Oh dear @wfphil

Therefore any files in the child theme directory cannot be repaired as they will be treated as unknown files.

That is not how Wordfence works in the scan. It identifies all occurrences of infected files. Then it offers an individual repair option (which is not what I’m writing about). Or “REPAIR ALL REPAIRABLE FILES” which I have explained works absolutely fine and is truly wonderful.

Now, without examining the code in detail, I assume Wordfence creates an array of deletable files as it scans. Then when it repairs repository files it removes them from the deletable array. But child theme files stay in the array. So they get deleted (unless you follow my workaround).

Whoever deemed this “expected behaviour” needs to examine the logic. Or at least offer a realistic explanation. Because I cannot see any value in repairing a file only to subsequently delete it. And please don’t repeat that Wordfence cannot repair those files. Because I’ve examined them (before setting the Ignore flag in my workaround) and they are fully repaired – for which I thank you very much.

Thank you very much for your fast response @ryankienstra

I’ve since noticed that this also applies to pdf attachments. But I’m not familiar with GitHub. So I wonder if this should be commented there, or raised as a new issue?

Thanks again.

Thank you Joachim.

Hi, its:
Wordfence Security – Version 6.1.7

And it was updated as soon as it came out. I don’t know when the problem started though. I have not changed any Wordfence settings since 6.1.1 when the firewall started.

I’ve tried all the diagnostic options, including remote scan. It always fails at the same place I reported above – i.e. whatever it does after completing scan of plugins and themes.

Keith

Just a minor point. Integration fails with long messages. I’ve narrowed it down to:

if ( 500 < strlen( $excerpt ) ) {
			$excerpt = substr( $excerpt, 0, 500 );
		}

I don’t know what the actual limit is, and I cannot find it in the Slack docs. 160 works. 175 doesn’t. If it’s good enough for Twitter…

Thanks Brian,

I’ve thought about the logic of this, and I now think you have it absolutely right.

If I want email, all I have to do is remove the blacklisted usernames. Then they’ll trigger the lockouts from invalid username.

Maybe the documentation could be clearer, but it’s a minor point. Your plugin is fantastic. If I was only allowed 1 plugin it would be Wordfence.

Thanks for your pending action on the fake crawler alert.

Thanks for pointing out that alternative way for IP ranges.

I’m 99% sure that all the chaching is disabled/cleared, but I’ll keep an eye on it after every reboot (only a fool is ever 100% certain about caches)

Thanks for making things better – you’re very good at it.