The new hygienist is pretty cool.

Dentist: Are you feeling numb?

Me: For the last decade or so.

I guess she meant my mouth, specifically.

Conventional commits aren't reliable for semver automation, don't result in good release notes, and carry a faint stench of agent. Write your commit summaries for people.

There should be a name for a run where you suddenly regain your former pace, and realize the struggles over the last several weeks were not the irreversible decline into old age, but just a summer that was too damn humid. It is chilly and gray and I feel alive again.

I am stepping back from http4s-1.x maintenance, including security patches. Withdrawals will continue until burnout improves.

https://github.com/http4s/http4s/discussions/7945

#http4s

There's a fork late in the morning walk path, and I let my coonhound decide whether we go left or right. Sometimes she gets confused and gets her leash wrapped around a tree. This morning she confidently plowed forward and stuck her nose in a pile of poo. This ritual helps prepare me for another day in the modern software industry.

I've been up late the last three nights triaging security advisories. I received eight more in the last hour. I don't review standard PRs anymore, and now functionality is going to start getting cut.

https://github.com/http4s/http4s/discussions/7931

The most sure way to secure something is to delete it. Increasingly, it's the only way.

Since enabling `completion-eager-update` in Emacs 31, I've been mercilessly dinged during non-strict completions. Try this:

```
(define-advice minibuffer-completion-help (:around (orig-fn &rest args) suppress-ding)
"Silence the bell during minibuffer completions that may not be strict."
(cl-letf (((symbol-function #'ding) #'ignore))
(apply orig-fn args)))
```

There's also `completion-fail-discreetly`, but that also takes out the "[No match]" help.

#Emacs

I remember `scalaz.\/`, which was in many ways nicer than `Either`. My first contribution to Cats was the short-lived `cats.data.Xor`. It didn't survive to 1.0, because having multiple representations of a foundational type was worse than standardizing on the lesser. Here we go again.

https://contributors.scala-lang.org/t/pre-sip-a-new-type-for-optionals-and-error-handling-2/7533

#Scala

A key difference between me now and me 25 years ago is that now I'm compiling Emacs at 6:30am because I'm up early instead of compiling Emacs at 6:30am because I'm up late.

If the IndieWeb is punk, writing a manifesto with Claude is Good Charlotte.

Always check the bucket for guppy fry before watering the house plants.

Whereas I will be salty forever.

Spent a hot afternoon indoors answering 100 Questions for Webmasters:

https://rossabaker.com/blog/100-questions-for-webmasters/

Assisted-by: deez nuts

I initially thought Codeberg's updated terms of service against vibe coding and crypto was a good change, but after reading several critiques, I like it even more.

#Codeberg

Getting ready to spend Friday evening packaging an old Ant build for Nix. Can you believe some people say I'm no fun?

#Nix #Java

Usually: JAR size doesn't matter.

Releasing: watch helplessly while a Sonatype employee carefully chisels each bit into the Maven Central granite.

2fa deez nuts

Talking about memory safety and remembering the joke,

> To those from elsewhere, a Yankee is an American.
> To Americans, a Yankee is a Northerner.
> To northerners, a Yankee is an Easterner.
> To easterners, a Yankee is an New Englander.
> To New Englanders, a Yankee is a Vermonter.
> And in Vermont, a Yankee is somebody who eats pie for breakfast.

... and realizing that Rustaceans are the Vermonters of software.

#Rust