| 1 | | That's just the data above. Here's an interpretation. |
| 2 | | |
| 3 | | Wth the mini plugin the situation did not improve as it took 7.6 to 8 seconds: |
| 4 | | |
| 5 | | WP_Comments_List_Table::comment_type_dropdown() does not count the comments, it only checks to see if there is any comment of the desired type. That kind of a check (if the comments are being counted, then do not use order by) is does not help improve performance. |
| 6 | | |
| 7 | | With the proof of concept pull request the situation did not improve, the query would take 7.1 to 7.8 seconds: |
| 8 | | |
| 9 | | In your pull request @peterwilsoncc you tried to adjust the get_comments() calls in different places to remove the ordering where it's not necessary. But you omitted the exact single place where it matters for our issue: https://github.com/WordPress/wordpress-develop/pull/4571/files#diff-1018e4974931b2cea88b413dadc533d3cf535f977c7f3a5f310e0bc7376233a8R522 |
| 10 | | |
| 11 | | So our fix it still needed. This is how much faster it gets - 0.7 to seconds or it does not even appear in the list of slow queries. |
| 12 | | |
| 13 | | Let's get ''this'' performance bottleneck solved gentlemen. Arguments about whether WordPress performance bottlenecks matter (of course they do, except for toy sites and small personal sites) are anti-productive. |
| | 1 | duplicate |