Closed (fixed)
Project:
Search API
Version:
8.x-1.x-dev
Component:
General code
Priority:
Major
Category:
Bug report
Reporter:
Anonymous (not verified)
Created:
21 Nov 2016 at 15:52 UTC
Updated:
13 Nov 2025 at 13:05 UTC
Jump to comment: Most recent, Most recent file
Comments
Comment #1
Anonymous (not verified) commentedSchnWalter created an issue. See original summary.
Comment #2
Anonymous (not verified) commentedComment #3
Anonymous (not verified) commentedComment #4
borisson_Yep, that makes sense.
Comment #5
drunken monkeyThanks a lot for reporting this problem! You're right, once again we didn't think of this.
However, your patch would also lead to the node's comments not being updated in indexes, which could be a security problem. It's rather improbable, but in theory there's nothing keeping you from excluding a node from the index, but indexing its comments. And in this case, the comments would still need to know about the node's access records changes.
The attached patch seems better suited to me. Please test/review!
Comment #6
borisson_yep
Comment #8
drunken monkeyGood we agree.
Committed.
Thanks again, everyone!
Comment #9
drunken monkeyComment #11
simon georges commentedI apologize for coming back to this relatively old issue.
My problem with moving the "skip" in the final patch is that the line
!$index->isValidProcessor('content_access')is executed. In our case, it means instantiating a lot of plugins, which account for 10% of the total time of the migration (see attached blackfire screenshot), even when we didn't want to index the content. That don't happen when using the first patch of the thread.I'm wondering if there is a way to avoid that, or simply if there is a bug somewhere else in my code resulting in instantiating as many plugins.