[Wikitech-l] Re: Sponsored-by

[Lars Wirzenius]

On Thu, May 27, 2021 at 08:04:23PM +0200, David Caro wrote:
> I'm willing to help with some scripts/hooks to set it up if needed
> (though I'm suspicious they already exist xd).

I added a line to my git commit template, and it suffices for me so
far. I'm sure this can be made smoother.

-- 
WMF release engineering team | he/him or they/them
"Imagine a world in which every single human being can freely share in
the sum of all knowledge."
___
Wikitech-l mailing list -- wikitech-l@lists.wikimedia.org
To unsubscribe send an email to wikitech-l-le...@lists.wikimedia.org
https://lists.wikimedia.org/postorius/lists/wikitech-l.lists.wikimedia.org/

[Wikitech-l] Sponsored-by

[Lars Wirzenius]

I figure this might be of interest to some here: we could make
sponsorship of contributions more evident by adding a git commit
pseudo-header:

Sponsored-by: Name Of Sponsoring Entity

Obviously this would be optional, and only used by those who want to.

Spelled out in a little more detail:
https://blog.liw.fi/posts/2021/05/26/sponsored-by/

-- 
WMF release engineering team | he/him or they/them
"Imagine a world in which every single human being can freely share in
the sum of all knowledge."
___
Wikitech-l mailing list -- wikitech-l@lists.wikimedia.org
To unsubscribe send an email to wikitech-l-le...@lists.wikimedia.org
https://lists.wikimedia.org/postorius/lists/wikitech-l.lists.wikimedia.org/

[Wikitech-l] The deployment of 1.37.0-wmf.3: rocky, but good now

[Lars Wirzenius]

This is a summary of this week's deployment of the 1.37.0-wmf.3 branch
of MediaWiki and its extensions (also known as "the train"). The
primary person in charge this week was me, Lars Wirzenius, with Jeena
Huneidi as backup, both from the Release Engineering team.

The summary task for this week is
https://phabricator.wikimedia.org/T278347

This week was a little more rocky than usual. I achieved three
personal records: largest number of rollbacks in a week (two), largest
number of errors in logstash after a deployment (63367), and largest
number of deer seen from my window while conducting the train (eight,
in pairs on three different days).

The new version is running all sites. There were a total of six
blockers reported and resolved:

- T281180 - DiscussionTools: Precondition failed: This Title instance
  does not represent a proper page, but merely a link target  
  https://phabricator.wikimedia.org/T281180
  
- T281184 - PHP Fatal Error: Access level to
  CentralAuthPrimaryAuthenticationProvider::$userNameUtils must be
  protected (as in class
  MediaWiki\Auth\AbstractPasswordPrimaryAuthenticationProvider) or
  weaker  
  https://phabricator.wikimedia.org/T281184
   
- T281226 - PHP Notice: Only variables should be assigned by reference  
  https://phabricator.wikimedia.org/T281226
  
- T281361 - TypeError: Argument 2 passed to
  
Wikibase\Client\DataAccess\Scribunto\WikibaseLanguageIndependentLuaBindings::trackUsageForSitelink()
  must be an instance of Wikibase\DataModel\Entity\ItemId, instance of
  Wikibase\MediaInfo\DataMo  
  https://phabricator.wikimedia.org/T281361
  
- T281405 - Wikimedia\Rdbms\DBQueryError: Error 1054: Unknown column
  'img_actor'   
  https://phabricator.wikimedia.org/T281405
  
- T281480 - SqlBlobStore no longer caching blobs (DBConnectionError
  Too many connections)  
  https://phabricator.wikimedia.org/T281480
  
The last task is still open, as it was worked around quickly, on
Thursday, to avoid a service outage (read: taking down all sites), but
is waiting for a proper fix later.

As usual, a whole bunch of people helped to find, triage, analyze,
fix, or work around problems this. Release Engineering thanks
everyone, without help we wouldn't be able to deploy MediaWiki.

- Aaron Schulz
- Addshore
- Bartosz Dziewoński
- Daimona
- DannyS712
- Gergő Tisza
- Giuseppe Lavagetto
- Jaime Crespo
- Jakob Warkotsch
- James D. Forrester
- Kosta Harlan
- Lucas Werkmeister
- Martin Urbanec
- Pchelolo
- Taavi Väänänen
- Timo Tijhof
- Tim Starling

There may have been other people, and if so, I apologize for not
including them on the list above.

Have a good weekend. Be well. Be safe.

For more information, please see:

- https://phabricator.wikimedia.org/T278347
- https://wikitech.wikimedia.org/wiki/Heterogeneous_deployment/Train_deploys
- https://versions.toolforge.org/

--
WMF release engineering team | he/him or they/them
"Imagine a world in which every single human being can freely share in
the sum of all knowledge."

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

[Wikitech-l] [Train] 1.37.0-wmf.3 status update

[Lars Wirzenius]

The 1.37.0-wmf.3 version of MediaWiki is blocked[0].

The new version is deployed to {group(s){0,1,2}}[1], but can proceed no
further until these issues are resolved:

* T281361 TypeError: Argument 2 passed to
  
Wikibase\Client\DataAccess\Scribunto\WikibaseLanguageIndependentLuaBindings::trackUsageForSitelink()
  must be an instance of Wikibase\DataModel\Entity\ItemId, instance of
  Wikibase\MediaInfo\DataMo  
  

Once these issues are resolved train can resume. If these issues are
resolved on a Friday the train will resume Monday.

Thank you for your help resolving these issues!

-- Your humble train toiler

[0]. 
[1]. 

-- 
WMF release engineering team | he/him or they/them
"Imagine a world in which every single human being can freely share in
the sum of all knowledge."

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

[Wikitech-l] Summary of this week's deployment of 1.37.0-wmf.1: all good!

[Lars Wirzenius]

This is a summary of this week's deployment of the 1.37.0-wmf.1 branch
of MediaWiki and its extensions (also known as "the train"). The
primary person in charge this week is Jeena Huneidi, with Dan Duvall
as backup, both from the Release Engineering team.

The summary task for this week is
https://phabricator.wikimedia.org/T278345

This was the first week of the 1.37.0 series. The 1.36.0 release is
being made separately, and our sites are moving on to run weekly
releases of the new series. It's not a momentous change. We start a
series couple of times a year. The version number changes, but
otherwise the changes this week are fairly mundane and usual.

(There's no train deployment next week, due to a holiday on Thursday.)

The new version is running all sites. There were a total of three
blocker reported and resolved:

- T279008 - mw-parser-output now clears after block, which is a
  noticeable change for Category pages where infoboxes used to render
  floated alongside the automatic content  
  https://phabricator.wikimedia.org/T279008
  
- T279113 - Unable to parse title at new Title in Special:MediaSearch  
  https://phabricator.wikimedia.org/T279113
  
- T280082 - Wikimedia\Rdbms\DBQueryError: Error 1146: Table
  'mediawikiwiki.discussiontools_subscription' doesn't exist
  (10.64.16.7)  
  https://phabricator.wikimedia.org/T280082

As usual, a whole bunch of people helped to find, triage, analyze,
fix, or work around problems this. Release Engineering thanks
everyone, without help we wouldn't be able to deploy MediaWiki.

- Anne Tomasevich
- Bartosz Dziewoński
- Carly Bogen
- Cormac Parle
- Innovative
- JJMC89
- Jon Robson
- Lucas Werkmeister
- Mike_Peel
- Rummana Yasmeen
- Tacsipacsi
- Timo Tijhof
- TTO
- Umherirrender
- Urbanecm
- Zabe

There may have been other people, and if so, I apologize for not
including them on the list above.

Have a good weekend. Be well. Be safe.

For more information, please see:

- https://phabricator.wikimedia.org/T278345
- https://wikitech.wikimedia.org/wiki/Heterogeneous_deployment/Train_deploys
- https://versions.toolforge.org/

-- 
WMF release engineering team | he/him or they/them
"Imagine a world in which every single human being can freely share in
the sum of all knowledge."

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

[Wikitech-l] Summary of this week's deployment of 1.36.0-wmf.38: not there yet

[Lars Wirzenius]

(This is almost a repeat of last week's train summary. It's again a
short week and the train is still running.)

This is a summary of this week's deployment of the 1.36.0-wmf.38
branch of MediaWiki and its extensions (also known as "the train").
The primary person in charge this week is Dan Duvall, with Mukunda
Modell as backup, both from the Release Engineering team.

The summary task for this week is
https://phabricator.wikimedia.org/T278344 .

This week's deployment is still ongoing. The train is at group 1, and
will hopefully move to group 2 later today, but I'm writing this ahead
of time due to me being in an unfortunate time zone (hello from the
future!). As of writing this email, there are no blocker tasks. In
fact, there doesn't seem to have been any this week. None I say. NONE!
Well okay, a couple of things were added tentatively, but dropped as
blockers after further investigation.

Amir Sarabadani reported a risky change in this train. Special thanks!
Forewarned is better than surprised.

As usual, a whole bunch of people helped to find, triage, analyze,
fix, or work around problems this. Release Engineering thanks
everyone, without help we wouldn't be able to deploy MediaWiki.

- Amir Sarabadani (WMDE)
- Anne Tomasevich
- Bartosz Dziewoński
- Carly Bogen
- Cormac Parle
- C. Scott Ananian
- Jon Robson
- Lucas Werkmeister
- Umherirrender

There may have been other people, and if so, I apologize for not
including them on the list above.

Have a good weekend. Be well. Be safe.

For more information, please see:

- https://phabricator.wikimedia.org/T278344
- https://wikitech.wikimedia.org/wiki/Heterogeneous_deployment/Train_deploys

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

[Wikitech-l] Summary of this week's deployment of 1.36.0-wmf.36: not there yet

[Lars Wirzenius]

This is an early summary of this week's deployment of the
1.36.0-wmf.37 branch of MediaWiki and its extensions (also known as
"the train"). The primary person in charge this week is Mukunda
Modell, with Antoine Musso as backup, both from the Release
Engineering team.

The summary task for this week is
https://phabricator.wikimedia.org/T278343

This week's deployment is still ongoing. The train is at group 1, and
will hopefully move to group 2 later today, but I'm writing this ahead
of time due to a time zone issue (hello from the future!). There were
some problems with getting a late patch included in the train branch,
initially, but the train got started pretty much on time. As of
writing this email, there is one blocker task:

- T279008: mw-parser-output now clears after block, which is a
  noticeable change for Category pages where infoboxes used to render
  floated alongside the automatic content  
  https://phabricator.wikimedia.org/T279008

There has been some back and forth about tasks as issues have been
investigated and various tasks have either been raised to being
blockers or demoted from that status without the issue being resolved.
For details, please see the summary task. (I'm trying to keep these
summaries short, but each week could be turned into a suspense novel.)

As usual, a whole bunch of people helped to find, triage, analyze,
fix, or work around problems this. Release Engineering thanks
everyone, without help we wouldn't be able to deploy MediaWiki.

- Amir Sarabadani
- Bartosz Dziewoński
- C. Scott Ananian
- Daniel Kinzler
- DannyS712
- Dylsss
- James D. Forrester
- João Alexandre Peschanski
- Jon Robson
- Mike_Peel
- Pchelolo
- RhinosF1
- Subramanya Sastry
- Timo Tijhof
- Urbanecm

There may have been other people, and if so, I apologize for not
including them on the list above.

For more information, please see:

https://phabricator.wikimedia.org/T278343
- https://wikitech.wikimedia.org/wiki/Heterogeneous_deployment/Train_deploys

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

[Wikitech-l] Summary of this week's deployment of 1.36.0-wmf.36

[Lars Wirzenius]

This is a summary of this week's deployment of the 1.36.0-wmf.36
branch of MediaWiki and its extensions. The primary person in charge
this week is Antoine Musso, with Ahmon Dancy as backup, both from the
Release Engineering team.

The summary task for this week is  
https://phabricator.wikimedia.org/T274940

This week's deployment was not successful. The train got to group 2,
but had to be rolled back all the way to group 0, which is where it is
at right now. The current problems are:

- T278478: FlaggedRevs: PHP Notice: Undefined index: status  
  https://phabricator.wikimedia.org/T278478

- T278429:
  Translate\PageTranslation\ParserOutput::sourcePageTextForRendering()
  must be an instance of Language, instance of StubUserLang given  
  https://phabricator.wikimedia.org/T278429

- T277795: User not found by actor ID: [id]  
  https://phabricator.wikimedia.org/T277795

The general policy is to NOT deploy on Fridays. If we deploy something
on a Friday and the sites break on Saturday, multiple people may need
to be alerted to diagnose and fix or work around the problem. Thus,
when fixes are found, we tend to deploy them on Monday instead. Then,
on Tuesday, the next train can start.

Issues that have already been dealt with along the week:

- T277722: TypeError: this._languageCodes is undefined at
  getLanguageNameMap  
  https://phabricator.wikimedia.org/T277722

- T278375: Class 'GlobalUsageHooks' not found  
  https://phabricator.wikimedia.org/T278375

- T278379: Argument 1 passed to
  ProofreadPage\Index\IndexTemplateStyles::__construct() must be an
  instance of Title, null given, called in
  
/srv/mediawiki/php-1.36.0-wmf.36/extensions/ProofreadPage/includes/Page/PageContent.php
  on line 284  
  https://phabricator.wikimedia.org/T278379

- T278400: PHP Warning: array_key_exists(): The first argument should
  be either a string or an integer  
  https://phabricator.wikimedia.org/T278400

We saw some people highlight especially risky patches in the week's
summary task. That was very welcome! Thank you, Amir Sarabadani and
Roan Kattouw.

As usual, a whole bunch of people helped to find, triage, analyze,
fix, or work around problems this. Release Engineering thanks
everyone, without help we wouldn't be able to deploy MediaWiki.

- Abijeet Patro
- Adam_WMDE
- Amir Sarabadani
- Asartea
- Daniel Kinzler
- DannyS712
- James Forrester
- Johan Jönsson
- Jon Robson
- Klaas Skelte van der Werf
- Max Binder
- Niklas Laxström
- Pchelolo
- Roan Kattouw
- Silvan_WMDE
- Sohom Datta
- Timo Tijhof
- Tpt
- Urbanecm

There may have been other people, and if so, we apologize for not
including them on the list above.

For more information, please see:

- https://phabricator.wikimedia.org/T274940
- https://wikitech.wikimedia.org/wiki/Heterogeneous_deployment/Train_deploys

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

[Wikitech-l] Summary of this week's deploymentt of 1.36.0-wmf.35

[Lars Wirzenius]

This is a summary of this week's deployment of the 1.36.0-wmf.35
branch of MediaWiki and its extensions. The primary person doing this
was Ahmon Dancy, with Brennen Bearnes as backup, both from the Release
Engineering team.

Overall this week was not entirely smooth, but not too bad. The branch
was created automatically early on Tuesday, UTC time, as usual. Ahmon
deployed it to group0, but had to roll back due to T277362. All wikis
are now on 1.36.0-wmf.35. Next week is a new train.

There were several problems found that blocked the train from
progressing to the next group:

- T277056: TypeError: Argument 1 passed to
  Wikimedia\Rdbms\TransactionProfiler::recordConnection() must be of
  the type string, null given
  https://phabricator.wikimedia.org/T277056

- T277061: Many WikibaseLexeme tests suddenly failing in
  quibble-vendor-mysql-php72-{selenium,noselenium}-docker
  https://phabricator.wikimedia.org/T277061

- T277094: Bug in client error logging stops any errors from being
  logged in group 0 wikis
  https://phabricator.wikimedia.org/T277094

- T277161: Minify v2.2.0 incorrectly minifies vue.js (faulty line
  break resulting in early return)
  https://phabricator.wikimedia.org/T277161
  
- T277795: User not found by actor ID: [id]
  https://phabricator.wikimedia.org/T277795
  
  - an attempt at fixing was made, but the fix is stiill being
assessed

- T277302: Hatnote and ambox recognition is poor and essentially only
  works for enwiki
  https://phabricator.wikimedia.org/T277302

- T277362: Deprecation warning client-repo wikitext link
  https://phabricator.wikimedia.org/T277362

- T277593: ErrorException: Invalid cross-wiki page Q[id]. Expected:
  wikidatawiki, got local (duplicate of T277593)
  https://phabricator.wikimedia.org/T277593

Several people helped to find, triage, analyze, or fix of problems
during this week. Release Engineering thanks everyone, without
help we wouldn't be able to deploy MediaWiki.

- Aaron Schulz
- Addshore
- Amir Sarabadani
- Antoine Musso
- Bartosz Dziewoński
- Daimona
- Daniel Kinzler
- Ed Sanders
- FriedhelmW
- Gergő Tisza
- James D. Forrester
- Jon Robson
- Legoktm
- Leszek Manicki
- Lucas Werkmeister
- PerfektesChaos
- RhinosF1
- Roan Kattouw
- Sam Smith
- Tarrow
- Timo Tijhof
- Tobias Andersson
- XanonymusX

There may have been other people, and if so, we apologize for not
including them on the list above.

For more information, please see:

- https://phabricator.wikimedia.org/T274939
- https://wikitech.wikimedia.org/wiki/Heterogeneous_deployment/Train_deploys

-- 
WMF release engineering team | he/him or they/them
"Imagine a world in which every single human being can freely share in
the sum of all knowledge."

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

Re: [Wikitech-l] Train status: on all wikis, looking good

[Lars Wirzenius]

Train has reached all wikis about 1.5 hours ago, and things look OK.
Unless something bad happens soon, this week's train deployment is
over and I shan't say more. If something does come up, there will be
further communications.

If all goes well, have a good Friday and weekend. Be well. Be safe. Be
happy.

-- Your humble train toiler

https://phabricator.wikimedia.org/T274937
https://versions.toolforge.org/

-- 
WMF release engineering team | he/him or they/them
"Imagine a world in which every single human being can freely share in
the sum of all knowledge."

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

[Wikitech-l] Train status: at group1 and OK

[Lars Wirzenius]

The train blocker[0] mentioned earlier was worked around and train was
unblocked. It's now at group1. Watching logs to see if there's any
further breakage.

Thank you to Majavah, Urbanecem, and Vlad.shapik (who came up the
patch) for unblocking train.

Train will move forward tomorrow to group2, if all goes well.

-- Your humble train toiler

[0] https://phabricator.wikimedia.org/T276316
https://phabricator.wikimedia.org/T274937
https://versions.toolforge.org/

-- 
WMF release engineering team | he/him or they/them
"Imagine a world in which every single human being can freely share in
the sum of all knowledge."

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

[Wikitech-l] [Train] 1.36.0-wmf.33 status update: blocked at group0

[Lars Wirzenius]

The 1.36.0-wmf.33 version of MediaWiki is blocked[0].

The new version is deployed to group0[1], but can proceed no further
until these issues are resolved:

* T276316 - Argument 1 passed to getCanonical() must be of the type
  string, null given, called in CentralAuthGroupMembershipProxy.php
  - 

Once these issues are resolved train can resume.

Thank you for your help resolving these issues!

-- Your humble train toiler

[0]. 
[1]. 

-- 
WMF release engineering team | he/him or they/them
"Imagine a world in which every single human being can freely share in
the sum of all knowledge."

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

[Wikitech-l] Train status: at group0 and OK

[Lars Wirzenius]

A short status update about this week's train, 1.36.0-wmf.33.

I've deployed MediaWiki version 1.36.0-wmf.33 to group0. Everything
seems to be OK after an hour. There are currently no blockers. Thank
you everyone involved in MediaWiki development and running Wikimedia
servers for the lack of bumps this week.

Train will move forward tomorrow to group1, if all goes well.

-- Your humble train toiler

https://phabricator.wikimedia.org/T274937
https://versions.toolforge.org/

-- 
WMF release engineering team | he/him or they/them
"Imagine a world in which every single human being can freely share in
the sum of all knowledge."

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

[Wikitech-l] [Train] 1.36.0-wmf.9 status update

[Lars Wirzenius]

The 1.36.0-wmf.9 version of MediaWiki is blocked[0].

The previous train blocker was resolved, and train moved to group1.

The new version is deployed to groups{0,1}[1], but can proceed no
further until these issues are resolved:

* T263034 PHP Notice: Trying to get property 'textContent' of non-object -
  https://phabricator.wikimedia.org/T263034
* T263047 Uncaught TypeError: Cannot read property 'node' of undefined -
  https://phabricator.wikimedia.org/T263047

Once these issues are resolved train can resume. If these issues are
resolved on a Friday the train will resume Monday.

Thank you for your help resolving these issues!

-- Your humble train toiler

[0]. https://phabricator.wikimedia.org/T257977
[1]. 
-- 
WMF release engineering team | he/him or they/them
"Imagine a world in which every single human being can freely share in
the sum of all knowledge."

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

[Wikitech-l] [Train] 1.36.0-wmf.0 status update

[Lars Wirzenius]

The 1.36.0-wmf.9 version of MediaWiki is blocked[0].

The new version is deployed to nowhere at all, including not even on
testwikis, but can proceed no further until these issues are resolved:

* T262900 Rebuilding l10n cache fails for train - 
https://phabricator.wikimedia.org/T262900

Once this issue is resolved train can get moving at all.

Thank you for your help resolving these issues!

-- Your humble train toiler

[0]. https://phabricator.wikimedia.org/T257977
[1]. 

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

[Wikitech-l] scap sync now called scap sync-world

[Lars Wirzenius]

Hi,

I tagged a new release of Scap (3.15.0) yesterday, which will
hopefully be installed on the various servers next week. One of the
changes is that "scap sync" now gives and error, and directs the user
to use "scap sync-world" or "scap sync-file", depending on what they
meant to do.

This change was made because Release Engineering had noticed that
people accidentally run "scap sync" when they meant "scap sync-file".
We decided to have a "flag day"[0] for this because merely warning
about using "scap sync" is very easy to miss, especially if in an
automated system somewhere. We also assume this is not actually used a
lot, and mostly interactively, in which case the error message is
hopefully clear enough to guide the user in the right direction.

 [0] https://en.wikipedia.org/wiki/Flag_day_(computing)

We've fixed and are fixing places in scripts and CI jobs that we know
of where "scap sync" is used. If you find more, please report them to
https://phabricator.wikimedia.org/T250302 .

Sorry about any inconvenience.

-- 
WMF release engineering team | he/him or they/them
"Imagine a world in which every single human being can freely share in
the sum of all knowledge."

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

[Wikitech-l] scap sync --canary-wait-time option (new)

[Lars Wirzenius]

The 3.14.0 release of Scap (current version) added the
--canary-wait-time option to the sync subcommand (which the next
version of Scap will rename to sync-world). This sets the time how
long Scap waits for code to run on canary servers. The default is 20
seconds, but if that is too long or too short, you can now adjust it.

As usual, please use carefully. Don't shorten the time just to make
your deployment go faster.

https://phabricator.wikimedia.org/T217924

-- 
WMF release engineering team | he/him or they/them
"Imagine a world in which every single human being can freely share in
the sum of all knowledge."

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

[Wikitech-l] [Train] 1.35.0-wmf.37 status update

[Lars Wirzenius]

The 1.35.0-wmf.37 version of MediaWiki is blocked[0].

The new version is deployed to group0[1], but can proceed no
further until these issues are resolved:

* Watchlist star gone on Vector -
  https://phabricator.wikimedia.org/T255574

Once these issues are resolved train can resume. Normally, there would
be plenty of time, today being only the first day of the three-day
train week. However, as Friday is a holiday, the week is shorter, and
Release Engineering would like to finish the train on Wednesday
instead of Thursday. If the issue is resolved by tomorrow, we can
resume the train. Otherwise, train can continue on Monday.

Thank you for your help resolving these issues!

-- Your humble train toiler

[0]. 
[1]. 
-- 
WMF release engineering team | he/him or they/them
"Imagine a world in which every single human being can freely share in
the sum of all knowledge."

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

[Wikitech-l] "scap sync" getting renamed to "scap sync-world"

[Lars Wirzenius]

The next release of Scap, our deployment tool, renames the "scap sync"
subcommand to "scap sync-world". The sync command should mostly only
be used by Release Engineering during train deployments, and SWAT
folks updating l10n; everyone else almost certainly wants "scap
sync-file" instead. The change is to make it harder to use the wrong
command by mistake.

If you know of another use of "scap sync", especially in a script
rather then interactive use, please tell me as soon as possible, by
leaving a comment on this Phabricator task:
https://phabricator.wikimedia.org/T250302
  
The next release will probably happen within a few weeks. After that
"scap sync" will be an error, with a message suggesting "scap
sync-world" or "scap sync-file" as a replacement. (Which is OK for
interactive use, but might be problematic for scripts.)

-- 
WMF release engineering team | he/him or they/them
"Imagine a world in which every single human being can freely share in
the sum of all knowledge."

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

Re: [Wikitech-l] [Ops] [Train] 1.35.0-wmf.3 status update

[Lars Wirzenius]

The blockers were quickly dealt with, thanks!

We're at 1.35.0-wmf.3 on all wikis now.

On Thu, Oct 24, 2019 at 03:30:54PM +0300, Lars Wirzenius wrote:
> The 1.35.0-wmf.3 version of MediaWiki is blocked[0].
> 
> The new version is deployed to groups 0 and 1[1], but can proceed no
> further until these issues are resolved:
> 
> * Call to undefined method
>  Wikibase\Lib\EntityTypeDefinitions::getPrefetchingTermLookupCallbacks()
>  <https://phabricator.wikimedia.org/T236370>
> 
> * Too few arguments to function
>  Wikibase\DataAccess\SingleEntitySourceServices::__construct(), 9
>  passed in /extensions/Wikibase/repo/includes/WikibaseRepo.php on line
>  2470 and exactly 10 expected
>  <https://phabricator.wikimedia.org/T236372>
> 
> Once these issues are resolved train can resume. If these issues are
> resolved on a Friday the train will resume Monday.
> 
> Thank you for your help resolving these issues!
> 
> -- Your humble train toiler
> 
> [0]. <https://phabricator.wikimedia.org/T233851>
> [1]. <https://tools.wmflabs.org/versions/>
> 
> On Tue, Oct 22, 2019 at 05:13:14PM -0600, Brennen Bearnes wrote:
> > The 1.35.0-wmf.3 version of MediaWiki is no longer blocked[0], and was
> > deployed to group 0 earlier today at 19:27 UTC, using the American
> > deploy window.
> > 
> > Thanks to andrewbogott for supplying helpful context in unblocking the
> > train, as well as hashar, thcipriani, mutante, and paladox for dealing
> > speedily with Gerrit migration issues.
> > 
> > The train should proceed to group 1 as regularly scheduled, during the
> > European deploy window.[1]
> > 
> > -- Your unassuming train minion
> > 
> > [0]. <https://phabricator.wikimedia.org/T233851>
> > [1].
> > <https://wikitech.wikimedia.org/wiki/Deployments#deploycal-item-20191023T1300>
> > 
> > ___
> > Ops mailing list
> > o...@lists.wikimedia.org
> > https://lists.wikimedia.org/mailman/listinfo/ops
> 
> -- 
> WMF release engineering team | he/him or they/them
> "Imagine a world in which every single human being can freely share in
> the sum of all knowledge."

-- 
WMF release engineering team | he/him or they/them
"Imagine a world in which every single human being can freely share in
the sum of all knowledge."

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

Re: [Wikitech-l] [Ops] [Train] 1.35.0-wmf.3 status update

[Lars Wirzenius]

The 1.35.0-wmf.3 version of MediaWiki is blocked[0].

The new version is deployed to groups 0 and 1[1], but can proceed no
further until these issues are resolved:

* Call to undefined method
 Wikibase\Lib\EntityTypeDefinitions::getPrefetchingTermLookupCallbacks()
 

* Too few arguments to function
 Wikibase\DataAccess\SingleEntitySourceServices::__construct(), 9
 passed in /extensions/Wikibase/repo/includes/WikibaseRepo.php on line
 2470 and exactly 10 expected
 

Once these issues are resolved train can resume. If these issues are
resolved on a Friday the train will resume Monday.

Thank you for your help resolving these issues!

-- Your humble train toiler

[0]. 
[1]. 

On Tue, Oct 22, 2019 at 05:13:14PM -0600, Brennen Bearnes wrote:
> The 1.35.0-wmf.3 version of MediaWiki is no longer blocked[0], and was
> deployed to group 0 earlier today at 19:27 UTC, using the American
> deploy window.
> 
> Thanks to andrewbogott for supplying helpful context in unblocking the
> train, as well as hashar, thcipriani, mutante, and paladox for dealing
> speedily with Gerrit migration issues.
> 
> The train should proceed to group 1 as regularly scheduled, during the
> European deploy window.[1]
> 
> -- Your unassuming train minion
> 
> [0]. 
> [1].
> 
> 
> ___
> Ops mailing list
> o...@lists.wikimedia.org
> https://lists.wikimedia.org/mailman/listinfo/ops

-- 
WMF release engineering team | he/him or they/them
"Imagine a world in which every single human being can freely share in
the sum of all knowledge."

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

[Wikitech-l] [Train] 1.35.0-wmf.3 status update]

[Lars Wirzenius]

(Re-send to wikitech-l, first attampt had wrong address.)

The 1.35.0-wmf.3 version of MediaWiki is blocked[0].

The new version is not deployed anywhere[1], and can proceed no
further until these issues are resolved:

* scap sync failed, database error - https://phabricator.wikimedia.org/T236166

Once these issues are resolved train can resume. If these issues are
resolved on a Friday the train will resume Monday.

Thank you for your help resolving these issues!

-- Your humble train toiler

[0]. 
[1]. 

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

[Wikitech-l] 1.34.0-wmf.14 status update

[Lars Wirzenius]

The 1.34.0-wmf.14 version of MediaWiki is blocked[0].

The new version is deployed to {group0}[1], but can proceed no
further until these issues are resolved:

* TopicListBlock.php sort order params -
  

* RevisionFormatter.php: Unknown content format -
  

Once these issues are resolved train can resume. If these issues are
resolved on a Friday the train will resume Monday.

Thank you for your help resolving these issues!

-- Your humble train toiler

[0]. 
[1]. 

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

[Wikitech-l] Future CI planning, feedback requested

[Lars Wirzenius]

In the Release Engineering team we're preparing for a new CI system.
The current one needs to be replaced. It works well, but parts of it
are getting obsolete. In particular, the Zuul version we use is
obsoleted by upstream. The new version of Zuul is entirely different.
Because of this, we are taking the opportunity to re-think the whole
approach to CI. We would like to introduce the possibility of
continuous delivery and deployment, in addition to continuous
intergration.

Earlier this year, we started a working group to evaluate candidates
for software. In phase 1, we set up some criteria for evaluation, and
considered a large number of possibilities, and winnowed the list down
to three candidates: GitLab CI, Zuul v3, and Argo. For details and a
report, see [0].

We are currently writing up what the new CI system should look like in
more detail. The approach taken is to start with what's needed and
wanted, rather than what the tools provide. The document has had a
first round of internal review, to get rid of the worst issues, and v2
is now open for feedback from the whole movement. You can find it at
[1]. Those with a WMF Google account can comment directly on the doc,
everyone else please use email, either by responding to this email via
wikitech-l or directly to me.

[0] 
https://www.mediawiki.org/wiki/Wikimedia_Release_Engineering_Team/CI_Futures_WG
[1] 
https://docs.google.com/document/d/1EQuInEV-eY_5kxOZ8E1qEdLr8fb6ihwOD9V_tpVFWuU/edit

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

Re: [Wikitech-l] CI jobs run time

[Lars Wirzenius]

On Sun, Jun 16, 2019 at 08:07:02PM +0200, Amir Sarabadani wrote:
> In the past couple of weeks, jenkins jobs became considerably slow. There
> has been some actions to improve it.

Excellent! I'm very glad to hear about these improvements.

Software development is a set of concentric loops, the innermost of
which is the tight edit-build-test loop. The loop that expands this to
include CI is a little less tight, but its speed is also crucial. The
longer the loops, the longer it takes to make any change, so any
shortening of loops is highly likely to result in increased
productivity.

I'm especially glad to hear there's movement to separate actual unit
tests (individual components in the code based tested in isolation
from other parts of the code, and from the external world) from
integration tests that happen to be implemented using unit test
tooling (which is a fine way of implementing them, but they're not
unit tests).

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

Re: [Wikitech-l] Gerrit outage

[Lars Wirzenius]

On Wed, Mar 20, 2019 at 06:56:19PM +, Pine W wrote:
> I'd like to give a quick thanks to folks who have been dealing with
> turbulence.

Me too. It's inspirational and awe-inspiring to see people work on
this.

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

[Wikitech-l] Choosing tools for continuous integration

[Lars Wirzenius]

The Release Engineering team has started a working group to discuss
and consider our future continuous integration tooling. Please help!

The RelEng team is working with SRE to build a continuous delivery and
deployment pipeline, as well as changing production to run things in
containers under Kubernetes. We aim to improve the process of making
changes to software behind our various sites by making it take less
effort, happen faster, be less risky, and as automated as possible.
The developers will have a better development experience, be more
empowered, and more productive.

Wikimedia has had a CI system for many years now, but is based on
versions of tools that are reaching the end of their useful life.
Those tools need to be upgraded, and this will probably require
further changes due to how the new versions function. This is a good
point to consider what tools and functionality we need and want.

The working group is tasked to consider the needs and wants, and
evaluate the available options, and make a recommendation of what to
use in the future. The deadline is March 25. The work is being
documented at 
https://www.mediawiki.org/wiki/Wikimedia_Release_Engineering_Team/CI_Futures_WG
and we're currently collecting requirements and candidates to
evaluate.

We would welcome any feedback on those! Via IRC (#wikimedia-pipeline),
on the talk page of the working group's wiki page above, or as a
comment to the blog post below. Or as a reply to this email.

(Also at
https://phabricator.wikimedia.org/phame/post/view/148/choosing_tools_for_continuous_integration/)

___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

Re: [Wikitech-l] Collecting UI feedback for PolyGerrit - Gerrit

[Lars Wirzenius]

I've implemented some like that for my personal issue tracker, based
on the PGP word list (https://en.wikipedia.org/wiki/PGP_word_list).
Each ticket id is converted into hexadecimal, and the first six hex
digit are used to look up word. For example, I have tickets such as
"REPAY RETRACTION MERIT" and "BEFRIEND CELEBRATE DROPPER". For my
issue tracker, ticket ids are random 64-bit digits, which helps. I'm
not sure how well this would work for sequential integers.

On Tue, Oct 16, 2018 at 6:56 PM C. Scott Ananian  wrote:
>
> Just as a wishlist item, I wonder if we could use something like
> https://tools.ietf.org/html/rfc1760 to convert gerrit #s (and phab #s) to
> short word strings.  I find myself typing sequences of six+ digits over and
> over again during the workday, and the difficulty of getting these exactly
> right (esp transpositions) makes me resort to copy-paste etc.  And of
> course in oral meetings reciting these digit strings is especially fun.  It
> would be much nicer to say `git review -d "top huh"` rather than `git
> review -d 467470`...
>   --scott
>
> On Tue, Oct 16, 2018 at 11:34 AM Paladox via Wikitech-l <
> wikitech-l@lists.wikimedia.org> wrote:
>
> >  I must note that with the branching of 2.16, it means that GWTUI is going
> > away. With the removal of GWTUI in
> > https://gerrit-review.googlesource.com/c/gerrit/+/116790
> >
> > On Tuesday, 16 October 2018, 16:30:53 BST, Paladox via Wikitech-l <
> > wikitech-l@lists.wikimedia.org> wrote:
> >
> >   Upstream have just branched 2.16 with a rc release a few hours away! So
> > a stable release should be done shortly with all the improvements to
> > polygerrit's ui and a inline editor too!
> > On Friday, 5 October 2018, 19:26:30 BST, Paladox via Wikitech-l <
> > wikitech-l@lists.wikimedia.org> wrote:
> >
> >   I have filed it upstream at
> > https://bugs.chromium.org/p/gerrit/issues/detail?id=9815
> > On Friday, 5 October 2018, 06:54:25 BST, Dalba 
> > wrote:
> >
> >  I have issues with copying the text of changed files:
> >
> > `ctrl+a` does not work as it used to anymore: While using the
> > side-by-side diff view, go to a changed file, click on the old/new
> > revision, press `ctrl+a`. In the old UI only the text of the selected
> > file would have been selected, but now the whole page gets selected
> > which is not that useful.
> >
> > Also, copying multiple lines results in extra whitespace being copied
> > in the middle of the lines. For example goto [1], select the text of
> > both lines and copy... It'll be copied as `# -*- coding: utf-8
> > -*-\n\t\n\t\n"""Package to hold all library tests."""` (note the
> > `\n\t\n\t\n` which should have been just a `\n`).
> >
> > [1]:
> > https://gerrit.wikimedia.org/r/c/pywikibot/core/+/463061/5/tests/library_tests/__init__.py#1
> >
> > On Thu, Oct 4, 2018 at 12:39 AM Paladox via Wikitech-l
> >  wrote:
> > >
> > > Hi, i am collecting feedback for Gerrit's New UI called PolyGerrit. It's
> > possible to use PolyGerrit on gerrit.wikimedia.org since 2.14. The new UI
> > has recently been made the default upstream. The Old UI is going away in
> > the next release after 2.16. Upstream have given PolyGerrit another update
> > that looks different to the one on gerrit.wikimedia.org. PolyGerrit now
> > includes a dark ui.
> > >
> > > To switch to PolyGerrit either click the "New UI" button on the footer
> > or put ?polygerrit=1 in the url.
> > >
> > > To switch back to GWTUI either click "Switch back to old ui" on the
> > footer or put ?polygerrit=0 in the url.
> > >
> > > Non dark mode:
> > >
> > > Here's how it looks like:
> > >
> > > Dashboard:
> > >
> > > https://phabricator.wikimedia.org/F26296230
> > >
> > >
> > > Change list:
> > >
> > > https://phabricator.wikimedia.org/F26296240
> > >
> > > Change screen:
> > >
> > > https://phabricator.wikimedia.org/F26296242
> > >
> > >
> > > https://phabricator.wikimedia.org/F26296257
> > >
> > >
> > > Dark mode: https://phabricator.wikimedia.org/F26296282
> > >
> > >
> > > And many other UI improvements across the app.
> > >
> > > You can play around the the new ui from the master branch that will
> > become 2.16 here https://gerrit.git.wmflabs.org/r/
> > >
> > > Please give feedback so upstream can make PolyGerrit even better! You
> > can either file your reports at
> > https://phabricator.wikimedia.org/project/view/330/ or reply to the email
> > with your feedback.
> > >
> > >
> > >
> > >
> > > It has a dedicated team on the UI with a design researcher behind the
> > scenes redesigning polygerrit constantly based on feedback.
> > > ___
> > > Wikitech-l mailing list
> > > Wikitech-l@lists.wikimedia.org
> > > https://lists.wikimedia.org/mailman/listinfo/wikitech-l
> > ___
> > Wikitech-l mailing list
> > Wikitech-l@lists.wikimedia.org
> > https://lists.wikimedia.org/mailman/listinfo/wikitech-l
> > ___
> > Wikitech-l