Notices by Brian Ó (blacksam@social.gibberfish.org), page 55

This pod, as well as several others that we know of have seen a large spike in "fake" accounts which have random-looking usernames and generally do not log in more than once. On this pod, we see batches of hundreds of accounts being created by individual IP addresses. This activity originates primarily from Russia and Romania.

We're not sure what the purpose of these accounts is, but they are wasteful of pod resources and clutter up the user lists. As a preventative measure, we have decided to block new signups from the two countries noted above until further notice. We have replaced the signup screen with a page explaining the situation and referring the user to https://podupti.me to choose a different pod.

To be clear, existing users from these countries are welcome and can continue to use our pod, and federation with pods hosted in these countries will still work. This only affects new user registration.

Apologies for any inconvenience this may cause.

#gibberfish #spam #registration

#Question for the #devs:

I've identified ~2,500 user accounts on my pod that are over 30 days old, have only signed in once ever, and have no posts, comments, likes, or PMs. They're either abandoned or (more likely) spam. Will anything be harmed if I delete them from the users table? I assume it will cascade to any linked tables, yes? Anything else to watch out for?

Thanks

Have any other #podmins noticed a huge uptick in accounts that seem to have gibberish usernames?

If you want to fight fascism, move left.

https://youtube.com/watch?v=agzNANfNlTs

#innuendostudios #politics #capitalism #democracy #conservatism

Here's my 5.5 gal dry-start at nearly 7 weeks. It doesn't show the same progress as the 20 gallon, but it's using an inert gravel substrate and for the first few weeks the lighting was sub-par. I'm still very happy with the progress. I'd say it's ready for water, but I put some small sprigs of monte carlo along the front as a late addition, and it hasn't gotten a chance to take hold yet, so I'll probably leave this dry for a few more weeks.

#aquariums #plants #aquaticplants #dry-start

#diaspora #devs, I have a #question: I recently migrated from mariadb to postgres, but it looks like the tag_followings table somehow got missed. I made a dump of the table from the old db and loaded it into the new one, so now there are thousands of rows, including some for my user id, but when I go to the followed tags list in the UI, it's empty. Any ideas what's wrong?

"Bowling juice"

Gibberfish Social has undergone a major overhaul this weekend.

First, the database has been converted to from MariaDB to PostgreSQL, which has led to a noticeable speed-up in stream loading time.

Second, the single server has been split into two: One which holds the database and user-uploaded content, which uses full-disk encryption. The other runs the Diaspora, Movim, and chat servers and holds no data of its own. Both new servers are ARM64-based systems, which do not have hardware backdoors like those designed into Intel and AMD chipsets.

This is not only a little bit cheaper to run than our old server, but allows us to easily scale by adding more application servers if the need arises. "Cheaper" doesn't mean free, though, so if you enjoy the services provided by this pod, please consider making a donation to help support it. Thanks!

#gibberfish #diaspora #movim #xmpp #postgresql #arm #arm64

https://youtube.com/watch?v=OL8a1YEhk_o

#thejuicemedia #australia #climatechange #politics #bullshit

I added a second light several days ago. The growth is accelerating.

#aquariums #plants #aquaticplants #dry-start

The Register wins the internet.

https://www.theregister.co.uk/2019/03/14/thats_numberwang_google_cloud_staffer_breaks_world_record_for_pi_calculations/

Soyuz Rollout to the Launch Pad

#nasa #photo #photography #space
posted by pod_feeder_v2

Not Scared

#comics #poorlydrawnlines
posted by pod_feeder_v2

Captain Marvel was kind of... flat? It had all of the things that a Marvel move is supposed to have, but for me it just didn't pop. I'm having trouble putting my finger on it. Maybe part of the problem was that it lacked a compelling villain? IDK.

#marvel #captainmarvel #mcu #movies

Are there any modern C&C/Starcraft style games that run on linux?

#games #linux #commandandconquer #starcraft #warcraft @[GamingOnLinux](/people/9555c7c09397013448590242ac110007)

@[Vermin Supreme](/people/6cd751f11339d0be) #trump #verminsupreme #crazy

Kiva is doing a 100% match on loans to women today.

https://www.kiva.org/lend/international-womens-day

#kiva #crowdfunding #microfinancing #womensday

Week 4 update on the dry start tank. I'm considering getting a second light for the tank, because the rocks cast shadows no matter where I position the light source, and I want 100% plant coverage of the substrate.

#aquariums #fish #plants #aquaticplants #dry-start #lighting

I swapped out the in-tank power filter for an air-powered sponge filter on my 5 gallon office tank. This air pump is super tiny, but it seems to do the trick!

The little sponge filter is only $5, but I think next time I'll just DIY one out of CPVC for fun.

#aquariums #fish #filtration

Fuuuck... that python face_recognition library was able to match the image on the left to the image on the right 😲

Try it yourself (where /tmp/foo.jpg is a picture with a face):
```
#!/usr/bin/env python3
import face_recognition, io, re, requests

searx_server = 'search.gibberfish.org'
search_term = 'bernie'
url = f'https://{searx_server}/?q={search_term}&categories=images&language=en-US&format=json'
hits =[]
misses = []
with open('/tmp/foo.jpg', 'rb') as sourcefile:
original = face_recognition.load_image_file(io.BytesIO(sourcefile.read()))
original_encoding = face_recognition.face_encodings(original)[0]
r = requests.get(url)
results = r.json()
count = 0
for item in results['results']:
img_url = item['img_src']
if re.match(r'https?://', img_url):
try:
r = requests.get(img_url, stream=True, timeout=3.0)
if r.status_code == requests.codes.ok:
r.raw.decode_content = True
unknown = face_recognition.load_image_file(io.BytesIO(r.raw.read()))
unknown_encoding = face_recognition.face_encodings(unknown)[0]
results = face_recognition.compare_faces([original_encoding], unknown_encoding)
if results[0]:
hits.append(img_url)
else:
misses.append(img_url)
count = count + 1
print(count)
if count > 25:
break
except:
pass
print("Hits:")
[print(h) for h in hits]
print("Misses")
[print(m) for m in misses]
```

#python #face_recognition #facialrecognition #bernie