“spoke out against” is an understatement:

• https://en.wikipedia.org/wiki/Metallica_v._Napster,_Inc
• https://en.wikipedia.org/wiki/Dr._Dre#Copyright_lawsuits

…besides contributing to the eventual shut down of napster, Metallica first compelled it to ban hundreds of thousands of their fans - who’s usernames they delivered to them. Lars Ulrich responded to the backlash by saying “If you’d stop being a Metallica fan because I won’t give you my music for free, then fuck you. I don’t want you to be a Metallica fan.”

hmm, i see my first edit to this comment got federated to piefed but subsequent edits did not; the version after subsequent edits can be seen here on the lemmy instance i’m posting from

oops, i read my test closer and realized it will never pass as-is due to being cribbed from another test that was using in instead of assertEqual (and the expected string not containing the closing tags) but I trust it conveys what i mean to

compare the rendering of the test output (which i also wrapped in backticks to tell the markdown rendering to render it as code) in my other comment on lemmy vs on piefed.

Essentially, it is not possible to reliably identify the URLs in code (in a language that is not known to either PyFedi or its markdown parser), because they sometimes are adjacent to otherwise-URL-valid characters which actually terminate the URL in whatever syntax is being used inside the code block. So, even though it is sometimes harmless to auto-linkify inside a code block, it is also often wrong and therefore should not be (and generally is not) done.

But also, code formatting should be available as a way to disable auto-linkification when a post or comment author wants to (as I did in my original comment).

URL perfectly clickable from PieFed

PieFed W

that is not a W… i surrounded the URL with backticks, which means it should be treated as code and not made into a link.

cc: PyFedi developer @rimu@piefed.social

I had a glance at the source and don’t see a very quick fix, but I think the solution involves getting rid of the auto-linkification here and instead have that be done by the markdown library. This issue indicates that the markdown library PyFedi is using is capable of doing that.

diff --git a/tests/test_markdown_to_html.py b/tests/test_markdown_to_html.py
index 329b19be..108276c5 100644
--- a/tests/test_markdown_to_html.py
+++ b/tests/test_markdown_to_html.py
@@ -37,6 +37,12 @@ class TestMarkdownToHtml(unittest.TestCase):
         result = markdown_to_html(markdown)
         self.assertTrue("<pre><code>code block" in result)
 
+    def test_code_block_link(self):
+        """Test code blocks formatting containing a link"""
+        markdown = "```\ncode block with link: https://example.com/ \n```"
+        result = markdown_to_html(markdown)
+        self.assertEqual("<pre><code>code block with link: https://example.com/ ", result)

====================================================== FAILURES =======================================================
_______________________________________ TestMarkdownToHtml.test_code_block_link _______________________________________

self = <tests.test_markdown_to_html.TestMarkdownToHtml testMethod=test_code_block_link>

    def test_code_block_link(self):
        """Test code blocks formatting containing a link"""
        markdown = "```\ncode block with link: https://example.com/ \n```"
        result = markdown_to_html(markdown)
>       self.assertEqual("<pre><code>code block with link: https://example.com/ ", result)
E       AssertionError: '<pre[24 chars]ink: https://example.com/ ' != '<pre[24 chars]ink: <a href="https://example.com/" rel="nofoll[61 chars]e>\n'
E       - <pre><code>code block with link: https://example.com/ 
E       + <pre><code>code block with link: <a href="https://example.com/" rel="nofollow ugc" target="_blank">https://example.com/</a> 
E       + </code></pre>
E       +

tests/test_markdown_to_html.py:44: AssertionError

HTH, and thanks for writing free software!

For some reason this post caused me to search to see if there are types of birdseed which are not suitable for ducks. From a few minutes of reading, I’m relatively sure now that at least all common types of birdseed are fine for them. But, along the way, I found this LLM slop https://plantnative.org/can-ducks-eat-bird-seed.htm (URL rendered unclickable since it is spam) which strongly implies that ducks are not in fact birds:

• “Yes, ducks do eat bird seed. Although bird seeds are specifically prepared to meet the nutrient demand of birds. But still, bird seed can be served as a snack to ducks in limited amounts.”
• “Bird seeds are loaded with nutrients and can be beneficial for ducks the way they are for birds.”

some more history:

April 2023: Substack CEO Chris Best Doesn’t Realize He’s Just Become The Nazi Bar

December 2023: Substack Turns On Its ‘Nazis Welcome!’ Sign

For chat, something with e2ee and without phone numbers or centralized metadata. SimpleX, Matrix, XMPP, etc - each have their own problems, but at least they aren’t centralizing everyone’s metadata with a CIA contractor like Jeff Bezos like Signal is.

For email, I’d recommend finding small-to-medium-sized operators who seem both honest and competent. Anyone offering snakeoil privacy features such as browser-based e2ee is failing in at least one of those two categories.

No, it isn’t about hiding your identity from the people you send messages to - it’s about the server (and anyone with access to it) knowing who communicates with who, and when.

Michael Hayden (former director of both the NSA and CIA) famously acknowledged that they literally “kill people based on metadata”; from Snowden disclosures we know that they share this type of data with even 3rd-tier partner countries when it is politically beneficial.

Signal has long claimed that they don’t record such metadata, but, since they outsource the keeping of their promises to Amazon, they decided they needed to make a stronger claim so they now claim that they can’t record it because the sender is encrypted (so only the recipient knows who sent it). But, since they must know your IP anyway, from which you need to authenticate to receive messages, this is clearly security theater: Amazon (and any intelligence agency who can compel them, or compel an employee of theirs) can still trivially infer this metadata.

This would be less damaging if it was easy to have multiple Signal identities, but due to their insistence on requiring a phone number (which you no longer need to share with your contacts but must still share with the Amazon-hosted Signal server) most people have only one account which is strongly linked to many other facets of their online life.

Though few things make any attempt to protect metadata, anything without the phone number requirement is better than Signal. And Signal’s dishonest incoherent-threat-model-having “sealed sender” is a gigantic red flag.

more important than expecting ip obfuscation or sealed sender from signal

People are only expecting metadata protection (which is what “sealed sender”, a term Signal themselves created, purports to do) because Signal dishonestly says they are providing it. The fact that they implemented this feature in their protocol is one of the reasons they should be distrusted.

it looks like it’s just some new in-browser advertising for Mozilla (Mullvad) VPN :/

one of the bugs linked from the bug that you linked to says this:

Feature callout that introduces the new VPN integration. Has two variants depending whether the user is signed in or signed out, prompting the user to either turn on the VPN or sign in for access. Targeting will need to reach users who have been selected for the Nimbus experiment rolling out VPN integration but do not have it turned on.

imo you should not promote signal, proton, or tuta for various reasons including the ones i mention in those three links.

Show proof that azov are nazi

There are several ways to; afaik waypipe is the best one.

machine translation of a paragraph of the original article:

The police’s solution: It’s none other than a Trojan. Unable to break the encryption, they infect the traffickers’ phones with malware, subject to judicial authorization. This way, they gain full access to the device: apps, images, documents, and conversations. Obviously, GrapheneOS isn’t capable of protecting itself (like any Android) against this malware.

🤔

Any ideas where to post this

you can find appropriate lemmy communities for posts like this here or here or here.

i am deleting this post from asklemmy as it is offtopic here.

if you didn’t save your text elsewhere and want it back, let me know and i can paste it to you in a PM.

On the Effectiveness of Aluminium Foil Helmets: An Empirical Study

!eleven@lemmy.ml

(it’s either something like this, or OP elided the word “audience”)

not only a racist but also a misogynist, a present-day gamergater, who is literally in 2025 still complaining that “video games used to be made for guys and now they’re made for everyone” 🤦