News for hackers. Stuff that matters.

exploraz

Re:In less than a year, Google made scraping 10× harder on Friday September 04, 2026 @09:14PM (#49572109)
It used to be like that when they still used "/url?..." redirect links, which have the destination URLs clearly exposed. Maybe it's an intended feature for checking out where does the link leads to. That behavior can be prevented by browser extensions.

Now, on desktop Firefox (private browsing mode, logged out), I see that it always links to the opaque goto URLs, instead of only overwriting on interaction.

Using Google Search while logged in on Chrome Android still presents the older non-opaque /url?... links though.

It's unfortunate that we cannot immediately inspect what the link will lead to, whether the link is the intended (and safe/expected) destination or not.

... for some reason (inside a script tag for me)

It never got overwritten back to the original destination URL with the same browsing environment (see my other comment).

I encountered that on Firefox (with privacy setting both on strict, and I think also on default protection last time I encountered that) both on Android and desktop, and also Safari (iOS). Not sure about Chrome though.

Tell HN: Google Search to begin using opaque redirect URLs on search results

Recently, for a couple of months, I noticed that Google Search began to replace its /url?q=<encoded url> links into a more opaque one (/goto?url=<internal identifier>) randomly (indication of A/B testing) on my logged-out private browsing sessions.The old one (intermediate link with /url?sa=...q=...) will either be presented immediately after loading the page, or after the mouse has...

( Read More... | 8 comments | 4 points | on HN )
Re:Nitter project received cease and desist on Wednesday August 26, 2026 @07:53AM (#49447409)
... for pushing a commit.

Disregard. Now it's archived again

Re:Nitter project received cease and desist on Wednesday August 26, 2026 @06:58AM (#49446854)
... and now it isn't as I'm writing this
Other hypotheses:

- VPN/Proxy/Datacenter IP/Tor usage

- Censorship bypass tools triggering destination server's firewall (suspicious TLS fingerprint or connection characteristics, or bad proxy IP)

- Bad ISP (e.g. one ISP IP address or range of addresses shared by so many users that could create collateral damage if one or more user gone rogue)

- Actual botnet/hidden residential proxy exit node on the same IP/local network (e.g. TV boxes or SDKs embedded inside apps that users are not aware of)

- Strict privacy settings (unlikely to cause 403 straight away), or obscure browsers (also unlikely, but can trigger firewall because of connection fingerprint/characteristics)

- Actual bad behavior (scraping, etc)

Still working for me, your IP or browser request might be flagged by their firewall or anti-bot protection
Okay, I've figured it out, thanks to the "test distros online" thingy to test version 145 and 148 as I previously mentioned.[0]

From Firefox 148 release notes for developers:[1]

> The overflow, overflow-x, and overflow-y CSS properties can now be used on replaced elements such as images, in the same way that they are used with other elements. Prior to this, replaced elements were always clipped to their bounding container. (Firefox bug 1999100).

[0]: https://news.ycombinator.com/item?id=49272986 [1]: https://developer.mozilla.org/en-US/docs/Mozilla/Firefox/Rel...

Works for me on FF stable 148.0, 153.0.4, but does not with FF 145.
Found this open reviews platform while reading the 2026-07-16 CoMaps update change log,[1] which at a glance is a nice addition to CoMaps (still WIP I think).

[1]: https://codeberg.org/comaps/comaps/releases/tag/v2026.07.16-...

Feature request tracking: https://codeberg.org/comaps/comaps/issues/556

(relevant context in another submission: https://news.ycombinator.com/item?id=48859976)
Re:Proton Login Issues, anyone else impacted? on Friday July 10, 2026 @09:59AM (#48860039)
Yeah I also wanted to check out my inbox but it went down just as I wanted to sign in

Older »