Do you have a copy of the JSON?
This RfC is now ready for discussion, I've taken Ori's initial details and expanded it into a proposed implementation: https://www.mediawiki.org/wiki/Requests_for_comment/Proper_command-line_runner_for_maintenance_tasks
This is stalled on reprepro not working, so the existing packages can't be updated, let alone adding new ones.
I think most of what people want is documented on https://www.mediawiki.org/wiki/Manual:Extension.json/Schema now...is there still stuff that's missing?
Thu, Aug 16
I implemented an OAuth2 based PrimaryAuthenticationProvider, and ran into the same thing.
I had an old ~/.my.cnf that was apparently getting in the way. My bad :(
Should we deprecate the BeforeParserFetchTemplateAndtitle hook if it's usage is incompatible with extensions like Scribunto/TemplateStyles?
Rename finished, I think it was just a database timeout. Sorry about not seeing this ticket earlier :(
Sorry, looking into this now.
I'm not sure what kind of a useful answer you're going to get...I suspect each case has a different answer/reason. For ~/vagrant, it's used as a development tool on individual developer's laptops, so connecting to local services over HTTP should be fine.
Wed, Aug 15
What does "fail" mean? Is it outputting error messages? Or just failing to log you in?
I'm re-opening this given that the dependent pipeline is still a major source of problems. I regularly observe jobs failing in gate-and-submit (sometimes flakiness, often out of our control like composer failures), causing all test results to be thrown away and run again. mediawiki/core jobs take about ~20 minutes to run, so you can easily be looking at an hour or two to merge a core patch (which in itself is not an issue, the problem is that there's no good reason it takes so long).
Yep, looks like the zip extension is missing. For completeness:
--- 56.txt 2018-08-15 12:29:22.815823579 -0700 +++ 72.txt 2018-08-15 12:29:48.039927871 -0700 @@ -1,16 +1,11 @@ [PHP Modules] -apc apcu -bcmath -bz2 calendar Core ctype curl date -dba dom -ereg exif fileinfo filter @@ -19,14 +14,10 @@ gettext hash iconv -imagick intl json libxml mbstring -mcrypt -mhash -mysql mysqli mysqlnd openssl @@ -40,13 +31,12 @@ Phar posix readline -redis Reflection session shmop SimpleXML -soap sockets +sodium SPL sqlite3 standard @@ -60,7 +50,6 @@ xmlwriter xsl Zend OPcache -zip zlib
The above patch makes the TitleValue constructor twice as fast as Title::makeTitle in my benchmarking. I'm OK with that for now (though it seems clear there's more room for optimization).
Confusingly, Wikidata.org and WikimediaBadges have very similarly named images (resources/images/wb-badges-notproofread.png vs resources/images/badge-notproofread.png) except they're totally different.
Based on https://wikidata.beta.wmflabs.org/wiki/Special:DispatchStats it looks like dispatch hasn't run for 12 days?
I don't see any reason to archive it. It's an example extension. It could use a bit of cleanup though to meet modern standards.
Tue, Aug 14
api-feature-usage is exposed via Special:ApiFeatureUsage, which queries the log entries from elasticsearch, I'm not sure if that's dependent upon them being in logstash though.
Did she log into wikitech and set a real password instead of the temporary one? That would populate user_password.
And thank you for testing :-)
I think we need to backport a6f172ef3cced06f808e22c425bb272e4f5178c7 to 1.27.
It's still very much alpha quality for now, there's still some stuff missing. I hope to send out an announcement calling for beta testers later this week.
I only have access to the mediawiki account too.
This seems like a Continuous-Integration-Infrastructure issue that the nodes are running out of memory...
This was done in 1.27 with AuthManager.
Just to note, that editing the Gruntfile.js should not be necessary at all. Extensions should have their Gruntfile.js set to read the canonical list straight from extension.json (e.g. https://gerrit.wikimedia.org/g/mediawiki/extensions/AddPersonalUrls/+/master/Gruntfile.js#5), so it only needs to be updated in one place. If they aren't already doing this, this would be a good opportunity to migrate them.
Should we have a script to do this?
[22:50:35] <wikibugs_> (PS1) Legoktm: Parser: Add accessors needed by CodeMirror [core] - https://gerrit.wikimedia.org/r/452600
[22:50:37] <wikibugs_> (PS1) Legoktm: Parser: Call firstCallInit() in getTags/getFunctionHooks [core] - https://gerrit.wikimedia.org/r/452601
I think we should use the jessie packaged npm as the bootstrap to install a newer version of npm - this is what we do for CI.
Are these just php5.6 and nodejs tools only? Can you provide a list of tools where you've seen this issue?