#58895 closed enhancement (duplicate)
Make it easier to override default loading optimizations
Reported by: | joemcgill | Owned by: | |
---|---|---|---|
Milestone: | Priority: | normal | |
Severity: | normal | Version: | 6.3 |
Component: | Media | Keywords: | |
Focuses: | performance | Cc: |
Description
In [56037] (see #58853) a new set of functions were added to dynamically add loading optimization attributes, like loading="lazy"
and fetchpriority="high"
to images and iframes.
The main function responsible for this behavior is wp_get_loading_optimization_attributes()
, which is called directly in several other places in core, like wp_img_tag_add_loading_optimization_attrs()
and wp_iframe_tag_add_loading_attr()
. However, there doesn't seem to be a straightforward way to disable or override this functionality globally.
Similar functions that are dynamically run on rendered HTML, like wp_img_tag_add_srcset_and_sizes_attr()
and wp_img_tag_add_width_and_height_attr()
have filters in place that allow you to disable that functionality if you want to remove it or implement your own processing via the wp_content_img_tag
hook instead.
wp_img_tag_add_loading_optimization_attrs
has a hook to filter the return value, but not until after wp_get_loading_optimization_attributes
has been called, which
unfortunately produces side effects by calling wp_increase_content_media_count()
or changing the value of wp_high_priority_element_flag()
.
Ideally, you could disable just the optimizations applied to rendered markup via a filter in wp_img_tag_add_loading_optimization_attrs()
in case you needed to override just that functionality. However, wp_get_loading_optimization_attributes()
should also be easier to override or turn off entirely so implementors could enhance core's defaults with their own optimization techniques.
Change History (11)
#2
@
17 months ago
I saw that right after submitting mine 🤣. I think it makes sense to handle filtering wp_get_loading_optimization_attributes()
via #58893, but let's leave this open to discuss the ability to override the optimizations added during wp_filter_content_tags()
specifically.
#3
@
16 months ago
With [56347] committed, this effort is now unblocked. Please refer to the latest codebase for any changes to wp_get_loading_optimization_attributes()
.
This ticket was mentioned in Slack in #core-media by antpb. View the logs.
15 months ago
This ticket was mentioned in Slack in #core-performance by pereirinha. View the logs.
15 months ago
This ticket was mentioned in Slack in #core-performance by mukeshpanchal27. View the logs.
15 months ago
This ticket was mentioned in Slack in #core-performance by mukeshpanchal27. View the logs.
15 months ago
#8
@
15 months ago
- Milestone changed from 6.4 to 6.5
We need a patch and there is no time before Beta 1 to make and test it, so, I am moving this ticket into the 6.5 milestone. If someone produces the patch on the spot, it can be rescheduled back, but we need to be realistic.
This ticket was mentioned in Slack in #core-performance by thekt12. View the logs.
15 months ago
#10
@
15 months ago
- Resolution set to duplicate
- Status changed from new to closed
Newly introduced filter pre_wp_get_loading_optimization_attributes
in #58893 is able to disable optimization logic completely without the side effects mentioned in the description.
Test code to disable optimization logic.
add_filter( 'pre_wp_get_loading_optimization_attributes', 'disable_wp_get_loading_optimization_attributes' ); function disable_wp_get_loading_optimization_attributes( $loading_attr ) { return array(); }
The same filter can be used to set any value for loading
and fetchpriority
enabling overriding of the default value.
@joemcgill I just opened a somewhat related #58893, which is about adding a filter to
wp_get_loading_optimization_attributes()
for this purpose. I think adding a filter to that function makes the most sense as that is the low-level point / source of truth where the actual changes are controlled (rather than e.g. one of thewp_img_*()
functions that call it, and are semantically tied to images).So I think that ticket addresses part of the problems reported here. But I agree there is likely more to bring in order related to that, so I think this ticket is good to have this discussion. I definitely wouldn't consider it a duplicate, but rather overarching.