso
___
QA mailing list
QA@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/qa
Hello,
Yesterday labs instances have lost access to the LDAP authentication
server. The internal certificate authority had expired which prevents
the secure connection from establishing.
For instances having puppet running properly, the certificate has been
refreshed. In some case we had to
Hello,
Following a reboot of the beta cluster instances that host the MariaDB
databases, the instances are unable to come back up. Most probably due
to a faulty DNS config for some reason.
So meantime, the beta cluster is mostly unavailable :-(
https://phabricator.wikimedia.org/T168569
--
Hello,
I have updated our Jenkins Job Builder fork in
integration/jenkins-job-builder.git to catch up with a latest upstream
changes and add a custom patch to it:
https://review.openstack.org/#/c/471030/
That fix the conditional-step plugin when several steps are used. I
required it for
Hello,
Changes made to integration/config required to be fast forward only.
I have finally made Gerrit to attempt a rebase for us and allow content
to be transparently merged.
From now on when one CR+2 and tests pass, Gerrit will automatically
rebase the change and merge it. Thus we no more
Hello,
I have updated our Jenkins Job Builder fork in
integration/jenkins-job-builder.git to catch up with latest upstream
changes.
The update is force updated with master pointing to:
edebce7f98a8e041537b9f0569a99ca9050a6066
I have updated all the jobs and the log is kept in T162674
Hello,
PhantomJS 2.1.1 is available in jessie-backports since March 8th and is
now available in the jessie Jenkins slaves.
Hopefully repositories relying on it would be a bit faster since they
would no more have to download it.
--
Antoine "hashar" Musso
Hello,
When running skins structure tests or selenium tests, we had a need to
clone a skin and have it properly loaded by MediaWiki. However we had
no support to do up until now.
The way it is implemented, one edits the 'dependencies' dictionary in
integration/config
Hello,
I have granted Tobias Gritschacher (WMDE) Code-Review +2 to the
integration repositories. He has been involved in QA since ever so it
kind of make sense :-}
--
Antoine "hashar" Musso
___
QA mailing list
QA@lists.wikimedia.org
On 26/10/16 04:46, Legoktm wrote:
> Note: Stephen can not deploy Zuul changes though. That currently
> requires shell access on the server. Eventually we will get rid of that
> requirement.
Is there any reason we can't give Stephen access to gallium?
Hello,
I quickly chatted with Stephen,
Hello,
I have granted Stephen Niedzielski Code-Review +2 for the
integration/config repo.
Stephen has been instrumental in baby sitting the Jenkins jobs for the
Android mobile app :] He already deploy and live check them on
production, hence there is no need to further hold changes. I
Hello,
A few hours ago I have changed the deployment server on the beta cluster
to be deployment-mira. It is based on Jessie (instead of Trusty).
That is the achievement of a lot of work with Elukey and Moritz. We
caught lot of issues but it seems to be almost in good shape right now.
Some
On 27/07/16 18:34, Loic Dachary wrote:
Hi,
I'd like to get tests configured for a particular gerrit repository hosting a
wikidata bot. Is there a document or an example I could read to know more about
it ? The bot I'm writing is about 50 lines of code[1] but I'd like to go in the
right
/package_builder/templates/pbuilderrc.erb for
the change https://gerrit.wikimedia.org/r/#/c/300830/
It is enabled again, sorry.
Antoine Musso
___
QA mailing list
QA@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/qa
is to have all generated material under
the same domain doc.wikimedia.org and have integration.wikimedia.org
solely to expose the backend services (Jenkins/Zuul). That will let us
move each domains to be hosted on different hosts later on.
--
Antoine Musso
On 08/06/16 18:47, Antoine Musso wrote:
Le 08/06/2016 à 15:02, Antoine Musso a écrit :
The operation team has worked hard this European morning to backup
files, investigate the raid issue and setup a new host.
We are in the process of reinstalling everything on the new host and
bring back
Le 08/06/2016 à 06:58, Greg Grossmeier a écrit :
> There is a possible disk failure of our main CI host, 'gallium'. It is
> being investigated at:
> https://phabricator.wikimedia.org/T137265
>
> We are very sorry for the inconvenience.
>
> Major updates will be sent to the lists with minor
Hello,
I have refreshed integration/jenkins-job-builder.git repo to upstream
version 1.5.0 and rebased a couple patches on top of it.
The new commit is 10f2bcd . Please refresh your local copy.
--
Antoine "hashar" Musso
___
QA mailing list
Hello,
Ganglia is a monitoring tool which Wikimedia uses in production but is
not available on labs.
On the beta cluster and CI infrastructures, I have mass purged the
packages and removed /etc/ganglia/ directories entirely.
The result is that whenever a puppet class ends up trying to inject a
Hello,
Thanks to Alex Monk, the Jenkins jobs that updates the beta cluster are
now sending mail notifications to a dedicated mailing list. Hence one
might want to subscribe to betacluster-alerts:
https://lists.wikimedia.org/mailman/listinfo/betacluster-alerts
Ref:
Hello,
Zeljko needed support in JJB for the Yaml Axis Plugin which I wrote and
proposed upstream. In the process I have upgraded our fork in
integration/jenkins-job-builder to the latest upstream master then
cherry picked the Yaml axis related patches.
Please upgrade your local JJB
Hello,
It was no more possible to authenticate on the beta cluster from roughly
11:50am UTC till 18:50 UTC.
The root cause was a configuration change having a nasty side effect
which caused the backend storing the user sessions to no more be reachable.
Related tasks in order of investigation:
Le 11/02/2016 18:02, Antoine Musso a écrit :
> Hello,
>
> I was looking at what gate-and-submit run for mediawiki/core and I think
> there are unnecessary jobs/duplicates. Specially looking at phpunit:
>
> mediawiki-phpunit-php55
> mediawiki-phpunit-parsertests-php55
repository, I am all for it.
cheers,
--
Antoine Musso
Le 05/02/2016 19:40, Guillaume Lederrey a écrit :
> Message below cross posted
> from o...@lists.wikimedia.org
> <mailto:o...@lists.wikimedia.org>.
>
> Seems that the discussion might be interesti
Hello,
I was looking at what gate-and-submit run for mediawiki/core and I think
there are unnecessary jobs/duplicates. Specially looking at phpunit:
mediawiki-phpunit-php55
mediawiki-phpunit-parsertests-php55
I have extracted the parsertest for Zend to speed up the phpunit
reporting. HHVM
Hello,
For a month or so, the Beta cluster user sessions might have suffered
from redis server being down / flapping:
https://phabricator.wikimedia.org/T124677
It should be all solved now. I have retriggered manually a bunch of
browser tests and they started passing again.
--
Antoine "hashar"
Le 18/01/2016 07:58, Legoktm a écrit :
> Hi,
>
> In order to prepare for raising the MediaWiki core version requirement,
> which'll happen at some point in the near future (I hope!), I have
> renamed[1] all of the uses of "zend" in the jenkins and zuul config to
> the more specific "php53". This
Hello,
The MediaWiki Qunit jobs run by CI were semi randomly failiveng for the
last three days. The reason is we had set $wgHTTPProxy to a service
which is no more usable and a timeout race condition would cause the
jobs to fail.
I have disabled the proxy configuration and the jobs are now
Hello,
Zeljko wrote about it another thread:
We have decided to merge the CI and browser tests checkin reusing the
later time slot. Hence:
- Merging CI and browser tests meeting
- agreed: the new merged meeting will be on Tuesdays, 16-17 UTC
Or
5pm-6pm CET
8am-9am PST
--
Antoine
Hello,
I have update our for of JJB in integration/jenkins-job-builder to
upstream version 4eb33b2
Due to some issue with python-jenkins module, I have added a patch to
pin it to 0.4.8:
https://gerrit.wikimedia.org/r/#/c/255182/
Please refresh your local install to point to 816e1c1b11e
Or
Le 28/10/2015 22:27, Jon Robson a écrit :
> QA team
> We're struggling to identify why this is happening and it's spamming our
> inboxes
> https://gerrit.wikimedia.org/r/246801 seems to fix the issue for unknown
> reasons.
> Any help you can provide on the patch would be much appreciated.
Hello,
Hello,
Looking for some advices regarding maintaining a cache store for the
various package managers. The Nodepool instances have cold caches and
that causes npm/pip/gem/composer/gradle/maven to fetch everything over
the internet.
When a job start, I would like the cache to be populated from
Le 13/10/2015 18:57, Željko Filipin a écrit :
> I have sent this to the list a few days ago, but I have forgot to change
> the subject to something more descriptive. Resending with a better subject.
>
> Željko
>
> On Fri, Oct 9, 2015 at 2:44 PM, Željko Filipin
>
Le 05/10/2015 18:28, Greg Grossmeier a écrit :
> Antoine might have a better idea, but for US hours there weren't any
> major issues, CI-wise, that I remember.
>
> I don't see anything obvious here:
> https://tools.wmflabs.org/sal/releng?p=0==2015-09-23
> nor here:
Hello,
We have skip last week meeting (Sep 22th) but just completed this week
meeting. Please find the minutes on
https://www.mediawiki.org/wiki/Continuous_integration_meetings/2015-09-29
We talked about:
* Update CI scaling (more jobs are being migrated)
* Jenkins console and collapsible
Le 14/09/2015 23:01, Antoine Musso a écrit :
> Hello,
>
> The next continuous integration checkin is tomorrow Tuesday 15th at
> 14:00 UTC (16:00 CEST) in #wikimedia-office *and Google hangouts*:
>
> https://plus.google.com/hangouts/_/wikimedia.org/ci-weekly
>
> You wou
Le 01/09/2015 17:12, Antoine Musso a écrit :
> Hello,
>
> The next meeting will be on Tuesday September 8th at 14:00 UTC (16:00
> CEST) in #wikimedia-office *and with Google hangouts*:
>
> https://plus.google.com/hangouts/_/wikimedia.org/ci-weekly
>
> Next
Hello,
This is a mail I sent a month or so ago to the Release Engineering
internal mailing list. Kunal Mehta pointed out it is worth publishing
for later reference.
It follows up the previous mail 'Jenkins API and Zuul Gearman'
https://lists.wikimedia.org/pipermail/qa/2015-September/002372.html
Hello,
Paladox found a rather nice issue in Zuul which is that when you +2 two
patches on repository that is fast-forward only, only the first change
is merged in and the other is rejected.
We are using the fast-forward merge strategy on integration/config.git
repo since we want to manually
Hello,
I forgot to announce today checkin on 2015-09-01. That was Antoine, Jan
and Zeljko.
We talked about:
* Wikidata browser test job failling
* "CI isolation" project renamed to "CI scaling"
* CI scaling update
* Wikidata and composer
* Jobs compatibility with older release branches
* Unique
Le 01/09/2015 17:11, Antoine Musso a écrit :
> Hello,
>
> I forgot to announce today checkin on 2015-09-01. That was Antoine, Jan
> and Zeljko.
>
> We talked about:
>
> * Wikidata browser test job failling
> * "CI isolation" project renamed to "CI s
Hello,
The next meeting will be on Tuesday September 8th at 14:00 UTC (16:00
CEST) in #wikimedia-office *and with Google hangouts*:
https://plus.google.com/hangouts/_/wikimedia.org/ci-weekly
Next meeting agenda:
https://www.mediawiki.org/wiki/Continuous_integration/Meetings/2015-09-08
Please
Le 20/08/2015 01:52, Greg Grossmeier a écrit :
Hello all,
This is a heads up and discussion starter about the upcoming Gerrit
Cleanup Day (see: https://phabricator.wikimedia.org/T88531).
tl;dr: On Wed Sept. 23rd there will be a concerted effort to reduce the
backlog of open patchsets in
Le 28/07/2015 16:57, Antoine Musso a écrit :
Hello,
Due to vacations and travelling, the next CI check-in will be on Tuesday
August 25th at 4pm CET (2pm UTC).
Meeting agenda is at:
https://www.mediawiki.org/wiki/Continuous_integration/Meetings/2015-08-25
The meeting has been made
Hello,
The beta cluster puppet master is now deployment-puppetmaster and is
running Ubuntu Trusty.
https://phabricator.wikimedia.org/T106649
--
Antoine hashar Musso
___
QA mailing list
QA@lists.wikimedia.org
Hello,
Marko Obrovac enquired to learn more about Jenkins and Zuul for the
various MediaWiki services (parsoid, graphoid, restbase etc). Those
repositories share a very common structure:
- javascript
- dependencies declared via npm
- a deploy.git repo for production usage
We have just
Hello,
Following the overnight labs NFS outage, the CI Jenkins slaves in labs
no more uses NFS.
So /data/project is gone, we never used it anyway.
Homedirs are now local to the instance. If you want your favourite dotfiles:
operations/puppet.git
./modules/admin/files/home/your username/...
Hello,
I have created a very dumb Jenkins slave for Jessie. It is not able to
run all our CI jobs since we are missing a bunch of packages, but can be
used for corner case that would run without all the ci packages
dependencies.
The reason is role::ci::slave::labs includes too many things which
Le 08/06/2015 16:29, Antoine Musso a écrit :
Hello,
The next meeting will be on Tuesday June 9th at 14:00 UTC (16:00 CEST)
in #wikimedia-office *and with Google hangouts*:
https://plus.google.com/hangouts/_/wikimedia.org/ci-weekly
Next meeting agenda:
https://www.mediawiki.org/wiki
Hello,
The next meeting will be on Tuesday June 9th at 14:00 UTC (16:00 CEST)
in #wikimedia-office *and with Google hangouts*:
https://plus.google.com/hangouts/_/wikimedia.org/ci-weekly
Next meeting agenda:
https://www.mediawiki.org/wiki/Continuous_integration/Meetings/2015-06-09
Please add
Le 03/06/2015 11:05, Antoine Musso a écrit :
Hello,
On Monday we updated the Jenkins git plugin which comes with some
slightly different XML config format.
We have updated integration/jenkins-job-builder.git with a more recent
version that supports the new git format and I have adjusted
Hello,
On Monday we updated the Jenkins git plugin which comes with some
slightly different XML config format.
We have updated integration/jenkins-job-builder.git with a more recent
version that supports the new git format and I have adjusted our yaml
configuration files:
Le 13/05/2015 16:52, Antoine Musso a écrit :
The next meeting will be on Tuesday May 19th at 14:00 UTC (16:00 CEST)
in #wikimedia-office. The intended audience is anyone having an interest
in continuous integration and having actions run automatically on code
changes.
Next meeting agenda
Hello,
The next meeting will be on Tuesday May 19th at 14:00 UTC (16:00 CEST)
in #wikimedia-office. The intended audience is anyone having an interest
in continuous integration and having actions run automatically on code
changes.
Next meeting agenda
tests to test everything, just
need a few end user scenario. The rest should be tested via integration
tests.
If we could get developers from over teams to help on speeding up
cucumber that would be already a huge improvement!
cheers,
--
Antoine Musso
On 27/03/15 12:58, Antoine Musso wrote:
Hello,
Starting this Monday, Timo and I will hold a weekly meeting for
Continuous Integration.
Next meeting is Monday March 30th, at 13:00 UTC in #wikimedia-office
Hello,
The next meeting will be on *Tuesday* April 7th at 13:00 UTC due to
Easter
Le 02/02/2015 19:16, Legoktm a écrit :
Hi,
We're currently working on converting[1] all WMF-deployed extensions to
use the new extension registration[2] system that was introduced in 1.25.
I filed T86359[3] to track the updates we need to make to jenkins to
accommodate the new system:
*
Le 06/01/2015 02:55, Chris McMahon a écrit :
In Jenkins I clicked prepare for shutdown, then cancelled the
operation, hoping to unstick jenkins. I saw the beta-scap-eqiad job run
after that.
Then I disabled/enabled gearman
I made a trivial update to a gerrit patch but did not see zuul
Le 18/12/2014 09:11, Yuvi Panda a écrit :
Has been for about 6 days, is that still intentional?
While at it, does anyone have any idea what this instance is for?
--
Antoine hashar Musso
___
QA mailing list
QA@lists.wikimedia.org
Hello,
The Parsoid version used on the beta cluster has been stalled since Oct
28th. Additionally, the cache in front of it was down.
We are now back to an up-to-date and automatically updated Parsoid
version which is functional.
Details:
In October, parsoid switched from node 0.8 to 0.10
Le 29/10/2014 16:28, Antoine Musso a écrit :
Hello,
We run the browser tests twice per day and keep the history of builds ad
vitam.
A side effect is that the Jenkins view for browser tests takes a couple
minutes to load on first hit. That is because Jenkins lamely load the
while history
Le 29/10/2014 16:28, Antoine Musso a écrit :
Hello,
We run the browser tests twice per day and keep the history of builds ad
vitam.
A side effect is that the Jenkins view for browser tests takes a couple
minutes to load on first hit. That is because Jenkins lamely load the
while history
Hello,
The Varnish cache that serves upload.beta.wmflabs.org (ie uploaded
files) was down for some time. I have restarted Varnish there.
More details on the task:
https://phabricator.wikimedia.org/T75922
--
Antoine hashar Musso
___
QA mailing
Le 20/11/2014 01:38, Dan Duvall a écrit :
Ironically enough, it has failed the Travis CI check because of an
overzealous `GuardClause` check. (Seriously, who can't read an `if`
statement ... :)
http://www.rubydoc.info/github/bbatsov/rubocop/Rubocop/Cop/Style/GuardClause
Well, one could name
Le 20/11/2014 01:38, Dan Duvall a écrit :
snip
IMO, we shouldn't make rubocop voting until we iron out all the
overzealousness of RuboCop. I've submitted a pull request to help us
create base configuration that's more congruent with the actual style
guide.[2]
snip
[2]
Le 19/11/2014 19:14, Krinkle a écrit :
It seems I can't access Build Artefacts from jobs run on labs.
I've almost finalised migration of mediawiki/qunit jobs to labs slaves,
and applied it to mediawiki-core-qunit and mwext-VisualEditor-qunit, but
it seems there's no longer the option to
Le 17/11/2014 07:15, Yuvi Panda a écrit :
What exactly is deployment-lucid-salt used for? It's oranging up my shinken!
:)
It has been created by Ariel apergos Glenn. Probably related to some
salt migration work going on in production.
--
Antoine hashar Musso
Le 13/11/2014 15:54, Željko Filipin a écrit :
I think it is the time to make rubocop jobs voting. I did not notice any
problems. Does anybody disagree?
Reviewing the changes to oojs/ui there were some oddities:
Style/WordArray:
Le 06/11/2014 07:05, S Page wrote:
I'm not sure that a nightly updating once a day is much benefit. By the
time we figure out master is broken and switch to it as a fallback the
fix is nearly ready.
The idea behind nightly is to provide a wiki that does not unexpectedly
break because of a
Le 30/10/2014 18:26, Gabriel Wicke a écrit :
For projects like RESTBase it would also be great to be able to use
Gerrit with Travis pre-merge testing. Does anybody have experience
with hooking up Travis with Gerrit or Jenkins?
Zuul is a scheduler that listens for Gerrit events and trigger
Hello,
We run the browser tests twice per day and keep the history of builds ad
vitam.
A side effect is that the Jenkins view for browser tests takes a couple
minutes to load on first hit. That is because Jenkins lamely load the
while history of all jobs to generate the [dashboard].
S Page
Le 23/10/2014 13:17, Željko Filipin a écrit :
If you are familiar with RuboCop, I could use some help.
http://stackoverflow.com/q/26508803/17469
Željko
That is an issue in rubocop. It traverses the tree hierarchy using:
Dir[somepath/**/*]
Which is the equivalent of:
Dir.glob(
Le 21/10/2014 22:55, Antoine Musso a écrit :
Le 20/10/2014 21:42, Yuvi Panda a écrit :
Hello!
I've been slowly ramping up Labs monitoring, and currently have hosts
up/down availability on shinken (shinken.wmflabs.org -
username/password: guest/guest).
deployment-soa-cache01 seems
Hello,
The upstream Jenkins Job Builder has landed a change that will abort the
run whenever it find duplicate jobs definitions. We do use that feature
to override some MediaWiki extensions related jobs.
Since duplication detection is turned on by default, please adjust your
jenkins_jobs.ini
Le 10/10/2014 21:32, S Page a écrit :
On Fri, Oct 10, 2014 at 7:45 AM, Antoine Musso
hashar+...@free.fr
mailto:hashar+...@free.fr wrote:
Before releasing it, can we enforce the use of ruby 2.0 and no more
depends on 1.9.x?
Ubuntu 12.04 ships Ruby 1.8 and Ubuntu 14.04 ships
Le 15/10/2014 20:27, Matthew Flaschen a écrit :
It looks like there is a new Ruby style checker called Rubocop. It is
running as a non-voting job on Flow (and I assume other repositories).
If this is considered ready to run, please document this at
1 - 100 of 169 matches
Mail list logo