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

Add an admin pointer for version 1.0 #1271

Merged
merged 5 commits into from
Jul 26, 2018
Merged

Add an admin pointer for version 1.0 #1271

merged 5 commits into from
Jul 26, 2018

Conversation

kienstra
Copy link
Contributor

@kienstra kienstra commented Jul 18, 2018

Request For Review

Hi @hellofromtonya,
Thanks for your PR #1274. I'm going to review it next.

It would be great if you could review this. Then, maybe @westonruter could give it another look and merge when it's ready.

  • The admin pointer should appear on checking out this branch and loading /wp-admin.
  • Mainly follows Weston's mockup in Issue 1254.
  • Enqueues the script only if this pointer hasn't been dismissed.
  • Based on this tutorial
  • Copies some of the structure of amp-block-validation.js

Closes #1254

amp-admin-pointer-image

Mainly follow Weston's mockup in Issue 1254.
Enqueue the script only if this hasn't been dismissed.
Based on:
https://code.tutsplus.com/articles/integrating-with-wordpress-ui-admin-pointers--wp-26853
Also uses the scaffold of amp-block-validation.js.
*
* @var string
*/
const TEMPLATE_POINTER_ID = 'amp_template_mode_pointer_10';
Copy link
Contributor Author

@kienstra kienstra Jul 19, 2018

Choose a reason for hiding this comment

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

When you're testing this, you might have to change this value locally. Once you dismiss the pointer, it won't display again. Unless you manually update the user meta value.

function amp_admin_pointer() {
$admin_pointer = new AMP_Admin_Pointer();
$admin_pointer->init();
}
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 bootstrapping is repetitive. It'd be nice if amp_post_meta_box(), amp_editor_core_blocks(), and amp_admin_pointer() could be combined into one function, like amp_admin_bootstrap().

But in the unlikely case that a user unhooked one of these from wp_loaded, this would break that.

Copy link
Contributor

Choose a reason for hiding this comment

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

Then the questions are:

  • Do we need to provide the ability to unhook any of those components?
  • Do we need to provide the ability to unhook any of these components?
	add_action( 'wp_loaded', 'amp_editor_core_blocks' );
	add_action( 'wp_loaded', 'amp_post_meta_box' );
	add_action( 'wp_loaded', 'amp_editor_core_blocks' );
	add_action( 'wp_loaded', 'amp_add_options_menu' );
	add_action( 'wp_loaded', 'amp_admin_pointer' );

Changing that to add_action( 'wp_loaded', 'amp_admin_bootstrap' ); and then letting amp_admin_bootstrap() manage the order and initializations results in less callbacks having to be processed, i.e. only 1 instead of 5. It'll be slightly faster.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Good questions. I think it'd be best to avoid renaming the functions above, as a user could have unhooked them.

But it's not a strongly-held belief, and there probably aren't many users who would have unhooked them.

Another option is renaming amp_admin_pointer() to amp_admin_bootstrap(), without making any other change.

Then, any future admin classes could add their bootstrapping there.

Copy link
Contributor

Choose a reason for hiding this comment

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

Another option is renaming amp_admin_pointer() to amp_admin_bootstrap(), without making any other change.

This option is future-thinking as you are planning for future bootstrapping functionality. That's good.

Now, one could argue that confusion may be introduced when a dev is deciding where to bootstrap. Adding an inline comment above the action hook may elevate that confusion.

Copy link
Contributor Author

@kienstra kienstra Jul 20, 2018

Choose a reason for hiding this comment

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

Though amp_editor_core_blocks() has never been in a released version. It was added in version 1.0.

amp_post_meta_box() was added in 0.6. I'm not sure why someone would want to hide the meta box section entirely, and remove the ability to change the AMP enabled status:

amp-meta-box-section

Maybe I could only combine amp_editor_core_blocks() and amp_admin_pointer() into amp_admin_bootstrap().

Copy link
Contributor

Choose a reason for hiding this comment

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

I'd vote to review how we want to bootstrap and consider what callbacks we can consolidate. That said, I think it's outside the scope of this particular PR. We may want to open a ticket for it and begin the discussion process beyond ours here.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Sounds good, Tonya.

Copy link
Member

Choose a reason for hiding this comment

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

Yes, I think we should try to get away from bootstrapping like this because the class instantiation variable that gets created is not accessible anywhere. We should instead of an AMP plugin manager that is responsible for instantiating classes and then keeping around the references to the class instances for the plugin (and others) to access later.

),
'position' => array(
'edge' => 'left',
'align' => 'middle',
Copy link
Contributor Author

@kienstra kienstra Jul 19, 2018

Choose a reason for hiding this comment

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

I think 'edge' => 'left' is probably correct, but the 'align' value might need improvement.

Copy link
Contributor

Choose a reason for hiding this comment

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

@kienstra I think both left and middle works well.

pointer: data.pointer.pointer_id,
action: 'dismiss-wp-pointer'
} );
}
Copy link
Contributor

Choose a reason for hiding this comment

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

@kienstra Right now, the pointer scrolls instead of sticking in a fixed position with the AMP menu.

amp-pr1271-pointer-scrolling

One way to fix this is to apply CSS position: fixed;. A possible solution is to leverage the show function and then set the CSS:

                    show: function( event, t ) {
                        t.pointer.css( 'position', 'fixed' );
                    }

amp-pr1271-fixed

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Thanks, Tonya! That's a great suggestion to use show. Commit e84e22 does that.

* @since 1.0
*/
public function enqueue_pointer() {
$dismissed = explode( ',', strval( get_user_meta( get_current_user_id(), 'dismissed_wp_pointers', true ) ) );
Copy link
Contributor

@hellofromtonya hellofromtonya Jul 19, 2018

Choose a reason for hiding this comment

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

When no pointers have been dismissed, an empty is returned from get_user_meta. In that case, there's no need to run the rest of the code , i.e. strval(), explode(), and in_array(). We can then optimize this code by separating it out.

To keep the method focused on the enqueueing tasks, consider abstracting it into a new method:

	protected function is_pointer_dismissed() {
		$dismissed = get_user_meta( get_current_user_id(), 'dismissed_wp_pointers', true );
		if ( empty( $dismissed ) ) {
			return false;
		}
		$dismissed = explode( ',', strval( $dismissed ) );

		return in_array( self::TEMPLATE_POINTER_ID, $dismissed, true );
	}

Copy link
Contributor Author

@kienstra kienstra Jul 20, 2018

Choose a reason for hiding this comment

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

Hi @hellofromtonya,
Thanks, that's a good suggestion. Self-documenting functions like has_pointer_been_dismissed() help 😄

But I think it only optimizes this if $dismissed is empty...if the user has never dismissed a pointer before.

Once the user dismisses this pointer, $dismissed should at least have the string of this pointer. Unless there was an issue in the POST request in the JS file.

What do you think?

Copy link
Contributor

@hellofromtonya hellofromtonya Jul 20, 2018

Choose a reason for hiding this comment

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

Correct. If the user has never dismissed a pointer, then the user meta does not exist or is empty. In that case, it optimizes the code by bailing out and not running strval(), explode(), or in_array().

Copy link
Contributor

Choose a reason for hiding this comment

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

I also find it expresses the intent more clearly to help us more quickly read and understand what's going on.

Copy link
Contributor

Choose a reason for hiding this comment

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

@kienstra You could name that guard clause method something less passive such as

is_pointer_dismissed()

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Thanks, commit e60f0 adds is_pointer_dismissed().

$dismissed = explode( ',', strval( get_user_meta( get_current_user_id(), 'dismissed_wp_pointers', true ) ) );

// Exit if the pointer has been dismissed.
if ( in_array( self::TEMPLATE_POINTER_ID, $dismissed, true ) ) {
Copy link
Contributor

@hellofromtonya hellofromtonya Jul 19, 2018

Choose a reason for hiding this comment

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

Here, this can now be simplified to this:

		if ( $this->is_pointer_dismissed() ) {
			return;
		}

These changes make it a little bit easier to read and quickly know what's going on in the guard clause.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Thanks, this is applied in commit e60f0.

Fix issue where pointer didn't appear alongside the AMP menu item.
Uses the show function, as Tonya suggested.
@kienstra
Copy link
Contributor Author

kienstra commented Jul 20, 2018

Hi @hellofromtonya,
Thanks for your great suggestions here. Using show was a great idea.

There are still a few open discussions here. I'm happy to add has_pointer_been_dismissed() if you'd like.

@hellofromtonya
Copy link
Contributor

hellofromtonya commented Jul 20, 2018

I'm happy to add has_pointer_been_dismissed() if you'd like.

@kienstra your code works well. Great job!

I'd vote for abstracting the dismissed guard clause for the following reasons:

  1. It clear communicates intent of what that code is doing.
  2. It's more performant when the user has not dismissed pointers.
  3. It's more readable and quickly understandable.

I suggest naming the method name is_pointer_dismissed() instead of the passive name I originally suggested.

I'll leave it up to you whether you wish to change it or not. I'm approving the PR as your code does work.

Abstract logic from enqueue_pointer() into this.
First check that the user meta is not empty.
Then, return whether it is in the array.
Props @hellofromtonya.
@kienstra
Copy link
Contributor Author

Thanks, Applied Suggestion

Hi @hellofromtonya,
Thanks for reviewing this, and adding so much thought into your comments.

Commit e60f0 adds is_pointer_dismissed().

@westonruter westonruter added this to the v1.0 milestone Jul 26, 2018
@westonruter westonruter merged commit 3f09a43 into develop Jul 26, 2018
@westonruter westonruter deleted the add/admin-pointer branch July 26, 2018 03:21
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants