rom174

Should I forget getting an answer about this issue ?

Hello,
Just to know if you have good news about it ?

Hello,

I just had the warning and replaced line 6249 in index.php as follow :

To be replaced :

'author' => ($a['author_name'] != NULL) ? $a['author_name'] : FALSE,

Replaced by :

'author' => array(

	'@type' => 'Person',

	'name' => ($a['author_name'] != NULL) ? $a['author_name'] : FALSE

),

Tried it in rich result and it’s all green.

No problem for the delay.

Okay I will make it through screenshots instead.

1/ So first go to:
/wp-admin/options-permalink.php
Set a custom name for tags prefix

View post on imgur.com

2/ Then create a new tags, check that the custom prefix is there with the tag name.
That creates a 3 parts URL

View post on imgur.com

3/ Load the page associated to this tag, check that the custom prefix is in the URL bar and see there is no breadcrumb associated to the custom tag prefix.

View post on imgur.com

4/ Check the source and see that the json is only generated with 2 positions instead of 3 while the url has the 3 parts

View post on imgur.com

Let me know if it is better description of the issue.
Thanks

In fact, it is pretty easy to reproduce, just add a string in the tag field in the permalink structure page in order to customize it. Then create a tag.
The corresponding page of this tag will have a breadcrumb trail (and markup) that doesn’t reflect the url. Missing the custom string added.

Customizing the permalink turns the URL for tags into a 3 parts trail but the plugin seems to interpret it as if it hadn’t been customized (2 parts).

I created a page corresponding to this custom string in order to see if it “activates” the breadcrumb but no success.

OK thank you for explaining this.
In fact, I changed the default behvior thanks to custom permalinks in order to respect this logic :

https://home.com/
|___directory1/
|___index.php –> (1) url : https://home.com/directory1/
|___page-name1.php –> (2) url : https://home.com/directory1/page-name1

(1) Index of the directory never displayed so the url ends up with a trailing slash

(2) Page of the directory but not a subdirectory. As extensions of files are not displayed, it ends up with a url without extentions and without trailing slash.

But I understand the weak point of this way to work is that no automatic redirect can be set up either to add or remove the trailing slash following the settings. Only the canonical tag indicates the correct url but this seems to be not handled by WP for categories and tags pages in order to avoid duplicates (a behavior that I don’t understand). Maybe here is why it is one or the other.

Thanks again Corey

Absolutely no problem with this, you are helping.

is_wrong_permalink_structure was the last occurence of string I searched yesterday across all files of the plugin but I would have not imagine deleting this part.

Thank you very much, it works perfectly.

If this scenario (mixing both URLs) has not been implemented before, can I ask you if there is a deep reason from an SEO point of view or something else for having to choose to use either the trailing slash or not on URLs of a whole site ?
Can it be wrong to mix it ?

Following your answer to these questions, yes it would be a great to have a setting for using with or without or mixed trailing slash.

Thank you for your reply which gives true clues on where to look, unfornately I didn’t manage to bypass the cache.

Do I have to remove the setting lines in settings file :

  'use_trailing_slashes'
  'permalink_structure'

or to set it up in some ways ?
Is it okay to place the filters in functions.php setting it to false ?

Hello guys,

I post my question at the end of this thread because it seems I have the same issue, except that mine looks to be more consistent.

First, I am using cache enabler along with autoptimize. I do not use adinsert plugin.

Here is the behavior : every ending pages, by ‘ending pages’ I mean last part of the trail/url/current hierarchy, are well cached but not the ‘middle’ ones, the directory/archive pages that are categories. In fact all urls with a trailing slash are not cached. They do not print the https-index.gzip signature and are not present on the server (while I am seeing the generated cached pages if I am in the very last folder of a hierarchy).

Knowing this and having seen the permalink_structure setting line in the setting file in /settings/cache-enabler/, I wonder, if the issue is linked to this, what is the parameter I can try instead of the current one ‘no_trailing_slash’ ?

Do you think there is a link between the final slash and the fact those pages are excluded from caching ?
Can it be something else ?

Thanks a lot for your advice

EDIT: removing the trailing slash does generate the cache and found that putting has_trailing_slash works. But it is marked as deprecated, what can I do instead ?

• This reply was modified 3 years, 6 months ago by rom174.
• This reply was modified 3 years, 6 months ago by rom174.

Thank you @wfpeter

Hello @wfpeter

Thank you for getting back, making it clear and for your proposition !

Yes that would be awesome, this suggestion would use a behavior coming from websites admins :

After a certain amount of time, when they identify some trivial usernames always used by bots, like [admin], [NAME-OF-SITE-WITHOUTCCTLD], [AUTHOR-NAME-IF-DIFFERENT-FROM-ITS-USERNAME] it could be added in a section of the settings of Wordfence or auto-added after a number of failed logins. Then it could be added to a json or directly in the login page/login form source code to compare and disable the button without even having to request the database. In that way, bots would fall in a trap because even if, later, they are able to retrieve those blacklisted username, they wouldn’t know what to try. What do you think ?

I know it might be longer than just that because some new functions could be needed for instance a blacklist compare function when creating a new account or to alert that the blacklisted username is already used but if it is possible, that could be a great feature !

This scenario is only based on what I noticed about bots behavior from my end, and, I guess will only work if bots submit login form using the login button as humans.

Well, maybe I did not set it up correctly but in fact it seems it is not working as I thought as it is preventing them for trying again, while I would like to stop them before they even try to send the login request.
In fact, I realized that there are only 2 or 3 usernames that bots are using to attempt to login. Each time they do it, the form is submitted and the server comes back with fail or success response. If it fails, they are locked for a certain amount of time. But as soon as they change their IP or that another bot comes up, the login form is submitted again, what I wonder is to disable/hide/block the login button to avoid the form from even being sent if the username field doesn’t match an existing user OR if it matches one of the 2 or 3 usernames always used by bots.

The idea would be to pre-check during the process instead of waiting for the login click and the wrong server response that is always (with those usernames attempts) going to fail. During the process and just before the bot sends the form, we could imagine checking if the username is in an ‘allowed list’ and not even let them submit the form if it is already in this list.

Hello @wfpeter ,

Thanks a lot for your reply, I will try it.

Have a nice day!

• This reply was modified 4 years, 3 months ago by rom174.

So the better way is to put everything in the head ?

Well, thank you very much for your answers, your patience and your plugin !