Loggy

Problem solved. I checked another site on the same server (also of course running nginx and including the same fcgi file) and the Yoast indexing procedure worked.

The difference turned out to be a line:

rewrite “^/([^.]*[^/])$” /$1/ permanent;

which had been there for ages and was never a problem before. It clearly rewrites the URL by appending an automatic “/”. I commented it out and now Yoast works fine.

Jetpack still finds an error but I don’t need that – I have Akismet as a plugin anyway.

Eventually I will remove as many plugins as possible but at least I can now get on with other stuff.:)

I’ve got this message when I try to cross-reference my quite large site:

Oops, something has gone wrong and we couldn't complete the optimisation of your SEO data. Please click the button again to restart the process. If the problem persists, please contact support.

Below are the technical details for the error. See this page for a more detailed explanation.

Error details
Request URL
https://www.mysite.com/wp-json/yoast/v1/indexing/prepare

Request method
POST

Status code
404

Error message
No route was found matching the URL and request method.

The set up is v5.9.2 under nginx 1.18.0, yoast premium 18.2.1, wordfence 7.5.9, ithemes security nginx config patch – the current version (the old one also fails).

The error 404 should only occur (apart from a genuine missing page) when looking for readme.html or license.txt. All other issues give different return codes.

I have the permalinks set to /slug and both server and caching caching are switched off.

It’s a new site but has a lot of imported posts which need cross-referencing so Yoast should be the tool. Jetpack also won’t configure either so it seems to be a REST API issue.

It’s happening to me on a completely up-to-date site.

I don’t think it can be a browser issue but I did find that if I add tags via the Quick Edit, they don’t get corrupted.

This suggests that the issue is somewhere in the widget coding – probably the Javascript/JQuery.

Er – I can’t see where to register or login. I can login with my FB login but this only allows me to comment on posts rather than to start a new topic on the forum.

BTW this is the dump from theme-mods-twentyfourteen-child in the option table:

Array
(
    [0] =>
    [header_textcolor] => blank
    [background_color] => ffffff
    [background_image] =>
    [background_repeat] => repeat
    [background_position_x] => left
    [background_attachment] => fixed
    [nav_menu_locations] => Array
        (
            [primary] => 4
            [secondary] => 0
        )

    [featured_content_layout] => slider
    [header_image] => https://www.parrswoodpta.org/wp-content/uploads/2014/01/cropped-PTA-header-smaller.jpg
    [header_image_data] => stdClass Object
        (
            [attachment_id] => 976
            [url] => https://www.parrswoodpta.org/wp-content/uploads/2014/01/cropped-PTA-header-smaller.jpg
            [thumbnail_url] => https://www.parrswoodpta.org/wp-content/uploads/2014/01/cropped-PTA-header-smaller.jpg
            [height] => 178
            [width] => 1260
        )

    [fourteenxt_center_site] => 1
    [fourteenxt_primary_menu_float] => left
    [fourteenxt_hide_left_sidebar] => 1
    [fourteenxt_overall_content_width] => 1038
    [fourteenxt_fullwidth_blog_feed] => 1
    [fourteenxt_fullwidth_single_post] =>
    [fourteenxt_fullwidth_archives] =>
    [fourteenxt_fullwidth_searches] =>
    [fourteenxt_content_off_featured_image] =>
    [fourteenxt_content_width_adjustment] => 874
    [fourteenxt_feed_cat] => 0
    [fourteenxt_home_excerpts] =>
    [fourteenxt_excerpt_length] => 55
    [fourteenxt_fitvids_enable] =>
    [fourteenxt_fitvids_selector] => .post
    [fourteenxt_fitvids_custom_selector] =>
    [fourteenxt_featured_visibility] =>
    [fourteenxt_enable_autoslide] =>
    [fourteenxt_slider_width] => 1600
    [fourteenxt_slider_height] => 500
    [fourteenxt_slider_topmargin] => 0
    [fourteenxt_featured_bg_visibility] =>
    [fourteenxt_maximum_site_width] => 1260
    [fourteenxt_overall_image_height] => 572
    [fourteenxt_content_top_padding] =>
    [fourteenxt_sidebar_top_border] =>
    [fourteenxt_body_class_filters] =>
    [fourteenxt_layout_mobile] => grid
    [fourteenxt_slider_transition] => slide
)

Thanks for your prompt response – sorry I was out yesterday.

I have copied the HTML and also screen dumps for the three cases: 2014 child theme with plugin, 2014 child theme on its own and 2014 parent theme.

The child theme is very simple – I have just removed the “Category Archives” line in category.php, modified the footer in footer.php added the above function into functions.php and the following style.css:

/*
Theme Name: Twenty Fourteen Child
Template: twentyfourteen
*/
@import url("../twentyfourteen/style.css");

I also added the complete css directory from the parent which contains ie.css and editor-style.css but this doesn’t make any difference either.

Is there a better way to get the HTML and .gif images to you? Reply privately if you want. The latter are quite big because they include the header image.

PS I have copied the ie.css over from the parent theme BTW and added the following coding:

function mytheme_dequeue_styles() {
   wp_dequeue_style( 'twentyfourteen-ie' );
 }
 add_action( 'wp_enqueue_scripts', 'mytheme_dequeue_styles', 11 );
//add new from child theme
wp_enqueue_style( 'mytheme-ie', get_stylesheet_directory_uri() . '/css/ie.css', array( 'twentyfourteen-style' ), '1.0' );
$wp_styles->add_data( 'mytheme-ie', 'conditional', 'lt IE 9' );

Landing Page modifies the .htaccess file, which it doesn’t say in the documentation. This may be OK as long as (a) the file is writeable and (b) obviously apache is in use. I am using apache and nginx in reverse proxy and on installing this plugin got the message:

400 Bad Request
Request Header Or Cookie Too Large
________________________________

nginx

It also caused the Google Analytics Dashboard to fail.

I deactivated it and everything returned to normal but deactivation leaves the .htaccess file modified with the following three lines:

RewriteCond %{QUERY_STRING} !lp-variation-id
RewriteRule ^go/([^/]*)? /wp-content/plugins/landing-pages/modules/module.redirect-ab-testing.php?permalink_name=$1  [QSA,L]
RewriteRule ^langing-page=([^/]*)? /wp-content/plugins/landing-pages/modules/module.redirect-ab-testing.php?permalink_name=$1 [QSA,L]

In my configuration .htaccess file is correctly obeyed as nginx is only used to deliver static files and cached files.

If a plugin will only work with apache and if it modifies the .htaccess file in any way, this should be stated. It should also return .htaccess when deactivated. And as many people are now using nginx either raw or in reverse proxy, it may be worth providing some rewrites in that syntax.

Thanks @mc_will.

I add spf and rDNS records to my own nameserver but if the mail goes by API link to Mandrill and then out to the various recipients, won’t the next mail router down the line see a mail that says From: [email protected] but coming from the Mandrill IP number?

So don’t the Mandrill spf/rDNS records need changing?

Or doesn’t it work this way? ie does it leave the source IP number as my IP number which will match my spf/rDNS settings?

I updated the plugin and now liking with Facebook doesn’t (a) include the photo and (b) the meta description – it just repeats the link.

I’ve reverted to 2.9.5 (as modified by me to remove the extra line feeds) for the moment although any database change will not be reverted – I just moved the new coding and copied the old tree back.

Any idea why this is happening? I have to ask someone to unlike a post to see whether reverting will fix things but the picture (featured image) and meta description were copied to FB correctly yesterday.

Puzzled.

You can see if this fixes things at https://www.themoneyprinciple.co.uk on the latest post about nourishing your life.

Thanks @wt. I will wait the update and also the async buttons which IMHO will make RSFTS the share plugin of choice….

Another mod could be to store the current counters and only update when a Like button is clicked or every day using a cron job. This would mean that the counters could be delivered straight from the database with no calls at all. They would be out of step only if someone did a like by another route and cron could be sort that out.

I’m not too sure this is a good idea anyway.

A blog with just one user (hopefully not called admin) may well need an email address that is specifically not from the domain. That way, if the domain falls over (eg expires or there is a registration/nameserver problem) the admin can be warned.

Or if you have multiple users posting, they may not have or want to have [email protected] addresses.

A better (and I think simpler) solution would be to have a SENDFROM parameter in the setup that the user can alter. That way, the name server rDNS and spf records should be satisfied. And send from could comrise USERNAME and BLOG_DOMAIN.

I see there is already EMAIL but that is just the default.

I’ve noticed this over the last week for some users and have regressed TML to 6.1.4 too that I found on my backups.

The experience is that a user logs in on the home page of the site using the TML widget and that’s fine. Then going to the SP forum they need to login again except they just get redirected to the home page again.

And in the logs there is a slew of undefined variables which I initially thought was to do with the problem but maybe not.

I am not sure whether regressing TML will fix the issue but I can’t access the SP support at the moment.:-(

Actually we don’t really need the TML login at the moment – that is just for the future.

Well Chrome 12.07 on Linux. and FF 13.01 on XP don’t seem to work. An ancient FF on an old Linux system and Opera on XP do work as does IE 8 on XP. At least I can get something done on one machine or another.

Belay that!

It’s a problem with Chrome JS I think. Just opened the site in Firefox and it seems OK at least on one computer…:-)