browsertrix-behaviors
browsertrix-behaviors copied to clipboard
fix: facebook photos page scroll
Fixes #59
Facebook changed the XPath for selecting the first image in the photos page, which with this PR gets autopilot going again. However, because the photos list wraps around (clicking next enough times will get you back to the start) there's no real end condition for this behaviour...
Also in my tests I came across various different URL formats for photos pages depending on the source. They often don't match the regex we're looking for, which means the behaviour gets skipped. This PR doesn't deal with that case
The test fails because the profile saved in the DO Space has been logged out of facebook. I've uploaded a new profile to s3://dh-preserve/test-profiles/facebook/profile.tar.gz which only includes the sample logged in facebook account, but I'll need someone with permissions to the settings page in this repo to add it as a secret.