This month 130 users were eligible.

Old:

New:

It's based on the message ID. Posts before roughly when I posted that can be edited without updating the imgur links.

It looks like imgur is no longer blocking the image proxy. I didn't change anything on the forum's end, so imgur must've loosened up their limits. I guess it's not necessary to convert img images into links as long as imgur keeps working. But to prevent this from happening again, you'll get an error if you try to embed imgur images in posts from now on.

I'd like to create a page which recommends services for hosting forum images, and mention this in the error message you get when you try to use imgur. Other than TalkImg, which ones are good?

I logged into the account about a week ago to keep it active. If I have to keep doing this all the time, then I'll probably just let it expire, since it hasn't been used in a long time anyway.

My best guess at this point is that it was a random failure of some networking hardware. It doesn't look like any failure I've seen before, so it's difficult to diagnose. It doesn't particularly look like a DDoS or other attack, though I can't rule it out.

There won't be any downtime if/when I change anything regarding images.

It strikes me as very unlikely that anything related to this is going to cause widespread losses anytime soon, so I don't think that an Important Announcement is necessary.


I'm not as bothered by the whole idea of an opt-in centralized recovery thing as a lot of people seem to be, but this quote in particular is pretty damning because it shows that Ledger was/is either incompetent or lying.

I edited Ledger out of my "do not keep your money in online accounts" post. I still want to recommend some hardware wallet which is fairly easy-to-use, so I left Trezor in, even if it may not be perfect.

There are a lot of broken imgur links. How do you think this should be resolved? I've come up with these possible solutions:

1: Do nothing. People will have to edit past broken images and use different image-hosting services going forward.

2: Make it so that all past imgur embedded images show up as a clickable link instead of as an embedded image, as if the poster had just posted the imgur link instead of using the [img] tag. I could also maybe make it so that if you click the link, it will JavaScript-expand into an embedded image (using the specified width etc.) without going through the image proxy.

3: Exempt imgur links from the image proxy so they all just work. But this has privacy implications, since imgur will see forum users' IPs. Also, imgur has apparently banned a lot of Tor exit nodes etc. as well, so imgur images won't work for everyone.

4: Exempt imgur links from the image proxy as above, but prevent using imgur embedded images in any future posts, to sort of phase it out.

I didn't look into this deeply, but my understanding is that it's opt-in. Do we know that people using Ledger are being put at significant additional risk just by upgrading their firmware, if they don't opt into any backup stuff?

I think that what it is is that imgur added a new very restrictive per-IP limit on image downloads. So all ip.bitcointalk.org requests are blocked, but even many Tor exit nodes are being blocked, for example.

I'm not going to attempt to play some cat-and-mouse game with imgur, so people will just have to use a different service such as TalkImg.

This month 132 users were eligible.

Old:

New:

Added, thanks!


That was my thinking originally. But because SMF automatically translates multiple spaces (like    this) into an alternating sequence of regular spaces and non-breaking spaces in the preparsing step, any intentional usage of non-breaking space characters tends to get messed up, as you mentioned. So when SMF's automatic preparsing-level usage of non-breaking spaces doesn't do what you want, this new tag will be useful.

This month 132 users were eligible.

Old:
New:

Congrats! How long did it take you? I estimated that it'd take an hour or two to reverse engineer, and honestly I didn't expect anyone to go to the trouble in the 24 hours available. I'm glad it was interesting enough to somebody that I was proven wrong here!

April fools!

Thanks to PowerGlove for creating this cool game. I was really impressed that he was able to create it from scratch in less than a month with no dependencies and in only about 21kB. Thanks also to the mods who helped test it.

Here are the "high scores", though keep in mind that there were cheaters. karmamiu performed the best on every metric, and it doesn't look to me like he cheated.

I disabled the "new replies since you started writing" error globally.  (In ordinary times, you can disable this in your settings, BTW.)

In order to reduce the number the low-value posts, I had PowerGlove design a state-of-the-art, AI-enhanced CAPTCHA which will identify all low-IQ people. You'll have to complete the CAPTCHA before posting.

You can adjust the CAPTCHA's size with the +/- buttons in the upper left and right. If you post without solving it properly at first, you can go back and edit your post to try again. The CAPTCHA gets more difficult in more active topics, and it gets more difficult the more you post. If you experience poor performance, typing "\" changes a setting which may improve it.

Right now the CAPTCHA is in beta testing, so it only shows a warning banner for likely spammers, but in 2 weeks everyone who fails the CAPTCHA will be immediately banned, so you'd better get used to it!

KFC/AMC will be required for all members. Get ready to buy chicken and movie tickets.

Thanks for doing this contest, RickDeckard! There were a lot of great submissions.

nioctib_100 had the best-looking final result IMO. I love the way the letters seem to be emerging chaotically out of the apples, but yet you can still read them in the photo.

Dickiy, puloweh555, loopes, and Helena Yu also had fantastic-looking pies.

The most interesting recipe was Betwrong's. That's a pie I'd like to try making sometime.

This month 128 users were eligible.

Old:
New:

That quote predates the introduction of flags, and is no longer generally accurate. However, negative trust ratings created before the introduction of flags are "grandfathered in", and can generate a flag-like-warning, as a special case.


In other words, this type of warning can be eliminated if the grandfathered-in negative ratings are deleted, or if new positive ratings are created to offset them, or some combination of the two.