[Off-topic, but here goes]
My thinking was that the main issue is that YT now required versions of SSL/TLS that weren't supported by OpenSSL 1.0.0, the version supplied with the stock 2010-era firmware. Looking back, when the HD-Fox YT app was withdrawn, the APIs going https-only was given as the reason in these forums. However, the YT APIs used by the Humax app might also have been broken since, as I thought, YT had already cast off the app by requiring TLS1.2.
At least the first part of that thinking was not accurate.
Currently accessing YT with a browser that it doesn't like (eg, Humax UA "Opera/9.80 (Linux 7405b0-smp; U; HbbTV/1.1.1 (; Humax; HD-FOX T2; 1.03.02; 1.0; ); ce-html/1.0; en) Presto/2.10.250 Version/11.60") redirects to a browser-not-supported page on the first attempt, after which it sets the 'hideBrowserUpgradeBox' cookie flag. This works even with only TLSv1.0 enabled. However the site then serves JS that can only be understood by very recent browser versions (
the lunatics have taken over the asylum). The pages returned by the YT API that the Humax API used wouldn't have demanded such advanced JS (if any).
What actually happens with
the new-portal version of the YT app is that the left-hand menu appears but anything that requires getting data from YT fails with either 404 or a message that execution of an API request failed, quoting ('Most Popular')
Code:
Execution of request failed: https://gdata.youtube.com/feeds/api/standardfeeds/most_popular?start-index=1&max-results=10&paid-content=false&embed=allowed&syndicate=allowed&v=2&format=6,1&time=today&safeSearch=strict&orderby=published
In my desktop browser faked to be a Humax HDR, the quoted URL also gives 404.
So I conclude that Humax decided not to upgrade the app to a new version of the YT API and that is why it doesn't work. SSL/TLS versions seem to be a red herring.
As to introducing a newer version of OpenSSL (which would be good even if it didn't save the YT app), in the Humax s/w the settop blob is dynamically linked against OpenSSL 1.0.0, ie TLS <=1.1. The Opera library seems to know about TLS up to 1.2, but isn't dynamically linked against OpenSSL. Introducing a newer version of OpenSSL would at least require a shim layer, which I think is a
research problem: I think that's a no.