Which js-encumbered sites to fix in the first place

You are awesome!!!

> "Is wrapping one of the links in a video tag sufficient? Or do you want something better (using other, higher-quality stream or sth?)"

Yes, wrapping the links in a video tag would do it. I can also just download.

>"From barely looking at page's source I know how to get an .mp4 link for Odysee video and I can write code to put that in a video tag."

That's incredible! Yes, search would be crucial, but this is a big start.

>"Making benchmarks re-appear on Phoronix is trivial. Will do. As to login and posting - haven't looked into it yet"

You are the best! Let me know if/when you'd like me to test things out?

Besides what I wrote in another post that you might have downloaded HTML instead of JSON, there's also 1 thing I should clarify: what should be copied into the javascript console is not the fix JSON itself, but rather the contents of its script. So, taking as an example my bank's credit card payment confirmation page fix (because it is shorter than the others), what you would be pasting into the console is:

/**
 * Copyright 2021 Wojtek Kosior
 *
 * Available under the terms of Creative Commons Zero.
 */

const submit_button = document.getElementById("submit");
submit_button.classList.remove("disabled");
submit_button.removeAttribute("disabled");

console.log(document.querySelectorAll(".noscript"));

for (const noscript_element of document.querySelectorAll(".noscript"))
    noscript_element.remove();

and not:

[{"scentrum24 Santander 3D-Secure":{"url":"","hash":"","text":"/**\n * Copyright 2021 Wojtek Kosior\n *\n * Available under the terms of Creative Commons Zero.\n */\n\nconst submit_button = document.getElementById(\"submit\");\nsubmit_button.classList.remove(\"disabled\");\nsubmit_button.removeAttribute(\"disabled\");\n\nconsole.log(document.querySelectorAll(\".noscript\"));\n\nfor (const noscript_element of document.querySelectorAll(\".noscript\"))\n    noscript_element.remove();\n"}},{"phttps://acsv.centrum24.pl/ACS/servlet/ACSAuthoriz":{"components":["s","centrum24 Santander 3D-Secure"],"allow":false}}]

nor the HTML from CGit

I think I know what's the problem. The link I gave you leads to a cgit-generated HTML page of the JSON file with the fix.

You probably just wget'd or curl'd the link I gave you which I didn't assume you would. If you open it in a browser, you'll see a link to download the actual raw .json.

Anyway, the link for the raw JSON file you can just wget is as follows:
https://git.koszko.org/hachette_fixes_tmp/plain/phttps___odysee.com.json

Also, I noticed a mistake that causes the video not to appear on Mozilla-based browsers (I previously tested on Ungoogled Chromium). I'll push the fix within a few minutes.

EDIT: Another update. I implemented search for Odysee. Already pushed the enhanced fix to git.koszko.org (the same link as before). The fix now works as follows: if it detects the current page is an Odysee video page, it shows the video. Regardless of that, it always shows a search input. Search results appear without reloading, on the same page, below the search button. If my script at the start of execution detects the current page is an Odysee search page (i.e. one with url like https://odysee.com/$/search?q=something), it automatically performs the appropriate search as if you submitted it manually and shows the results. If you click on one of the results, it leads you to another page. If there aren't any fancy unicode characters in video name (not to be confused with video title), the video should appear there. For now, the fix only handles videos and not channels

This is truly awesome, thanks! And congratulations.

You might rename the add-on "Magic Hachette" :)