Most of us are Reddit refugees, and probably clicking more random links than we ever did before on websites we’ve never seen before. This whole experience feels like the old internet, but also throws up insane red flags with a modern internet perspective. What are the cybersecurity weaknesses we should all be looking for, and what are the best practices?
Here’s my reason for posting this. As I search for new communities across instances to follow, I sometimes end up clicking a link and I’m no longer logged in. In the corner, that could be a Sign In link or it could be phishing. It’s likely due to me not understanding how to properly navigate this system, but there’s nothing stopping someone from setting up a sight like this as far as I know.
Thoughts?
If you link directly to the full URL (including the instance), you’ll take anyone who clicks it to that instance, and they won’t be logged in. This is usually not what you want. Example: https://pawb.social/c/tech - This link will take you to my instance.
If you remove the instance URL, and just leave /c/communityname@instance - for example, /c/[email protected] - the link will still take you to the community, but you’ll still be on your instance. This is usually desirable.
Basically, instance -> community = link to that instance. Community -> instance = link to the community in whatever instance the user clicks it in.
You can also use ! instead of /c/ - I think this might work better for Kbin users (since they use /m/ instead of /c/ - can’t verify this). In that case, it’d be: [email protected]
I won’t get tired of posting this everywhere it applies :D
I made this userscript, which rewrites all links everywhere (not only on Lemmy) to always point to your home instance. So the link in your comment actually looks like this to me:
i.e. even though you tried to link to your instance, my script rewrote your link back to my instance so it’s working fine :D
But of course I can still hover over the icon to see how your link originally looked:
I Have to say that @[email protected] 's script is fantastic. I’ve been using for a few days now (I have another bug report for you@[email protected] )
Would be nice if third party apps implemented that functionality.
Or if there were bots that automatically identify those external links and reply to them with a link to the community/post in other popular instances.
This is great, and would make for a super useful Firefox extension.
Why would or should this be a Firefox extension when it already runs perfectly well on Firefox?
I’m curious, is there an advantage of running a script over an add-on? Like is it faster or takes less resources? Or did you just happen to code it like that? Not complaining though, it’s been working great for me so far.
The advantage is I don’t have to learn how to build an addon. It just runs code which I already can write. There’s also the advantage that any browser can run JavaScript. Idk if any browser can run Firefox (or whatever) extensions.
I guess I’ve never run scripts in Firefox.
As soon as you’ve installed your preferred user script extension like Violentmonkey it’s as simple as installing addons, you just click the “install” link on the script’s page.
There are lots of different useful ones.
So just out of curiosity, since you’re the first person I’ve seen actually point out optimal linking with the ! Symbol, I have to ask how you pronounce it. For me, ! will always be “bang”, so I’m just curious what the pronunciation is.
I use “Bang”, too, if I’m trying to verbally say it, though… that very rarely comes up. If I’m reading it, I don’t internally “pronounce” the symbol at all. If it was verbal, though, the above link would be bang tech at pawb dot social.
Thanks for the explainer! Doing some testing cause your example didn’t hyperlink on Memmy
c/[email protected] /c/[email protected] !c/[email protected] [email protected] test text /c/[email protected] /c/[email protected]
Weird. Not sure when your example didn’t link, because it did in my comment ¯\(ツ)/¯
Edit: I’m back on browser. Everything that hyperlinked works properly. It’s a Memmy issue
Okay I learned a few things, though they may be specific to Memmy.
Text testing #3 - confirmed, this returns the Null error.
Now without prior text test
I love the “show source” button which gives access to how the tests are made.