Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
remram
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
41 ms
·
751.
▲
by
remram
2y ago
How would that work? How do I get a patent troll to protect my patent, when a big company starts infringing it?
752.
▲
by
remram
2y ago
I agree with you in general but I really don't see it here. This app looks exactly like an email client.
753.
▲
by
remram
2y ago
MTA-STS is still vulnerable to MITM. If someone can tamper with DNS queries, it's useless. It's also apparently useless for the first message because the policy will not be in cache and only fetched afterwards. We could have added
754.
▲
by
remram
2y ago
That's fair. My issue with the requirement is that, if you fire up enough tasks, you only ever get to print a result or two before being blocked until the whole set is done anyway. But if you really want that behavior, my solution does
755.
▲
by
remram
2y ago
Yeah with enough scripting, you can rebuild a slow equivalent to a real backup program, that will also use 10x the disk space.
756.
▲
by
remram
2y ago
That's what WaitGroup is for. You can only get the full sorted array when every goroutine is done. You don't need to spin...
757.
▲
by
remram
2y ago
Usual problem, if you delete/corrupt a file and find out two days later, your daily backup is not going to help you. Having more than one snapshot is very valuable. http://www.taobackup.com/ etc Rsync is also very slow
758.
▲
by
remram
2y ago
The pipfile repo claims the opposite: > This repository contains the design specification of the Pipfile format, as well as a proposed implementation of a parser for the specification which can be used by Pipenv and, in the future, any o
759.
▲
by
remram
2y ago
This is great!
760.
▲
by
remram
2y ago
When I said "the author" I was not trying to mean myself. Agreed that in a perfect world I would never need this, but...
761.
▲
by
remram
2y ago
Pipfiles is what we are talking about.
762.
▲
by
remram
2y ago
Why use an array of channels when you can just have the goroutines write to a regular array?
763.
▲
by
remram
2y ago
I had forgotten about pipfiles. Like their name implies (/s) they are not for pip but pipenv, a separate tool. https://github.com/pypa/pipfile "This format is still under active design and development",
764.
▲
by
remram
2y ago
Can I have it find the requirements as of a specific date ? If I find a script/notebook/project with no requirements.txt, I usually know when it was created. Being able to get the versions that were selected back then on the auth
765.
▲
by
remram
2y ago
Isn't that true of anything a web browser can do?
766.
▲
by
remram
2y ago
Tasks that never get done, yes. In other words, tasks that wait forever.
767.
▲
by
remram
2y ago
That's right, this is true no matter the queue length. If input=output on average, there is no limit on how long your queue will grow, and therefore no limit on how long queued task will wait. I don't know what you mean by reverse
768.
▲
by
remram
2y ago
A task "shed" is one delivered with infinite latency. If that's fine for you then the theorem doesn't hurt you, do what's best for your domain. It's just something to be aware of.
769.
▲
by
remram
2y ago
I have the opposite complaint: you can't even reliably select text without also using the keyboard! Half the time I drag to select, it starts moving a block. Half the time I click to navigate, it starts a selection instead (as if I had
770.
▲
by
remram
2y ago
Why?
771.
▲
by
remram
2y ago
I find DataView to be clunky as well. Data entry is not easy (I'm currently using "DB Folder" and it helps but it's quirky) and each-row-is-a-file does not give great performance.
772.
▲
by
remram
2y ago
The fact that it has better write speed than read speed doesn't mean that it has bad read speed. It just happens to have even better write speed. It's like how I connect my phone to my home's cable connection to send a big fi
773.
▲
by
remram
2y ago
Those are great points! Another related law is from queuing theory: waiting time goes to infinity when utilization approaches 100%. You need your processes/machines/engineers to have some slack otherwise some tasks will wait forev
774.
▲
by
remram
2y ago
No trusted company behind it, not even a name or address. "About us" link leads nowhere. Is this an obvious scam to steal your servers?
775.
▲
by
remram
2y ago
Sounds like you might have wanted a poll: https://news.ycombinator.com/newpoll
776.
▲
by
remram
2y ago
In what circumstance do you imagine NYC tax money would go towards EU lawyers?
777.
▲
by
remram
2y ago
You are right of course, but the default policy can be changed using "ufw default allow incoming", making "deny" necessary. You might do it on a laptop/workstation (in a network firewalled from the internet), though
778.
▲
by
remram
2y ago
You're free to restart anything you want, of course. Reboot the whole machine every time you change anything, that way you'll be sure. I am not stopping you, and in some cases it can be useful to make absolutely certain. I only po
779.
▲
by
remram
2y ago
> The global scheduler is built on Cloudflare Workers, Durable Objects, and KV, and decides which Cloudflare location to schedule the container to run in. Each location then runs its own scheduler, which decides which metals within that
780.
▲
by
remram
2y ago
Please post the recording when it's up.
More ›