Trick: git merge --no-ff origin/dev

@Jdlrobson, @phuedx: Seems like an experiment hasn't been defined for Gather and it's initializing by default in beta. Should I simply remove the bucketing work and it can be reintroduced if such an experiment is defined?

@Jdlrobson: I'm not sold on having EventLogging as an optional dependency for QuickSurveys. Why are we worried at this point about it not being installed?

Scheduled for evening swat. https://wikitech.wikimedia.org/wiki/Deployments#Week_of_September_7th

@Jdlrobson I can push to a dev branch

First three responses +

response from tcpflow

@dduvall in efforts to provide more data i've ran client.log_in( un, pw ) and quickly aborted to retrieve the first few results.

@dduvall, my findings exactly. Even though I'm currently blocked on T111133. I was able to determine that I am seeing a JS version of the mobile site when running the nojs tests.

@Jdlrobson @bmansurov I can't even reach the failing step locally. Will try some more after lunch.

The above change is only partially complete. This patch does not resolve the issues with the text-indent hack. Need to research this more.

@Jdlrobson for some reason the first step in the tests is failing for me.

Given the quick survey test pages are installed # features/support/step_definitions/common_steps.rb:1
  unexpected HTTP response (404) (MediawikiApi::HttpError)

This just popped in the todo column. We didn't estimate this. Ok to work on ? @Jhernandez @Jdlrobson

@phuedx +1 Parsed on the server, payload would be reduced by only sending the needed survey modules to the client. Seems we could register modules and their i18n messages for enabled surveys via hook.

Blocked as it seems creating an instance is off the table.

sorry @bmansurov i'll ping you