sub-semalt

[wilderness]

anybody have a clue if this correct syntax?

#any two numbers
RewriteCond %{HTTP_REFERER} ^http://[0-9]{2}\.semalt\.com/

[iamzippy]

It works in Regex Buddy.
So does:

RewriteCond %{HTTP_REFERER} ^http://\d\d\.semalt\.com/

[wilderness]

Many thanks

[Pfui]

There are assorted subdomains (e.g.: http://semalt.semalt.com/) thus --

RewriteCond %{HTTP_REFERER} semalt
RewriteRule .* - [F]

-- works for me. (Ditto for fellow pest kambasoft.)

[ronin]

I'm still learning (and improving) my regex skills, but I am using this:

RewriteCond %{HTTP_REFERER} ^https?://([a-z0-9-]+\.)?semalt\.com [NC]
RewriteRule .* - [F]

[not2easy]

Couldn't these be combined like we do for UAs?

RewriteCond %{HTTP_REFERER} (kambasoft|semalt|whatever) [NC]
RewriteRule .* - [F]

[wilderness]

certainly.

[dupres01]

which is the better form to use (and, to help with my education, why)?
this one:
RewriteCond %{HTTP_REFERER} (kambasoft|semalt|whatever) [NC]
RewriteRule .* - [F]

or this one:
RewriteCond %{HTTP_REFERER} kambasoft [NC,OR]
RewriteCond %{HTTP_REFERER} semalt [NC,OR]
RewriteCond %{HTTP_REFERER} whatever [NC]
RewriteRule .* - [F]

[wilderness]

the combined line will be less server strain and slightly faster.

Both do however work.

I've one for "crawler" as well, however for simplicity sake and to possibly stop another stray bot, you could use "crawl".

[lucy24]

The form

#any two numbers
RewriteCond %{HTTP_REFERER} ^http://[0-9]{2}\.semalt\.com/

is syntactically correct, but I suspect it's easier on the server if you simply say

^http://[0-9][0-9]\.semalt\.com

That's assuming it will always be exactly two. Otherwise of course you'd go to

[0-9]+

Or-- my preference-- \d for a savings of three bytes ;)

fwiw, mine simply says

SetEnvIf Referer semalt keep_out

It's in mod_setenvif because this rule is in my shared htaccess used by all sites. If it were for a single site it would be expressed as a RewriteCond along with assorted other referer-based lockouts.

My impression is that semalt works 100% via infected human browsers, because they always ask for favicon and stylesheet. Robots normally don't. Did anyone ever figure out what they want?