Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Disregard transient cache in perflab_query_plugin_info() when a plugin is absent #1694

Open
wants to merge 9 commits into
base: trunk
Choose a base branch
from

Conversation

b1ink0
Copy link
Contributor

@b1ink0 b1ink0 commented Nov 22, 2024

Summary

Fixes #1617

Relevant technical choices

When a new feature plugin is published, it may not yet exist in the transient. Therefore, whenever one of the expected feature plugin is absent, we should act as if the transient was not set in the first place so that we can obtain the latest plugin info right away.

Logic for this to work:

  • If the plugin slug is not found in the transient then instead of returning a error we proceeded to fetch information from API.
  • If the plugin slug is also not found in the API it is also stored in transient but with key as plugin slug and value as false.
  • If the plugin slug is found in the transient but if its value is equal to false raise the error.

This logic will make sure that the multiple request are not made for the plugin which is absent.

@b1ink0 b1ink0 marked this pull request as ready for review November 22, 2024 14:07
Copy link

github-actions bot commented Nov 22, 2024

The following accounts have interacted with this PR and/or linked issues. I will continue to update these lists as activity occurs. You can also manually ask me to refresh this list by adding the props-bot label.

If you're merging code through a pull request on GitHub, copy and paste the following into the bottom of the merge commit message.

Co-authored-by: b1ink0 <[email protected]>
Co-authored-by: westonruter <[email protected]>
Co-authored-by: felixarntz <[email protected]>
Co-authored-by: swissspidy <[email protected]>

To understand the WordPress project's expectations around crediting contributors, please review the Contributor Attribution page in the Core Handbook.

@westonruter westonruter added this to the performance-lab n.e.x.t milestone Nov 22, 2024
@westonruter westonruter added [Plugin] Performance Lab Issue relates to work in the Performance Lab Plugin only [Type] Bug An existing feature is broken [Type] Enhancement A suggestion for improvement of an existing feature and removed [Type] Bug An existing feature is broken labels Nov 22, 2024
__( 'Plugin not found in WordPress.org API response.', 'performance-lab' )
);
// Cache the fact that the plugin was not found.
$plugins[ $current_plugin_slug ] = false;
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Instead of storing false here, what about storing the WP_Error instance? As it stands right now, there are two conditions resulting in false being logged, which conflates these two error scenarios which we tried to get rid of in #1651 to assist with debugging.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Overall this makes sense to me. We probably shouldn't store the WP_Error instance though, but just its data, since storing PHP class instances is a bit unpredictable with what would happen in different object cache implementations.

@@ -101,9 +102,14 @@ function perflab_query_plugin_info( string $plugin_slug ) {
}
}

if ( ! isset( $plugins[ $plugin_slug ] ) ) {
// Cache the fact that the plugin was not found.
$plugins[ $plugin_slug ] = false;
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

See above about storing a WP_Error instance instead.

Comment on lines 27 to 33
if ( false === $plugins[ $plugin_slug ] ) {
// Plugin was requested before and not found.
return new WP_Error(
'plugin_not_found',
__( 'Plugin not found in cached API response.', 'performance-lab' )
);
}
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If $plugins[ $plugin_slug ] is either an array or a WP_Error, then this can just return that.

Suggested change
if ( false === $plugins[ $plugin_slug ] ) {
// Plugin was requested before and not found.
return new WP_Error(
'plugin_not_found',
__( 'Plugin not found in cached API response.', 'performance-lab' )
);
}

// Cache the fact that the plugin was not found.
$plugins[ $plugin_slug ] = false;
}

set_transient( $transient_key, $plugins, HOUR_IN_SECONDS );
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Here I think it would be good to check if there was an error condition, and if so, set the expiration to 1 minute instead of HOUR_IN_SECONDS.

@westonruter
Copy link
Member

There are still these error scenarios that isn't getting cached:

if ( is_wp_error( $response ) ) {
return new WP_Error(
'api_error',
sprintf(
/* translators: %s: API error message */
__( 'Failed to retrieve plugins data from WordPress.org API: %s', 'performance-lab' ),
$response->get_error_message()
)
);
}
// Check if the response contains plugins.
if ( ! ( is_object( $response ) && property_exists( $response, 'plugins' ) ) ) {
return new WP_Error( 'no_plugins', __( 'No plugins found in the API response.', 'performance-lab' ) );
}

Copy link
Member

@felixarntz felixarntz left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks @b1ink0! This looks good, just one thing that needs to be fixed I think.

Comment on lines 116 to 120

set_transient( $transient_key, $plugins, MINUTE_IN_SECONDS );
} else {
set_transient( $transient_key, $plugins, HOUR_IN_SECONDS );
}
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This differentiation does not happen in the correct clause. We should use a shorter cache duration in case of any error, not just in the case of this specific error.

I think it would be most straightforward to implement this via a boolean flag like $has_errors that is set if anywhere in the (uncached) logic an error is set. Then the code here could be changed like:

Suggested change
set_transient( $transient_key, $plugins, MINUTE_IN_SECONDS );
} else {
set_transient( $transient_key, $plugins, HOUR_IN_SECONDS );
}
}
if ( $has_errors ) {
set_transient( $transient_key, $plugins, MINUTE_IN_SECONDS );
} else {
set_transient( $transient_key, $plugins, HOUR_IN_SECONDS );
}

Copy link
Member

@felixarntz felixarntz left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@b1ink0 There are two points of follow-up feedback, though some of them have overlap with #1691. But since this is broken too and affected by the same logic, maybe we should work towards a solution that addresses both?

cc @westonruter

Comment on lines 60 to +83
if ( is_wp_error( $response ) ) {
return new WP_Error(
'api_error',
sprintf(
/* translators: %s: API error message */
__( 'Failed to retrieve plugins data from WordPress.org API: %s', 'performance-lab' ),
$response->get_error_message()
)
$plugins[ $plugin_slug ] = array(
'error' => array(
'code' => 'api_error',
'message' => sprintf(
/* translators: %s: API error message */
__( 'Failed to retrieve plugins data from WordPress.org API: %s', 'performance-lab' ),
$response->get_error_message()
),
),
);

$has_errors = true;
}

// Check if the response contains plugins.
if ( ! ( is_object( $response ) && property_exists( $response, 'plugins' ) ) ) {
return new WP_Error( 'no_plugins', __( 'No plugins found in the API response.', 'performance-lab' ) );
if ( ! $has_errors && ! ( is_object( $response ) && property_exists( $response, 'plugins' ) ) ) {
$plugins[ $plugin_slug ] = array(
'error' => array(
'code' => 'no_plugins',
'message' => __( 'No plugins found in the API response.', 'performance-lab' ),
),
);

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think for both of these scenarios we need to set the error for every single plugin of ours. In other words, iterate through the list from perflab_get_standalone_plugins() and set the same error for every plugin's slug.

Otherwise, this can cause a problem where the API failing would cause the request to be issued for every single plugin, which would defeat the purpose of the transient cache and the combination of those plugin requests into a single API request (see #1542). This would effectively fix #1691, and since this PR is touching the exact same logic, I wonder whether we should go straight to a solution that addresses both issues.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

To make things clear as its getting confusing please answer these below question or my understanding of the solution needed.

  • If error is encounterd at line 57 and 76 we should add the error to the all perflab_get_standalone_plugins plugins so that if any one of the plugin is requested it should return error?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Correct. That's because a failing API response is by definition an error that affects all plugins.

Comment on lines 121 to 131
if ( ! isset( $plugins[ $plugin_slug ] ) ) {
// Cache the fact that the plugin was not found.
$plugins[ $plugin_slug ] = array(
'error' => array(
'code' => 'plugin_not_found',
'message' => __( 'Plugin not found in API response.', 'performance-lab' ),
),
);

// Enqueue the required plugins slug by adding it to the queue.
if ( isset( $plugin_data['requires_plugins'] ) && is_array( $plugin_data['requires_plugins'] ) ) {
$plugin_queue = array_merge( $plugin_queue, $plugin_data['requires_plugins'] );
$has_errors = true;
}
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Now that I look at this again, what is this particular check needed for? Wouldn't this already be catered for by the check above in line 101?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This check is needed because if the $plugin_slug is not in the perflab_get_standalone_plugins() then this $plugin_slug will not be processed inside the loop.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

In that case, I think we should make this check at the very beginning of the function, even before looking at the transient cache. This function should never be used for a plugin that is not among our plugins I think, so we could already check that before all the main logic in this function.

We wouldn't want that someone calling the function with e.g. jetpack would trigger this API request and logic to run for every single page load just because jetpack would (rightfully) never be included in the response.

Copy link
Contributor Author

@b1ink0 b1ink0 Nov 27, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If we only check the $plugin_slug inside perflab_get_standalone_plugins() it will also cause error for the dependency plugins. And if we want to dynamically check the dependency then we need to request the API.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why does the function need to be called for dependency plugins?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The perflab_render_plugin_card() function is calling the perflab_get_plugin_availability() function
and inside the perflab_get_plugin_availability() on here it is calling the perflab_query_plugin_info() for the dependency plugin.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ah, that's a good point. But in that case, I think at least this error should use a different message? I would try to clarify it as: If the plugin is not found at this point, we know it's not one of our plugins, since it's not in the API response, and the API response should always included all our plugins plus its dependencies.

'message' => __( 'Plugin not found in API response.', 'performance-lab' ),
),
);
continue;
Copy link
Contributor Author

@b1ink0 b1ink0 Nov 27, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Due to the recent commit ff1b440 this could cause incorrect behavour if $plugin_slug is not found in the all_performance_plugins it will add error for that $plugin_slug as we are now using the $has_errors flag which will not be set to true in this case. The below logic will fail.
Solution could be to check it using the isset( $plugins[ $plugin_slug ]['error'] ) condition but it will defect the usage of the $has_errors flag.

	if ( $has_errors ) {
		set_transient( $transient_key, $plugins, MINUTE_IN_SECONDS );

		return new WP_Error(
			$plugins[ $plugin_slug ]['error']['code'],
			$plugins[ $plugin_slug ]['error']['message']
		);
	}

Reason for the $has_errors flag not setting: As the errors could occur for other plugin ( not $plugin_slug ) currently code will cache the error but will return the requested $plugin_slug information.

Question: Should we return error even if the error is encountered for the other plugin ( not $plugin_slug ) ?
cc: @felixarntz

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

See my reply in #1694 (comment) - I think this problem would become irrelevant if we checked that the plugin slug is among our performance plugins in the beginning of the function?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Revisiting this: Why should we not set the $has_errors flag to true here? I think we should, given this is an error too.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Reason for the $has_errors flag not setting: As the errors could occur for other plugin ( not $plugin_slug ) currently code will cache the error but will return the requested $plugin_slug information.

Question: Should we return error even if the error is encountered for the other plugin ( not $plugin_slug ) ?

This above is reason. And question related to your question.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
[Plugin] Performance Lab Issue relates to work in the Performance Lab Plugin only [Type] Enhancement A suggestion for improvement of an existing feature
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Disregard transient cache in perflab_query_plugin_info() when a plugin is absent
3 participants