Yeah. It was already happening circa USB3. It’s not because of the connectors, but the broadening spectrum of requirements of client devices.
Maybe USB-C was a missed opportunity to address it, but it certainly didn’t “start the fire”.
If you’re not deliberately min-maxing the CAP Theorem or doing EDA, there’s no reason to use microservices and every reason not to.
It is not just an implementation detail or a matter of preference. There are fundamental UX implications.
That can be a net positive for users (and developers). But if you’re doing it “just cuz”, you’re gonna have a bad time.
You didn’t have to post this
I often want to know the status code of a curl request, but I don’t want that extra information to mess with the response body that it prints to stdout.
What to do?
Render an image instead, of course!
curlcat takes the same params as curl, but it uses iTerm2’s imgcat tool to draw an “HTTP Cat” of the status code.
It even sends the image to stderr instead of stdout, so you can still pipe curlcat to jq or something.
#!/usr/bin/env zsh
stdoutfile=$( mktemp )
curl -sw "\n%{http_code}" $@ > $stdoutfile
exitcode=$?
if [[ $exitcode == 0 ]]; then
statuscode=$( cat $stdoutfile | tail -1 )
if [[ ! -f $HOME/.httpcat$statuscode ]]; then
curl -so $HOME/.httpcat$statuscode https://http.cat/$statuscode
fi
imgcat $HOME/.httpcat$statuscode 1>&2
fi
cat $stdoutfile | ghead -n -1
exit $exitcode
Note: This is macOS-specific, as written, but as long as your terminal supports images, you should be able to adapt it just fine.
All of security is about trade-offs. “What does it protect me from, and what do I give up to gain that protection?”
If you need to remember a lot of passwords, then having some kind of system makes sense.
But most people don’t need to remember a lot of passwords. Most people can reasonably offload that job to a password manager.
So without knowing anything more, I’d guess it’s not good security for them.
If it’s shit, that’s bad.
If it’s the shit, that’s good.
Depends on the definition of “you”
The process is supposed to be sustainable. That doesn’t mean you can take one activity and do it to the exclusion of all others and have that be sustainable.
Edit:
Also, regretably, I’m using the now-common framing where “agile” === Scrum.
If we wanna get pure about it, the manifesto doesn’t say anything about sprints. (And also, you don’t do agile… you do a process which is agile. It’s a set of criteria to measure a process against, not a process itself.)
And reasonable people can definitely assert that Scrum does not meet all the criteria in the agile manifesto — at least, as Scrum is usually practiced.
It’s funny (or depressing), because the original concept of agile is very well aligned with an open source/inner source philosophy.
The whole premise of a sprint is supposed to be that you move quickly and with purpose for a short period of time, and then you stop and refactor and work on your tools or whatever other “non value-add” stuff tends to be neglected by conventional deliverable-focused processes.
The term “sprint” is supposed to make it clear that it’s not a sustainable 100%-of-the-time every-single-day pace. It’s one mode of many.
Buuuut that’s not how it turned out, is it?
Crispy can also mean you’re contemplating the nature of Mario’s 8-bit existence
Recently had a password that was acceptable for the account creation page on the website but too long for the login screen in the mobile app.
Took me a while to figure out that pasting into that field was just quietly dropping characters.
I’m out here trying to answer reasonable questions techie folks might have about the most promising possibility I’ve seen so far for getting our normie families off of Google.
What are you here for? Calling people naive pseudo-scammers? Get out of here.
It’s the #1 thing that drives me crazy about Linux.
It seems obvious. You’ve got a Windows/Apple/Super key and a Control key. So you’d think Control would be for control characters and Windows/Apple/Super would be for application things.
I can understand Windows fucking this up, cuz the terminal experience is such a low priority. But Linux?
There’s some projects like Kinto and Toshy which try to fix it, but neither work on NixOS quite yet.
There’s a link in the second paragraph to the technical details, including source code for the implementation and documentation for the required infrastructure.
But the tl;dr is that the tokens aren’t associated to your account. Unless you were able to snoop on the original request that generated the tokens (in which case, you’ve got bigger issues!), there’s no way to prove that a token is related to a specific account. A token only proves that an authorization server once granted access to some account.
Edit: Wikipedia has a good intro:
Non-interactive zero-knowledge proofs are cryptographic primitives, where information between a prover and a verifier can be authenticated by the prover, without revealing any of the specific information beyond the validity of the statement itself.
Edit 2: You should not be catching downvotes. You had a reasonable question.
Fine for you. I’m just glad there’s an option besides “sell your soul” and “invest hundreds of hours and dollars into self-hosting”.
I’m not against AI as a technology in principle, I’m no luddite.
Perhaps not a luddite, but a Luddite.
The actual followers of Ned Ludd weren’t opposed to technology. They were, in many cases, experts in the machinery — sometimes having built the machines they would later destroy.
They opposed the new social order that seemed to inevitably arrive with the machinery. The capitalists would make more money than before, the workers less, and also endure more dangerous working conditions.
Btw, your note about absorbing and repackaging counter-culture reminded me of Rebel Sell by Andrew Potter. There’s a good episode of You Are Not So Smart about it: https://youarenotsosmart.com/2012/10/08/yanss-podcast-episode-five
Beginner in IT:
“The problem is that there isn’t one”
Expert:
“The problem is that there isn’t one”