Search
1000 results for “kytta”
-
I've watched the #ChatControl session live, which was the first #EUParliament plenary session I've ever watched, and I have questions.
- If every MEP has a digital voting device, why do some votes happen by show of hands?
- For both types of votes: Why the hell is it so fast?
- Why are all proposals worded so badly, with layers upon layers of negation? "Ah yes, I vote against the amendment that would remove a paragraph from a motion to discard a paragraph from an act that forbids Foo" how do you even parse this?
Or, generally speaking: Are we sure that, even if the elected MEPs are the ones we voted for, they actually submit correct votes? Not because they're secretly evil, but because the whole voting process seems so confusing and user-hostile.
I really need to learn more about it all works
-
-
Having #Algolia widget for your docs is the worst thing you can do for your project's users. Only five results at any time, no results page so that you have to retype your search query every time, and the search is so fuzzy that it's impossible to find pages unless you type their full title.
-
Having #Algolia widget for your docs is the worst thing you can do for your project's users. Only five results at any time, no results page so that you have to retype your search query every time, and the search is so fuzzy that it's impossible to find pages unless you type their full title.
-
Having #Algolia widget for your docs is the worst thing you can do for your project's users. Only five results at any time, no results page so that you have to retype your search query every time, and the search is so fuzzy that it's impossible to find pages unless you type their full title.
-
Having #Algolia widget for your docs is the worst thing you can do for your project's users. Only five results at any time, no results page so that you have to retype your search query every time, and the search is so fuzzy that it's impossible to find pages unless you type their full title.
-
Having #Algolia widget for your docs is the worst thing you can do for your project's users. Only five results at any time, no results page so that you have to retype your search query every time, and the search is so fuzzy that it's impossible to find pages unless you type their full title.
-
Das gleiche nochmal auf Deutsch.
Es nervt mega, dass #MailboxOrg Spam-Mails nicht vernünftig filtert. If you know, you know. Ich überlege mittlerweile, meinen eigenen Spamfilter über den mailbox.org-Posteingang laufen zu lassen, aber ich weiß nicht, wie ich das am besten mache.
Meine Wünsche: Im Idealfall muss er E-Mails filtern, bevor ich sie lesen kann. Gleichzeitig sollte er auf mehreren Geräten gleich funktionieren, also am liebsten schon aufm Server greifen und nicht Client-Side.
Meine erste Idee wäre, einen einfachen Daemon zu schreiben, der entweder auf IMAP-Push-Benachrichtigungen achtet (oder alle 10 Minuten die Mails prüft) und dann die Spam-Mails in den Spam-Ordner verschiebt. Oder gibt es vielleicht eine einfachere Lösung? Wenn ihr auch Mailbox.org nutzt, wie geht ihr damit um?
-
I am sick of #MailboxOrg not filtering out spam and scam emails properly. If you know, you know. I am now playing with the idea of running my own spam filter on top of my mailbox.org inbox, but I can't figure out how.
My needs: Ideally, it has to filter emails before I get to read them. At the same time, it should be synced across devices (= happen on IMAP, not on-device).
My immediate idea was to set up a daemon that would either listen on IMAP Push notifications or check every 10 minutes, and then just move spam emails to spam. Am I missing a more obvious solution? If you are also a Mailbox.org user, how do you deal with it?
-
More #Vercel woes. Self-hosting #Share2Fedi didn't help – I guess a lot of people still use the v2 aka toot – and the usage now reached 400%, causing every website I host on Vercel to get blocked. It will be a fun train ride trying to self-host everything.
PSA to those using #Shareon share buttons or toot V2 for Mastodon sharing: You might wanna switch to the Fediverse button or Share2Fedi, respectfully.
-
Oh nein, die #Schrinkflation hat die 1 Liter Coca-Cola getötet, jetzt sind da nur 850 ml drin?
Und ich finds krass, wie sie es hinkriegen, dass die Flasche immer noch gleich groß aussieht
-
-
-
I like the DX of #Astro and #Vite, but man, I just can't survive any more "security" "vulnerabilities". Coupling any projects written in those frameworks with #Dependabot is a grave mistake. But it doesn't matter what dependency scanner you use.
#NPM taught us that vulnerabilities are everywhere, and we should use a scanner. But then, I'm getting a new "vulnerability" every other week. Yes, once a week I get an alert about yet another XSS/ReDoS. I use quotations the whole time, because the vulnerabilities can be exploited only when "you run Vite/Astro development server in production", which you should never do.
I'm developing an alert blindness because of this. If I see an alert about a new vuln in Vite/Astro, I straight up ignore it, because I can now deadass predict what it's gonna say.
-
I feel like this is a very common problem amongst #Python libraries. Most use #ReadTheDocs to host documentation, and the default template has no obvious way to include a link back to GitHub that is visible on every page. One can show the source code of a function/class in the API docs /
autodoc, but it's not the full repo. I'd prefer everyone to add an externaltoctreelink to the source code, but I haven't seen anybody doing this.As a result, when I search for a Python library on the web, the first search result is the RTD page, the second one, if I'm lucky, is the #PyPI page, but the source code is nowhere to be found, and I have to go look for it (often by clicking "Edit this page" in the docs, then manually trimming the URL) :/
-
I feel like this is a very common problem amongst #Python libraries. Most use #ReadTheDocs to host documentation, and the default template has no obvious way to include a link back to GitHub that is visible on every page. One can show the source code of a function/class in the API docs /
autodoc, but it's not the full repo. I'd prefer everyone to add an externaltoctreelink to the source code, but I haven't seen anybody doing this.As a result, when I search for a Python library on the web, the first search result is the RTD page, the second one, if I'm lucky, is the #PyPI page, but the source code is nowhere to be found, and I have to go look for it (often by clicking "Edit this page" in the docs, then manually trimming the URL) :/
-
I feel like this is a very common problem amongst #Python libraries. Most use #ReadTheDocs to host documentation, and the default template has no obvious way to include a link back to GitHub that is visible on every page. One can show the source code of a function/class in the API docs /
autodoc, but it's not the full repo. I'd prefer everyone to add an externaltoctreelink to the source code, but I haven't seen anybody doing this.As a result, when I search for a Python library on the web, the first search result is the RTD page, the second one, if I'm lucky, is the #PyPI page, but the source code is nowhere to be found, and I have to go look for it (often by clicking "Edit this page" in the docs, then manually trimming the URL) :/
-
I feel like this is a very common problem amongst #Python libraries. Most use #ReadTheDocs to host documentation, and the default template has no obvious way to include a link back to GitHub that is visible on every page. One can show the source code of a function/class in the API docs /
autodoc, but it's not the full repo. I'd prefer everyone to add an externaltoctreelink to the source code, but I haven't seen anybody doing this.As a result, when I search for a Python library on the web, the first search result is the RTD page, the second one, if I'm lucky, is the #PyPI page, but the source code is nowhere to be found, and I have to go look for it (often by clicking "Edit this page" in the docs, then manually trimming the URL) :/
-
I feel like this is a very common problem amongst #Python libraries. Most use #ReadTheDocs to host documentation, and the default template has no obvious way to include a link back to GitHub that is visible on every page. One can show the source code of a function/class in the API docs /
autodoc, but it's not the full repo. I'd prefer everyone to add an externaltoctreelink to the source code, but I haven't seen anybody doing this.As a result, when I search for a Python library on the web, the first search result is the RTD page, the second one, if I'm lucky, is the #PyPI page, but the source code is nowhere to be found, and I have to go look for it (often by clicking "Edit this page" in the docs, then manually trimming the URL) :/
-
Maybe #NixOS is something for me, after all, but for "wrong" reasons. I really don't care about reproducibility, learning what Flakes are, or ephemeral shells. All I want is one file that governs my whole system, with sane defaults for many services that I'd want to self-host, all without relying on Docker!
I've tried my luck with #Ansible and #pyinfra, but I've failed. I'm starting to see them as "Tailwind for Ops": You have to learn a non-standard language only to do stuff you could already do with a (de-facto) standard language (shell scripts). But, writing shell scripts is a hassle, too, just like applying changes on the server. And I will definitely forget where I've put what, or what technology governs which services.
The worst is probably user management. Creating system users, some of which should get subuids, other of which will run systemd units,
chmod,sudo... I'm just tired of typing the same commands over and over. Adding a line toconfiguration.nixthat will do the magic for me seems inviting. -
#Domain business is crooked.
There was a domain I really wanted to buy, and it was in deNIC's Redemption Period. I waited till the day it was supposed to become free, but it was snatched by a domain reseller; the homepage displayed some dubious links, and there was a form where I could make an offer. Looked the company up, found out it only accepts 4-digit dollar amounts and bigger, so I never bothered.
Until an update a few days ago? I wanted to check how long I'll have to wait, and I saw it was acquired by a different domain reselling company, and I can now get it from #GoDaddy for about £80? That's a lot less than I thought it would be, but it's still a lot for a .de domain (which I get for under $3 on #Porkbun). GoDaddy seems to charge a premium because the domain is "15 characters or shorter"??? Since when is the bar so high up? 😭 I mean, what did I expect from the literal Satan of domain registrars?
Now I'm at a crossroads. I really want the domain, but I will only use it for private things, and only a handful of people will ever visit the website, if at all... 80 quid is more expensive that all but one domain I own, together 😩
-
I was watching Mission: Impossible – Fallout and have noticed they've used #OpenStreetMap in one scene. I then proceeded to sit through the credits to see if they included the required attribution—they didn't :(
Wanted to fume about this online, but then I looked it up, and I'm not only the first person to notice, but there's apparently a #OSMWiki page that mentions this
-
#NowPlaying: Gareth Coker — Dashing and Bashing
The best ad for the #Ori games are their soundtracks! I definitely need to play them one day
-
#NowPlaying: Gareth Coker — Dashing and Bashing
The best ad for the #Ori games are their soundtracks! I definitely need to play them one day
-
#NowPlaying: Gareth Coker — Dashing and Bashing
The best ad for the #Ori games are their soundtracks! I definitely need to play them one day
-
#NowPlaying: Gareth Coker — Dashing and Bashing
The best ad for the #Ori games are their soundtracks! I definitely need to play them one day
-
#NowPlaying: Gareth Coker — Dashing and Bashing
The best ad for the #Ori games are their soundtracks! I definitely need to play them one day
-
https://git-repo.info/en/2020/03/agit-flow-and-git-repo/
Two takeaways:
- this is (almost) exactly what I was looking for and have been wanting to write a blog post about for some time
- Forgejo supports it!
Have just tried it out on a repo of mine — works like a charm! I feel like it's the best mix of
git-send-emailand the pull-request-based workflow. -
I was always struggling with wrapping in GNU #gettext PO files; #Django
./manage.py makemessagesdoes it one way, Poedit.app does it differently. To avoid dirty Git diffs, I used to re-wrap the strings by hand this whole time... until TIL that you can just usemsgcatwithout arguments :neocat_facepalm:msgcat -o django.po django.poTo re-wrap all PO files:
fd -epo . -x msgcat -o '{}' '{}' -
After procrastinating for way too long, I have finally set up regular backups of my laptop and achieved the 3-2-1 strategy 🙌
But the reason I was procrastinating this is because I couldn't decide on a good #restic wrapper! And, to this day, none of the options appeal to me
- #Backrest can't do stdin, --keep-within-*, and passwords from command
- #autorestic can't do stdin, and scheduled check/forget
- #resticprofile is too complicated and has a different mental model