Problem/Motivation
Split from #3437499: Use placeholdering for more blocks, if we add more pages from Umami, we get a better 'worst case scenario' for the current aggregation / library organisation to improve from.
Steps to reproduce
Proposed resolution
Visit an extra page during these two test methods. Because this loads a different, third, combination of libraries to the first two pages, it results in a different set of asset aggregates and therefore more files and overall bytes being downloaded.
Issues like #3437499: Use placeholdering for more blocks which attempt to reduce duplication between asset aggregates in this kind of scenario will then be able to show an improvement (or not) based on the existing test scenario.
Remaining tasks
User interface changes
API changes
Data model changes
Release notes snippet
Issue fork drupal-3445285
Show commands
Start within a Git clone of the project using the version control instructions.
Or, if you do not have SSH keys set up on git.drupalcode.org:
Comments
Comment #2
catchComment #4
smustgrave commentedAppears to have some failures.
Comment #5
catchComment #6
catchMR is green now.
Comment #7
smustgrave commentedTests are all green so going to lean on that.
Comment #8
catchComment #9
larowlanLooks to need a reroll
Comment #12
immaculatexavier commentedRerolled against 11.x
Comment #13
catch@immaculatexavier the 8144 MR had merge conflicts still, did you actually rebase against origin/11.x? Going to hide the MR, have rebased the original one here.
Comment #15
catchBit more to do re-rolling the original branch - Umami CSS got smaller in the meantime.
Comment #16
catchStraightforward changes, just adjusting the new assertion numbers to match the improved situation after the responsive grid issue landed.
Comment #17
larowlanStill one open thread on the MR regarding the chroot calls
Comment #18
catchArgh I'd missed that comment on the MR. Had temporarily applied #3445847: PHPUnit 10 behaves differently when invoked outside web root as a cherry-pick to get tests passing locally, committed that issue while I was working on this one (which meant it would have disappeared after rebase), but because it then got reverted again, the change was still in here. Removed those.
Comment #19
quietone commentedI read the IS, comments and the MR. Everything looks good to go here. There are no unanswered questions.
Comment #23
alexpottCommitted 506ccd4 and pushed to 11.0.x. Thanks!
Committed 8654d2b and pushed to 11.x. Thanks!
Didn't apply to 10.4.x so didn't work on seeing why as this is test only.