Recap: Modernizing Testbot Initiative meeting August 18th

We encourage users to post events happening in the community to the community events group on https://www.drupal.org.
hestenet's picture
Start: 
2015-08-18 18:30 - 19:30 America/Los_Angeles
Organizers: 
Event type: 
Online meeting (eg. IRC meeting)

The Testbot team is conducting public meetings. Want to join & help them, need support or just talk to the team, join the meetings! Come and share what you did last week & the plan for the weeks ahead.

To Know more about the testbot initiative check out the DrupalCI documentation.

Comments

Meeting Notes

hestenet's picture

Sorry this recap is about a week late! Here are the details from our last meeting:

Attendance: hestenet, isntall, drumm, basic, jthorson

Issue Review

Drupal 8 blockers:

#2467925

  • PGSQL
    • isntall - needed to contact the core devs about the upsert commit which further broke pgsql bots
  • SQLite
    • Looks like it was fixed and 100% passing as of Monday night
  • PHP7 
    • Current solution for build info is good enough, though it’s a bit hard to scroll through and find the version info.
  • Test Discovery
    • jthorson will try to get his core patches into d6, d7, d8
    • However, if necessary isntall/DA staff have work arounds
  • Email Notifications
    • Drumm can prioritize the email notifications based on the data we now collect in the database a little bit ahead of fixing the display direct on Drupal.org - gets this story done one item sooner.
  • Make simpletest fail with more than 64 MB memory usage
    • We agreed to just push back for any requests about this issue on the old bots - since the old bots use of opcache just breaks this kind of test. 
  • MySQL MariaDB
    • Agreed it’s out of scope as officially maria is not the supported environment - maybe something to look at after blockers are done

Disable PIFT/PIFR blockers

 
#2534132

  • D7 Core/Contrib
    • Separate job version will be available to commit if we want to
    • Jeremy will separately pursue the ‘different template’ version (see below)
  • PHP Fatals should be failures
  • Patches that fail to apply should result in failure
    • This is a more general issue: any build step failure should report a failure in a consumable junit xml format
  • Parsing results from D7 and D6 testing
    • Consume the simpletest mysql tables and format int he same way that we do with the sqlite db for d8
  • DrupalCI clean 
    • isntall will review/tweak the patch
  • Bus errors
    • Possibly fixed by the clean command - so do that first
  • Syntax linting - jthorson
    • Agreed it’s a regression/provides minimal testing for modules without tests - needs to be built.
  • Cancel a job
    • Minor in grand scheme, and depends on DruaplCI clean anyway
  • Email notifications betterer - see above re: drumm's idea about generating emails from the result data we now have directly on D.O.

Additional issues:

  • Ensuring that if any build-step fails the testis aborted (a more general case of - if a patch fails to apply that needs to report as a fail). 

Drupal.org Testing Infrastructure

Group notifications

This group offers an RSS feed. Or subscribe to these personalized, sitewide feeds:

Hot content this week