WordPress › Support » [Plugin: SEO Slugs] Setting post slug when posting via XML-RPC

Waddle
Member
Posted 2 years ago #
I'm new to both plugin development and XML-RPC.

I'm currently trying to re-work the seo-slugs plugin so that it removes the stopwords from the post slug when I'm posting via XML-RPC. Currently seo-slugs works properly when posting using the dashboard, but when I try to post using XML-RPC the stopwords aren't trimmed from the slug. I'm using the metaWeblog.newPost call to post via XML-RPC using a very simple script.

I've tracked the offending code to about line 2050 in the core wordpress file xmlrpc.php, inside the mw_newPost() method.
```
$post_name = "";
if(isset($content_struct["wp_slug"])) {
	$post_name = $content_struct["wp_slug"];
}
```
We eventually need to get the trimmed slug value into $post_name, which is used later in the wp_insert_post() function around line 2235. It would appear to me that the easiest way to do that is to set the value of $content_struct['wp_slug'] via a filter or action hook. $content_struct is generated from the $args fed into mw_newPost().

There is an action hook near the top mw_newPost():

do_action('xmlrpc_call', 'metaWeblog.newPost');

which I have been trying to hook in to so that I can modify $content_struct['wp_slug']. Here's the content of my modified version of seo-slugs with my attempt at a hook:
```
// Add hook
add_action("xmlrpc_call", add_slug, 0, 1);

// This sets the slug
function add_slug($args) {
	global $content_struct, $wpdb, $args;

	$slug = strip_slug($content_struct['title']);
	$content_struct["wp_slug"] = $slug;

	return $content_struct;

}

// Slightly modified version of original seo_slugs() function
function strip_slug($slug) {

	global $wpdb, $content_struct, $stopwords;

	$seo_slug = strtolower($slug);

	$seo_slug = preg_replace('/&.+?;/', '', $seo_slug); // kill HTML entities
    // kill anything that is not a letter, digit, space or apostrophe
    $seo_slug = preg_replace ("/[^a-zA-Z0-9 \']/", "", $seo_slug);

    // Turn it to an array and strip common words by comparing against c.w. array
	$array = explode(" ", $seo_slug);
	$array2 = array_diff($array, $stopwords);

    $seo_slug = join("-", $array2);
	$seo_slug = ereg_replace(' ', '-', $seo_slug);

	return "$seo_slug";
}

$stopwords = array(
	'this',
	'the',
	'a',
	// plus a whole list of stopwords

);
```
However, the problem I'm having is that content_struct isn't global outside the mw_newPost method and it can't be read or written to by add_slug(). If I edit xmlrpc.php manually to declare $content_struct to be global at the top of mw_newPost, then the code works and the slug is trimmed and saved correctly when posting via XML-RPC. However, I don't want to edit a core file, but would prefer to make a proper plugin so there are no problems with later updates to wordpress.

What am I doing wrong? How do I make the value of $content_struct readable and writeable without editing xmlrpc.php? I have also read about adding filters to xmlrpc methods (like http://josephscott.org/archives/2008/11/adding-xml-rpc-methods-to-wordpress/ ) - should I be trying to do this instead?
Joseph Scott
Member
Posted 2 years ago #

Why aren't you using the wp_insert_post_data filter? That is called by wp_insert_post and should work no matter how the post is added.
Waddle
Member
Posted 2 years ago #
Thanks for you reply.

I've delved into wp_insert_post and I think the best place to strip the stop words from the slug is about line 1650 in wp-includes/post.php.
```
if ( !isset($post_name) || empty($post_name) ) {
	if ( !in_array( $post_status, array( 'draft', 'pending' ) ) ) {
		/* HERE IS THE SPOT WE NEED TO SET THE POST NAME */
		$post_name = sanitize_title($post_title);
}	else {
		$post_name = '';
	}
} else {
	$post_name = sanitize_title($post_name);
}
```
Within the sanitize_title() function in wp-includes/formatting.php, there is a hook I want to use:
```
function sanitize_title($title, $fallback_title = '') {
	$raw_title = $title;
	$title = strip_tags($title);
	$title = apply_filters('sanitize_title', $title, $raw_title);

	if ( '' === $title || false === $title )
		$title = $fallback_title;

	return $title;
}
```
By setting the post slug here in sanitize_title, rather than later in wp_insert_post_data, it can be passed wp_unique_post_slug() later on so we can be assured of a unique post slug.

This is my modified plugin code:
```
add_filter("sanitize_title", strip_slug_san, 0);
function strip_slug_san($title) {
	global $title, $raw_title, $data, $post_title;
	$title = strip_slug($title);
	return $title;
}
```
However, the $title from sanitize_title() is not being passed to my strip_slug_san() function. I would have thought that it would be, considering that it's being called from within the sanitize_title() function. I feel I'm close, what am I doing wrong?

Waddle
Member
Posted 2 years ago #

I'm slightly further along and royally stuck.

Here's the complete plugin code I'm using:

add_filter("sanitize_title", strip_slug_san, 0);
function strip_slug_san($title) {

	$title = strip_slug($title);

	return $title;
}

function strip_slug($slug) {

	//global $wpdb, $content_struct, $stopwords;
	global $stopwords;

	$seo_slug = strtolower($slug);

	$seo_slug = preg_replace('/&.+?;/', '', $seo_slug); // kill HTML entities
    // kill anything that is not a letter, digit, space or apostrophe
    $seo_slug = preg_replace ("/[^a-zA-Z0-9 \']/", "", $seo_slug);

    // Turn it to an array and strip common words by comparing against c.w. array
	$array = explode(" ", $seo_slug);
	$array2 = array_diff($array, $stopwords);

    $seo_slug = join("-", $array2);
	$seo_slug = ereg_replace(' ', '-', $seo_slug);

	return "$seo_slug";
}

$stopwords = array(
    'a',
    'the',
    'and',
    'etc'
);

I'm hooking into sanitize_title and the slug is being stripped of its stopwords successfully. However, when I click on any of the links to the posts on the index page, I get a 404. If I comment out
$title = strip_slug($title);
then posts are added normally (without the stop words being stripped) and can be retrieved normally.

I've checked the wp_posts table and the post_name values are set correctly when I'm trimming the stopwords.

Any ideas?

Joseph Scott
Member
Posted 2 years ago #

Sounds like a rewrite problem.

Waddle
Member
Posted 2 years ago #

I got it working. I abandoned trying to hook into sanitize_title, because it appears to be called multiple times per page and the scope was difficult. I tried using wp_insert_post_data and it works fine. Even if you make two different posts with the same title, the slugs are still appended with numbers correctly despite where wp_insert_post_data appears in the script.

Here's my working code in case it helps anyone else trying to set post slugs using XML-RPC:

add_filter("wp_insert_post_data", strip_slug_san, 0);
function strip_slug_san($data) {
//	global $data;

	$slug = $data['post_name'] ;

	$slug = strip_slug($slug);

	$data['post_name'] = $slug;

	return $data;

}

function strip_slug($slug) {

	//global $wpdb, $content_struct, $stopwords;
	global $stopwords;

	$count = count($stopwords);

	$seo_slug = strtolower($slug);

	$seo_slug = preg_replace('/&.+?;/', '', $seo_slug); // kill HTML entities

    // kill anything that is not a letter, digit, space or apostrophe
    $seo_slug = preg_replace ("/[^a-zA-Z0-9- \']/", "", $seo_slug);

    // Turn it to an array and strip common words by comparing against c.w. array
	$this_array = explode('-', $seo_slug);

	$array2 = array_diff($this_array, $stopwords);

    $slug2 = implode("-", $array2);

	return "$slug2";
}

$stopwords = array(
    'a',
    'the',
    'and',
    'etc'
);

Thanks for your tips Joseph.

Topic Closed

This topic has been closed to new replies.

WordPress.org

Forums

[Plugin: SEO Slugs] Setting post slug when posting via XML-RPC (6 posts)

Topic Closed

About this Topic

Tags

Code is Poetry