Re-checking all pages or re-checking selected links misses broken links

  • Resolved abitofmind

    (@abitofmind)


    3 years, 2 months ago

    ?? Screen recording for a convenient and detailed impression of what’s described below as text.

    ENVIRONMENT
    – WordPress 5.8
    – Broken Link Checker 1.11.15
    – Plugin “Redirection” was once installed for some tests. But is now inactive. The whole website is in construction and currently behind a login wall. I have no external incoming link or SEO concerns yet. But maybe some redirections are still persisted somewhere? In .htaccess itself nothing extra-ordinary is persisted.
    – open_basedir was ON during install time and during the initial broken link checker run. Later I turned it off in my webhost setting but it nevertheless seems to be effectively on many hours later.

    REPRODUCTION I
    1) Target page: Created at /skills/tools/snagit
    2) Source page: At / “Home” created a link with text “Snagit” to /skills/tools/snagit
    3) Target page: Changed slug and therefore path /skills/tools/snagit to /skills/tools/snackit which may become necessary somewhen in the life of a pretty-url.
    4) All incoming links are now broken, i.e. the one from the homepage.

    Maybe the scheduled run would have caught those correctly. But for testing the plugin I did not want to wait for the next scheduled run but instead wanted a manual re-check. Both offered options failed to detect the broken link. Even a re-install didn’t help.

    ATTEMPT 1:
    – Opened: Tools > Broken Links > Detected Links > All > Sorted by source.
    – At source / “Home” I saw a target link /skills/tools/snagit/ which we know to be broken.
    – Clicked its “Re-check” action
    –> Still falsely shows as 302 OK.

    ATTEMPT 2:
    – Settings > Link Checker > Advanced
    – There I clicked “Re-Check all pages” which I let run until completed.
    – Opened freshly: Tools > Broken Links > Detected Links > All > Sorted by source.
    – At source / the link to target /skills/tools/snagit/ falsely still shows as 302 OK.
    Despite when webbrowsing / and clicking /skills/tools/snagit/ you get a 404 not found.

    ATTEMPT 3:
    – Uninstall Broken Link Checker.
    – I checked in phpMyAdmin and all tables were indeed deleted.
    – Reinstalled, activated, waited for initial run to complete.
    –> Still no broken links gets caught.

    REPRODUCTION II
    – Additionally renaming in the middle part of the page hierarchy:
    /skills/tools/snackit/ gets /skills/toolz/snackit/
    – Both attempt 2+3 catch nothing.

    REPRODUCTION III
    – Changing slug itself back but the middle part of the page hierarchy remains renamed
    /skills/toolz/snackit/ gets /skills/toolz/snagit/
    –> Now redirection works!
    – Clicking link /skills/tools/snagit/ redirects to /skills/toolz/snagit/
    – Only the middle part of the link differed to the current URL, and start and end of the URL were correct.
    – WP core’s fuzzy logic seems to handle this scenario at least.

Viewing 6 replies - 1 through 6 (of 6 total)
  • Thread Starter abitofmind

    (@abitofmind)

    Btw, ??Broken Link Checker was successful with finding broken external link during initial scan a video which I recorded some days ago before today running into the issues with re-checking failing to detect broken internal links.

    Plugin Support Nithin – WPMU DEV Support

    (@wpmudevsupport11)

    Hi @abitofmind,

    I gave a quick test with such a permalink structure but wasn’t able to replicate this in my system. Once I changed the slug of the page and re-checked the URL in the plugin it did pick the URL as 404.

    Seems like it’s a site-specific issue, I could also notice you mention the Nested pages plugin could you please check whether you could replicate the same if the Nested Pages plugin is temporarily deactivated? A similar URL structure can be created using Parent pages.

    If the issue still exists even after that then please do let us know how the Broken Link Checker plugin behaves if you temporarily deactivate all the plugins except the Broken Link Checker plugin and also switch to a default WordPress theme and see how that goes?

    This would give a better idea of what might be causing it. Please do let us know how that goes.

    Looking forward to your response. Have a nice day.

    Kind Regards,
    Nithin

    Thread Starter abitofmind

    (@abitofmind)

    Dear Nithin, thanks for your info, I will try to isolate the bug by first disabling NestedPages, then further plugins, and last with the default theme.

    And after each configuration change I will run “Re-check all pages” (a.k.a. the “nuclear option” as you call it drastically ?? in the Settings).

    And then tell you what behavioral changes I noticed after the configuration changes.

    Thread Starter abitofmind

    (@abitofmind)

    FOUND THE PLUGIN CONFLICT

    See as a ?? screen recording or read the description below.

    I use the plugin My Private Site which hides the entire website behind a login wall for anonymous users.

    Ideal for WordPress pages which are in the early building process and not yet ready for the eyes of the public and the spiders of the search engines.

    It redirects all anonymous requests for example.com/* to https://example.com/wp-login.php?redirect_to=YourRequestedURL with HTTP status code 302 “Found” (also known as “Moved temporarily”).

    And Broken Link Checker seems to conduct its link tests simply as an anonymous user.

    As soon as “My Private Site” is de-activated, it detects the broken internal links correctly.

    PROPOSAL FOR BETTER COMPATIBILITY

    Settings > Link Checker > Advanced

    Gets this new setting:

    “Check links as”: [ Anonymous User ? ]

    Offers choices: Anonymous User | Admin | Editor | Author | Contributor | Subscriber | Specific User

    When choosing “Specific User” then you get a text field nearby with the hint text “username”.

    I imagine that WordPress has an API/calls for “request as…” and maybe also some sort of “sudo” hence no need to specify the password too. If that’s not the case then also offer a password field.

    • This reply was modified 3 years, 2 months ago by abitofmind.
    • This reply was modified 3 years, 2 months ago by abitofmind.
    Plugin Support Dmytro – WPMU DEV Support

    (@wpmudevsupport16)

    Hey there @abitofmind

    Awesome troubleshooting job, and glad you found the conflict.

    And that’s a pretty cool feature idea! I’ve sent it over to the developers for consideration for a future update. ??

    Cheers!
    Patrick

    Thread Starter abitofmind

    (@abitofmind)

    Also glad that I was able find it and glad to read that you are interested in the fix for it.

    Until then I will as a workaround turn off the login wall temporarily, run the test job, heal broken links if needed, then put the login wall up again. Hopefully the website will soon anyhow be in a good enough state to open to the public.

    Please let me know what your team plans to do about it!

Viewing 6 replies - 1 through 6 (of 6 total)
  • The topic ‘Re-checking all pages or re-checking selected links misses broken links’ is closed to new replies.