Yes, I noticed this. Perhaps due to different handling of logins on both the sites. This could be handled in the backend by not redirecting as well, but therd should have been design choices there.
> Currently, the script looks like this
>
> #!/bin/sh
> 0x $1|yy084|yy030|yy073
>
> "0x" is another Almquist shell script that can search about 63 different servers that return search results, e.g., Google, DDG, and so on.^1
>
> yy___ are UNIX filters written in C.
>
> 0x uses some yy proggrams as well, e.g., yy025, along with sed and a TCP client, e.g., tcpclient, netcat, socat, bssl, openssl, etc.
>
> yy084 outputs SERPs as SQL.
>
> This makes it easy to create simplified "mixed" SERPs with results from different servers.
>
> Where possible 0x allows for "continuation search". Going past page 1 of SERPs is discouraged or even prevented in recent times, all focus is on "the top result",^2 and some www search engines actively try to block exhaustive research and discovery. By continuing searches over time, e.g., page 1 of results on day 1, page 2 on day 2, page 3 on day 5, etc., one can sometimes avoid being blocked when doing exhaustive searches.
>
> 1. This is an ongoing experiment. Sometimes a site will "break" if the site operator changes something but this does not happen too often. Majority have remained stable over time.
>
> 2. This coincidentally benefits an advertising services racket.
How can I get more details about these filters. Are there existing implementations somewhere online which I can test? Seems pretty interesting.