Greasy Fork

Was broke due to google, someone in the comments of this found the issue cause and gave a solution that works

What browser are you using, as well as what addon to run it? Things are still working for me on Firefox and Chrome with Tampermonkey.

Yes, the script doesn't work anymore. Chrome (70.0.3538.110) + TM.

I'm using Chrome 85.0.4183.102. Try updating first and see if that resolves the issue.

Alas, in recent versions of Chrome, the flags I need are missing. I checked on portable version 85.0.4183.83. Does not work.

I cannot replicate it no matter how hard I try. Can you give a page example? As well as any and all addons you have running and other scripts that are also running on said google page?

I checked it on a clean browser. Installed only TM and this script. A regular search page.

That is extremely peculiar. Even in a fresh browser myself, I can't replicate this. If I had to guess, it's selective releasing. They did the same thing when they pushed the changes that made this script exist in the first place out. They did it slowly with specific users and IPs before eventually pushing it on everyone.

This makes it extremely difficult to diagnose the problem though if you're currently unaffected by said change as it hasn't reached you yet. May I request from you a saved HTML complete zip for both the search page with the script active and not active? I'm not all that experienced I'll be perfectly honest, especially if working off of local files, but I'm determined to keep this script working as I too hate the changes. Additionally, the creator of the script hasn't touched it in quite some time. I'm not sure if we can rely on them updating it.

Oh, additionally, does it continue to not work if you are not logged into google? Often these selective changes are user-based. Beforehand, not being logged in returned it to a previous state. It's at least worth a shot. Not suggesting this as a fix, mind you. I may need it for reference, however. Please make sure that the files you provide are while searching "test search" like in your example.

Yes, your account linking theory turned out to be correct. After cleaning the portable version of Google Chrome (85.0.4183.83), I installed TM and Better Google. This option turned out to be working.

I also checked the IP theory just in case. I did not reveal any patterns. Nothing depends on IP.

I made some saved HTML pages. The name indicates the conditions with which the search was made. I hope this is what was needed?

test search.zip

P.S. Yes, I also hate the changes that Google has been doing lately. The way they messed up the image search ... No, just no. Ew!

You were close, but I need complete HTML files. These usually come with a folder that holds images and stuff by the same name. It saves as complete HTMLs by default, so you probably just forgot to include them. Try again when you get the chance.

I know there should usually be a resource folder. But this time something went wrong. I first saved the page via the context menu and "save as ..." but the folder was not created. Therefore, I made copies of the pages through the "Save Page WE" extension. Alas, the folders were not created this time either. If you tell me how exactly you need to save the page for your requirements, I will do it.

P.S. Hmm, the art site saves images to a separate folder, but when you save the Google Search page, this doesn't happen...

That's odd. It should still create a folder due to scripts since they go in the folder too. What I do, with Chrome, is right click anywhere on the page that isn't a link, click "Save As..." And make sure where it says "Save as type:", it says "Webpage, Complete".

Yes, I get it, that's right. I had a different "Save Type" selected as the default. I changed it and now saves it with a folder. Tomorrow I'll make copies of the web pages, once I get to my home PC, and send it to you.

Do you know when you might be getting those?

If you want the "new" layout, browse to google.nl with this NID cookie: 204=fX1YTSuN4FebGoYQ-1XLjiMUJLbPsYuErcpt9z72qyDdnDDLa317_naULUXZHIskm4CaK3rV0ugJeIDU2nssORgKFsMg_kZMFAlmRSgTv6V5qEyJtBLztTb7mzeBcHI9hUKYMJJpaHTxtwifEuDCGqEU-oVriEYd1EKccMY9RCA

I have a couple of NID on other TLDs with the same effect.

From what I have seen after updating a stylesheet similar to "better google", they only changed several "class" names. The html seems identical. "better google" uses .r which is renamed in this particular layout.

If you need help fixing the code, I'll be glad.

Thanks for the assistance Achernar, it seems that the .r was the only thing changed. Got the script updated with the cookie you provided for use. Because I have no way to push an update to this script, I will upload the working version on my own account until the original script writer comes to update their own.

https://greasyfork.org/en/scripts/411886-better-google-updated

I proposed my help because I know an easy way to fix the code.

Just edit var linkEl = el.querySelector('.r > a');
to: var linkEl = el.querySelector('.r > a, .rc > .yuRUbf > a'); (with the .rc in case the new classname is elsewhere too)

and below, the line: el.querySelector('.r).appendChild(betterEl);
becomes: el.querySelector('.r, .rc > .yuRUbf').appendChild(betterEl);

querySelector will return the first matching element.

I have tested right now, and it works.

I'll update the script accordingly then. Thanks for the help, didn't realize you could do that sort of thing in JS.

What can be done in CSS selectors in stylesheets can be used in querySelector and querySelectorAll. Plus one or two extra commands.

I apologize, only today I got home from a business trip. I had to urgently leave the city for three days, because of which there was no access to a PC at all.