Closed (duplicate)
Project:
Drupal core
Version:
9.3.x-dev
Component:
node system
Priority:
Normal
Category:
Bug report
Assigned:
Unassigned
Issue tags:
Reporter:
Created:
6 Nov 2014 at 10:36 UTC
Updated:
11 Apr 2022 at 20:30 UTC
Jump to comment: Most recent, Most recent file
Comments
Comment #2
lauriiiWhy are we removing documentation for variables that still exists?
Comment #3
berdirMade a few improvements, fixed bartik_preprocess_node(), and made sure that uid/created are always hidden. And updated node.html.twig in bartik.
@laurii: What are we removing that still exists? teaser no longer exists, the others are just moved.
Comment #5
berdirFixed that test. Switched to test both page and article to test different settings, only those that show the author now add the relevant user tags.
Comment #6
lauriiiI think I misread the patch because there's too many spaces before date and authorname on the documentation
Comment #9
dotsi commentedRerolled the patch from #5. There were some conflicts.
Comment #11
star-szrNeeds another reroll I suppose :)
Comment #12
lauriiiReroll
Comment #14
star-szrThat time again apparently. Needs reroll.
Comment #15
joelpittetHere's a re-roll it fell back to 3-way auto merge so nothing to resolve.
Comment #16
joelpittetIs moving this hunk needed?
We don't need to drupal_render these variables, let the template do that.
Because of the duplication we have for templates in core from classy this also needs to happen in classy's node template.
There is some coding standards for the indent and 80 chars from what dreditor is saying.
This looks like it snuck in and may not exist? Or is this a clean-up of something that is missing?
Does the user: need to be removed?
That change is not needed as it is Bartik's.
Comment #18
joelpittetOk I think the drupal_render() is needed... but it really shouldn't be, RDF is concatinating it and using SafeMarkup... would be nicer if they just did prefix/suffix on the render element or something. That should be a follow-up.
I've put the other bits in place and fixed the failing test but turning on display_submitted (NodeTestBase has it off be default).
Let's see if testbot agrees.
Comment #20
joelpittetHmm 2 different errors, text showing that shouldn't be...
Comment #21
kgoel commentedLooking into fails and working on it.
Comment #22
kgoel commentedComment #23
kgoel commentedMoved this hunk down as the test was failing because author_name and date variable should not be included in $variables['content'] . we render them before building the 'content' array so they don't show up there.
In #18 patch $variable['date'] and $variable['author_name'] was set after building the content variable.
Replaced deprecated drupal_render with \Drupal::service('renderer')->render.
Comment #24
kgoel commentedI am going to try do some profiling tomorrow.
Comment #26
kgoel commentedHere is profiling result -
=== 8.0.x..8.0.x compared (556d0efcc18be..556d13edd0b22):
ct : 610,204|610,204|0|0.0%
wt : 2,045,741|2,013,482|-32,259|-1.6%
mu : 58,569,312|58,569,344|32|0.0%
pmu : 59,801,216|59,801,624|408|0.0%
=== 8.0.x..2370145-a compared (556d0efcc18be..556d153c4efc5):
ct : 610,204|586,877|-23,327|-3.8%
wt : 2,045,741|1,773,106|-272,635|-13.3%
mu : 58,569,312|58,496,288|-73,024|-0.1%
pmu : 59,801,216|59,727,592|-73,624|-0.1%
http://www.lionsad.de/xhprof-kit/xhprof/xhprof_html/?run1=556d0efcc18be&...
Comment #27
star-szr@kgoel you should include the lionsad.de links those are very important for being able to verify your results :)
Comment #28
joelpittetNice you fixed my other bug.
The cache tags are no longer needed for user because by default the tests have display_submitted false and the now won't get rendered and there for don't get tags for user context bubbling up.
Just need to remove the unneeded user cache tags from the test now.
Comment #29
kgoel commentedComment #31
joelpittetLooks you are on the right track @kgoel. The next one looks like it needs timezone cache key removed.
Comment #32
joelpittetMaybe this will finish them off? Didn't work locally (10 fails) running
Comment #33
joelpittetengage!
Comment #39
joelpittetWell those 10 fails are there! Well I'll be. That's an unexpected but nice to see that local is showing the same.
Comment #40
davidhernandezAssuming it needs a reroll at this point, but lets check.
Comment #43
joelpittetAll cache failures it looks like.
Comment #44
davidhernandezNeeded rerolling. Lets see if that worked.
Comment #56
larowlanIs this still relevant now that #2923701: Mechanism to disable preprocessing of node base fields so they can be configured via the field UI is in an we don't always process those fields if they're configured via manage-display?
Comment #57
berdirThere are still some micro-optimizations like removing the teaser boolean flag. Which is a BC break and I'm not sure that's worth removing, would need a BC layer and stuff. I'd say we can close it..
Comment #58
larowlanThanks, closing as duplicate of #2923701: Mechanism to disable preprocessing of node base fields so they can be configured via the field UI