Live data from Hacker News

Pastejacking

github.com

181–190 of 246 posts

Re: Pastejacking

#181
post #173
post #37

Earlier quoted context omitted.

Totally agree. On Firefox these are often modal to the whole browser so you can't even force close the tab.

What absolutely annoys me about this feature is that there are shitty ads which throw this dialog in a loop, making it virtually impossible to close them.

In which browser? Modern implementations shouldn't even make that possible.

Re: Pastejacking

#182

Earlier quoted context omitted.

I do too. But sometimes there's no text in source, just scripts :(

In Firefox use CTRL-SHIFT-I or use Firebug to inspect the code on the page after it's generated. I guess many times it's a transparent div covering the text, not javascript tricks like this, but I might be wrong. You can remove the div using Firebug. Selecting the text in Firebug will work, but may be a lot more work. How about printing to PDF, then selecting the text there?

remembers the times when computers were supposed to make things easier for us

Re: Pastejacking

#184

Earlier quoted context omitted.

Or to block the loading of the offending javascript!

What's stopping websites from minifying all their javascript into one file and preventing the site from functioning without JS?

Then you start writing GreaseMonkey scripts and patching the APIs they use.

Re: Pastejacking

#185
post #166

Earlier quoted context omitted.

I just turned this preference off in Firefox, and it didn't stop the demo from working, which makes sense since the preference says it only disables oncopy/cut/paste events, and this demo uses a different method.

This is because the Firefox preference is useless - it only disables the clipboard events, but not the clipboard access from any other event. So the demo simply hooks the keydown event instead. Have a look at the source, it's really quite straightforward. What browsers really should have are a standard "Ask & Whitelist" dialog for all of these security critical features[1]. It seems Firefox even used to have this fea…

I found one of the bugs and... wow:

https://bugzilla.mozilla.org/show_bug.cgi?id=38966

They had a pretty useful per-site configuration mechanism that wasn't UI-configurable so someone started to make a UI for it, but then some higher-up decided they should remove the whole thing completely! The screenshots they have there look so awesome:

https://bug38966.bmoattachments.org/attachment.cgi?id=63187

It's ironic that, meanwhile, IE gets this right.

Re: Pastejacking

#186

Earlier quoted context omitted.

> if you've got a concept of morality based on primarily being motivated to help others Ethically, self-interest is fine. Failing to help others isn't a huge deal either. It's the part where you're actively and deliberately harming others (by polluting search results - impairing both search engines, as a company, and their users) that we're objecting to.

I think the idea of "polluting search results" requires too much of a subjective concept of what good content is to say whether producing one kind or not is unethical. They aren't producing what they are because people aren't looking at it and I'm very sceptical of any notion that people don't generally do what they want. It has a very centrally planned feel to say "people want think pieces" or "people want to read a…

The techniques used to deliver it are a measure of its value.

Re: Pastejacking

#188
post #141

Earlier quoted context omitted.

Yes, it's called "advertisement". Sad welcome to 2016.

Very sad welcome because somehow working in advertisement industry is considered a respectable occupation in 2016. The cognitive dissonance of the society is sometimes mind-boggling.

"You know, you smell, and look ugly, and certainly are wearing yesteryears fashions. Buy our shit and you can be almost a human again!"

(hey it works in the advertising industry. I attack your humanity on multiple levels and then show you garbage that... kind of, almost, but not really restores your humanity. Of course, you really need next months update, or you're just subhuman again! )

Re: Pastejacking

#189
> It should also be noted, for some time similar attacks have been possible via html/css [1]

As it happens, this particular attack doesn't work in gngr [0]. The example uses an absolute positioned div to put extra text out of viewport, which is not picked up by gngr when selecting text.

gngr also doesn't enable Javascript by default, so attacks such as that described in OP are not possible from random site visits. (I recommend uBlock / uMatrix for other browsers).

However, the attack surface is really quite large here. CSS directives such as `opacity: 0.001` could be easily used to mask extra text.

  [0]: https://gngr.info/
       and https://github.com/UprootLabs/gngr
  [1]: https://thejh.net/misc/website-terminal-copy-paste

Re: Pastejacking

#190
post #13

Why do browsers not require explicit user permission before allowing a site to perform clipboard manipulations? In a similar tashion to how the HTML5 geo-location API is opt-in?

On a somewhat related note: why do browsers allow websites to prevent you from leaving via those annoying dialog boxes that ask you to click "cancel" or "leave"?

Funnily enough that API is already highly restricted: You can only show that single dialog box, you cannot modify the actions of the cancel and leave buttons and you cannot perform any asynchronous operations within the callback. (Technically you can, but the browser won't wait for them to finish)

All this was introduced to prevent abuse. Apparently it still wasn't enough though...

Post reply on HN