I flat-out refuse to do business with any that requires I use an app. I won't even scan a QR code for a restaurant menu; that's my cue to go eat elsewhere.
Those truck nuts are clearly aftermarket / gender-affirming care. Someone should point that out to the truck's owner.
Yep. That user is formerly:
And probably more. I was only paying attention since Dot, but I recognized the pattern retroactively for the two "101" accounts.
They post a LOT of stuff back to back, mostly "offbrand" news and blogs pretending to be news. They also start slipping in the propaganda news as well (as you pointed out). When they get called out, they delete their account with content removal set to true.
Now that you've called them out, expect them to nuke their account and all content soon. That's their standard play. They'll be back in a day or two with a brand new account on a different instance and starting the pattern all over.
FWIW, I locally banned their current alt (Cat) as soon as it popped up after they deleted their 000 one and started back with the same pattern. They leave a lot of abandoned comments in their wake which turns into database clutter and inaccessible conversations.
Lol, a sequel is a bit optimistic, innit?
If only there were some warning signs they could have noticed along the way to the edge of the cliff....
If one server is misbehaving, then they get defederated.
If the instance the spam is originating from is nothing but spam, yeah. Most instances only defederate from another as a last resort and/or if the offending instance is a total lost cause or dedicated to spam/trolling/etc.
Is there anything about Lemmy's architecture that will prevent this problem?
Yes. Applications for new registrations assuming admins can be arsed to turn them on. It won't 100% prevent it, but it will reduce it by probably 90%.
Most spam on Lemmy comes from instances with open registration (ones that do not require an application). Lemdro.id is probably the biggest offender and pain in my side. Email verification and CAPTCHAs are not effective barriers. They may slow down spam signups, but do absolutely nothing to stop them.
Instances that have 24/7 admin coverage do okay with allowing open signups (again, without application approval) and keeping spam to a minimum; some still slip through, but they're usually quickly dealt with due to having an admin available 24/7. Instances with round-the-clock admin availability are rare, though.
Instances without 24/7 admin coverage (roughly 99% of them) should, IMO, NOT have open signups and require applications. Some spam may get through, but the admins can at least have eyes on new registrations.
The one where Lisa keeps calling the Corey hotline? She's jonesing while Maggie is playing with her toy phone.
Fucken dumb ass Dems dont know how to not corporate with a dictator
They didn't. Not even one voted to confirm Patel.
From the article:
Democrats have unanimously considered Patel’s track record in the first Trump administration, his incendiary remarks criticizing the bureau he was nominated to lead and more generally his role in the classified documents case to be disqualifying.
[Patel] wins confirmation in 51-49 Senate vote
All Dems voted against as well as two independents and two republicans.
Source: https://www.nytimes.com/interactive/2025/02/20/us/politics/patel-senate-confirmation-vote.html
Intellectual curiosity: Encouraging critical thinking is essential. Without it, individuals risk being swayed by narratives that don't hold up under scrutiny.
I'm convinced this is why "AI" is getting shoved down our collective throats so hard.
I picked up after the dog at the park...FOR THE GLORY OF THE EMPIRE.
See post edit. I've already answered that twice.
Developers: I will never ever do that, no one should ever do that, and you should be ashamed for guiding people to. I get that you want to make things easy for end users, but at least exercise some bare minimum common sense.
The worst part is that bun is just a single binary, so the install script is bloody pointless.
Bonus mildly infuriating is the mere existence of the .sh TLD.
Edit b/c I'm not going to answer the same goddamned questions 100 times from people who blindly copy/paste the question from StackOverflow into their code/terminal:
WhY iS ThaT woRSe thAn jUst DoWnlOADing a BinAary???
The point is that it is bad practice to just pipe a script to be directly executed in your shell. Developers should not normalize that bad practice.
Power company recently outsourced their payment system and now have to pay a fuckin' fee to pay my goddamned bill. The only way to avoid that is autopay.
Further infuriating is I have to re-add my bank info to yet another third party system.
Fuck the modern world, man.
TNG 5x24: The Next Phase
Don't get me wrong, I love Ro Laren as a character, and I am presenting this scene out of context for humor. She did start with some less busy medical staff before bee-lining to the manager. She also never complained of a headache, but Geordi did, and they went through the same transporter accident, so I'm making an assumption here.
Cross-posted from "1.4.30 Released" by @ptz@dubvee.org in !tesseract@dubvee.org
This release incorporates several feature requests, lots of bugfixes, and the new ability to open posts and comment threads in modals. The changelog doesn't do the number of changes justice, and I've only highlighted the noticeable aspects.
Thank you to @db0@lemmy.dbzer0.com, @Kalcifer@sh.itjust.works, and @wise_pancake@lemmy.ca for submitting bugs I was able to fix. Additional thank you to @lucki@feddit.org for running the beta versions and reporting all new bugs.
ghcr.io/asimons04/tesseract:1.4.30ghcr.io/asimons04/tesseract:v1.4.30ghcr.io/asimons04/tesseract:latest"Moderator View" listing type was broken
Better truncation of modal titles
Tightened regex pattern to filter out links that are similar to user/community links.
The comment button on posts in the feed did not respect the "Open posts in new tab" setting.
Flairs should now detect if there are spaces before or after them in the post title. e.g. Check this out [Video] wouldn't have rendered [Video] as a flair since there was a space after.
Works better with Lemmy's stupid federated image proxy URLs; updated image/video/audio detection functions to account for that stupidity
Tesseract's (less stupidly implemented) image proxy can now handle more weird CDN formats if it has to un-proxy a thumbnail URL from Lemmy's stupid federated proxy URL
Custom emoji's weren't respecting aspect ratio
Nested list items were broken; added a regex to detect those and not trim those lines during pre-processing
If no spoiler title was provided, the default fallback "Spoiler" wasn't being applied
Re-ordered post type detection scripts so posts with embed_video_url aren't erroneously rendered as image posts (e.g. Imgur...somtimes)
Field for alt text shows up in post form if API is 0.19.4 or higher (was previously set for 0.19.5 as the minimum)
Don't attempt to mark dummy/preview post as read
Previewing content in modals is now properly contained when maximized (videos, embeds, etc)
Iframe link previews (previewing a link that allows access via iFrame) are now screen height (well, 80% to account for the modal wrapper/button bar).
On 0.19.4 or higher, you can now upload a custom thumbnail on posts
Post embed descriptions tweaked a bit
Badges that are clickable now have visual indicators on hover
"Old" and "Controversial" comment sort options now available
Too many minor tweaks to name individually (consequence of re-writing the post renderers)
You will now be prompted before navigating away if you have post or post edit in progress.
Moved listing type and sort selectors out of sub-navbar and into feed component
When expanding a post body in the feed, it only expands to a maximum of 50% of the viewport height and scrolls. Prevents opening a huge wall of text which requires a lot of scrolling in the feed to collapse again. This can be disabled in Quick Settings -> Scroll Post Body in Feed or Settings -> Feed -> Scroll Post Body in Feed
Scrollable area in the feed now includes the margins
Got rid of the feed margin container and just limit the width of the posts directly; width is toggleable with the same "Expand Margins" button and emulates the old behavior. Posts are slightly narrower now, but they're more consistent when resizing the window and less likely to need to expand the margins in odd, small width displays.
New view option: Wide Card (Card View + No Margins)
Moved the listing type and sort direction dropdown menus out of the sub-navbar and into the feed.
It's much more memory/network-efficient, most people don't seem to have embeds always enabled anyway, it doesn't work with Invidious/Piped, and some media has to be click-to-play (Loops, Dailymotion) for various reasons/limitations.
Making all media click-to-play has also greatly simplified the render logic.
The non click-to-play logic has been removed as well as the settings for handling those options.
This has been something I've wanted for a long time now, and it's finally here.
When a modal is active, you can now close it by navigating back: clicking the browser's back button, using the mobile OS 'back' button, gesture-swipe back, back key on mouse, etc. Anything that tells the browser to "go back" will close the modal. This also includes the zoomable images.
Links to Tidal albums, tracks, and playlists should now embed as interactive playlists. As with other embeddable media, you don't need to use any kind of special share link; just the link from the browser tab.
When clicking a Tidal link in the comments (or choosing 'Preview' from the post action menu on a Tidal post), the link preview modal will also show the album or playlist as an embed.
By default, posts open to the post page same as they always have. In addition to optionally opening them in a new tab, you can now load them in a modal.
The setting is in Quick Settings -> Open Posts in Modal or Settings -> Feed -> Open Posts in Modals
This is nice if you want to open posts and read/respond in the comments without leaving the feed.
Additionally, on comment items in the inbox and user profiles, there is a button to jump to the comment thread in a modal. Very useful for getting context without leaving your current spot.
Report items also have this ability in order to easily get context before making a mod decision on an item. It will even bring up the whole comment thread in the modal if the reported item is a comment.
When viewing a post in a modal, clicking the title will take you to its /post/ page (even if it's a remote post)
The modal does not automatically resolve foreign post/comment links to your home instance. It first loads it remotely, and there is a button to load it on your home instance. This could be automatic, however:
If you click into another post from within the modal (e.g. cilcking a crosspost item or another linked post), it will keep a history and back/forward arrows will appear in the top-right of the modal title bar. Use these like you would a browser's back/forward buttons to return to previous entries. Note, though, that these are not bound to the browser's history, so hitting "back" in the browser will close the modal.
Even with the "Open posts in modals" option disabled, cilcking the badge-ified post/comment links will open those in a modal. Useful for referencing what was linked without leaving your current position.
The post/comment badge buttons are also regular links. Right-clicking and choosing "Copy Link" or "Open in New Tab" work as you would expect. Middle-clicking also will open them in a new tab.
A bug was submitted that when posting a link that resolves to a Tesseract resource (e.g. https://tesseract.dubvee.org/post/lemmy.world/123456), the metadata would be the generic Tesseract info rather than the metadata for the content. I had been content to leave it at that (Photon and Alexandrite both behave the same way), but I figured I'd give it one more go.
Tesseract is fully client-side rendered, and I'm not a huge fan of SSR in general. That said, sometimes SSR is useful (like for providing metadata to non-browsers), but every "correct" way I've tried to implement partial SSR has met with failure.
What ended up working, and working quite well, is less SSR and more heavy use of server-side hooks to redirect non-browser user agents to an internal API route that returns a bit-banged, minimal HTML document with the meta tags populated for post, comment, user, community, and site details.
Which is fun because it, like Tesseract's other internal API endpoints, runs on top of my home made Express-like router framework that runs inside the SvelteKit server hooks.
It's not pretty (though it is elegant), but it works. At some point, I'm going to be basically re-writing the whole application in either Svelte 5 or React (haven't decided yet), and I may explore a more SSR-oriented design at that phase (or not).
Metadata is Generated for the Following: I have metadata generating for:
/post/[instance]/[post_id]/post/[post_id]/comment/[comment_id]/u/[username]/u/[username@instance]/c/[community_name]/c/[community_name@instance]/ (Metadata for the default instance)TL;DR: I've implement this client-side.
Added support "universal" links as well as badge-ifying links to posts and comments.
@<user>@instance.xyz: Has been implemented since at least 1.4.0 (forget when)!<community>@instance.xyz: Has been implemented since at least 1.4.0 (forget when)#<post_id>@instance.xyz~<comment_id>@instance.xyzLemmyVerse links will now be localized without having to hairpin to/from Lemmyverse. After being localized, they will be processed as if they were a regular-style link.
e.g. A LemmyVerse link will automatically/transparently turn into a user, community, post, or comment badge button and have the same abilities as first-party links.
Links to posts and comments (e.g. cross-posted from https://instance.xyz/post/12345) are detected, localized, and badgeified the same way.
Additionally, like users and communities, posts and comment badge links will now open in a modal for quick reference without leaving your current spot in the feed or another post. This also includes https://instance.xyz/post/%7BpostID%7D/%7BcommentID%7D formats which seem to just be a Lemmy-UI thing.
Clicking the post title in the modal will take you to the /post page for the item and close the modal.
The post/comment badge buttons are also regular links. Right-clicking and choosing "Copy Link" or "Open in New Tab" work as you would expect. Middle-clicking also will open them in a new tab.
Since the image proxy runs in the same NodeJS process as all of Tesseract, its image proxying can be greatly enhanced by adding another layer on top which is more suited to concurrent connections. I run Nginx, and the example is for that, but this should work with any suitable reverse proxy. I have an intentionally long cache period here; adjust according to your needs.
HTTP Config:
# Nginx proxy for Tesseract's Proxy Cache
# Adjust max_size from 200m as needed
proxy_cache_path /etc/nginx/conf.d/proxy_cache levels=1:2 keys_zone=imgcache:10m max_size=200m inactive=720h;
proxy_temp_path /etc/nginx/conf.d/proxy_cache/tmp;
proxy_cache_key "$scheme$request_method$host$request_uri";
Location Config:
location /image_proxy {
## You would probably want to put these proxy options and default headers into an
## include file since they're mostly redundant on the two locations. Shown here in
## both for clarity.
proxy_http_version 1.1;
send_timeout 5m;
proxy_read_timeout 360;
proxy_send_timeout 360;
proxy_connect_timeout 360;
proxy_max_temp_file_size 0;
# Set headers to send to backend server
proxy_set_header Host $host;
proxy_set_header X-Forwarded-Host $host;
proxy_set_header X-Forwarded-For $remote_addr;
proxy_set_header X-Forwarded-Proto $scheme;
proxy_set_header X-Forwarded-Uri $request_uri;
proxy_set_header X-Forwarded-Ssl on;
proxy_pass http://127.0.0.1:8080/image_proxy;
proxy_cache imgcache;
## Adjust proxy validity time from 720 hours accordingly
proxy_cache_valid 200 720h;
add_header X-Proxy-Cache $upstream_cache_status;
}
TNG 5x09: A Matter of Time.
Guest starring "Not Stephnie Weir" as "Scientist"
As an old, grizzled sysadmin who still deals with physical infrastructure and legacy systems, I can totally relate to this. I also like to think that all of O'Brien's hacks that save the day in DS9 were made possible by Ro Laren showing him the way.
Can't answer that, but it's possible they're simply not aware of their alt history. There's a similarly active user, Microwave, but they're legit; they always post quality stuff with no agenda. Perhaps they're giving "Cat" the benefit of doubt? They also give several other controversial posters a lot of leeway (won't mention them, but you probably already know the handful I'm referring to). Not suggesting any kind of agenda with the LW mods; just seems like they're trying a bit too hard to be "fair and balanced".
This user nukes their accounts, so unless one is familiar with the patterns of their previous accounts, it's difficult to correlate them with no history available. If they're on your instance and you have concerns about them (and they haven't self-destructed yet), you might be ahead to just ban them ahead of time (w/o content removal) so there's record for later comparison.