Mapa do Software Livre – Saiba mais sobre projetos no Brasil!

https://mapeamento.softwarelivre.tec.br/

#SoftwareLivre #mapa #brasil

Todd C. Miller has been maintaining the #sudo codebase for over 30 years. This is exactly one of those cases where an entire critical infrastructure is held together by the work of a single volunteer who apparently can’t find anyone willing to sponsor him for some financial support. #opensource #linux #foss #GNU

Apparently in Italian there is a word for:

"Men of retirement age who spend their time watching construction sites, especially roadworks – stereotypically with hands clasped behind their back and offering unrequested advice to the workers."

https://en.wikipedia.org/wiki/Umarell

With normal folks increasingly changing to Linux it’s vital that we all exclusively use tiling window managers to maintain an appropriate sense of superiority

@someodd A page with only <div>s and <span>s. The complete opposite of https://maurycyz.com/misc/make-up-tags/

@someodd it needs more <div>s
A clusterfuck of nested <div>s

It's Daisy.
Her soul lives in the BFG9000.

Cartoonish bunny nose, for comparison.

I just realized there's a cyberbunny in the center of BFG's sprite; and now i cannot unsee it.
The split fluffy nose.
The big bunny teeth.
The cyber-y red eyes.

#DOOM

Sometimes I wish I were a dev so that I could actually produce something that does stuff - rather than assemble and configure someone else’s something that does stuff.

@Anachron I did had backups of my maildirs. But after re{formating,installing} things i have still not configured automated backup yet; and have been postponing it...

@radhitya That is for the login manager only. KDE as a whole has not dropped FreeBSD support (yet?).

@radhitya Hi! Yes, i am. It has been my daily driver for 6 years.

I also have Fedora installed in dual boot for gaming. But i rarely use it.

For 10 months i've been deleting emails in my mail client thinking they would go to a trash mailbox as has always been the case.

But today, when looking for a mail i have trashed a few days ago, i realized that during some dotfile rewritting, past seninha, for some reason i cannot rember, removed the line defining a trash maildir from his ~/.muttrc, and that all my trashed mail has been permanently rm-f'd since then.

Not enough attention is paid to the fact that part of #Ubuntu's monetization strategy is to withhold security updates to non-paying users

#canonical #linux #ubuntuserver #ubuntults #ubuntu #sysadmin

@laxsill re: emacs vs vi

Back in 2015, when I was in a meeting with a user or stakeholder who said “I’d like feature X to be slightly different”, and I realized that it was just a matter of changing 1-2 lines of code, I could just grab my laptop, change a Perl or JS file in a single monorepo that I could easily grep, push it, trigger a pull+reload on our uWSGI server pool through Puppet, and that’s it - the feature was usually there within 5 minutes.

When I get the same request today, I know that it involves:

• Browsing through at least 4-5 different services or repos to even realize where the feature is, and how that change could impact a chain of theoretically-loosely-but-actually-tightly-coupled services downstream

• It often requires a round of Sourcegraph or Windsurf AI just to figure out these dependencies

• If the request touches any page or service with compliance requirements, go through a formal compliance request and wait for approval

• Wait 2-8 hours before one, two or three colleagues (depending on the project’s policy) review/approve my PR

• If the change is spread across multiple repos, multiply the time above by that number of repos

• Wait 30-180 minutes for all the batteries of tests to run, for all the environments to sync, for Kubernetes to take pods out and into the pool, for Terraform to rebuild the infra, etc.

What was often a 5 minutes round-trip to production before, and could be easily done within a single meeting, now turns easily into 1-2 full days (if everything goes smooth).

Of course, I’m not saying that all the compliance, validation and orchestration layers we’ve added are bad. Most of them (but perhaps not all) are there for a reason, because we learn from past mistakes, and commit to add intermediate steps to prevent those mistakes in the future.

And I’m not saying that we should all go back to the time of monorepos running on production boxes that we could just SSH into, git pull and reload. I’ve seen a whole company IT infra fall with a simple push+pull cycle on a Bash or AWK script.

But there must have been a middle ground between the two extremes that we’ve missed, while only listening to those who believe that problems of process overhead can be only solved by adding more process overhead.

If someone had told me 12 years ago that in 2026 I would spent much more time to deploy even simple changes to a production server, that the complexity and number of moving parts running different frameworks would have been so high that I would have needed an AI assistant to navigate it, and that the machines would have been 10x more powerful but would have spent 10x the time to ship changes, I don’t think that I would have felt very positive about my future.

you dont solve problems with docker.
you create them