seth's website

some old irc quotes

Wed, 18 Feb 2026 00:00:00 +0000

Came across this small collection of saved IRC quotes while clearing out old GitHub gists of mine. Figured I may as well post them here. IIRC most of these are from #proggit on FreeNode probably from 2010-2015.

<eck> QOTW, "Rob, I'm sure you're focused on recovering things at the comment.
      When the dust settles, you might want to ask your team how "PHP",
      "personal www dir" and "business critical app" were ever grouped together."
<eck> this is on a ticket i'm handling

<unknown> The strengths of Go to me seem to be its extensive standard library
          and the fact that you don't need to worry about writing elegant code
          because it's not really an option.

<ryan> I hope some day in the near future linus torvalds is on film taking a
       shit in front of the door to a canonical office pointing to it and saying
       "THIS IS WHAT YOU DID TO MY BABY!"

<thoughtpolice> FML
<thoughtpolice> i'm totally getting bitten by this linux leapsecond futex bug
<x[LGWs4x4i]uG2N0> ha ha
<x[LGWs4x4i]uG2N0> he uses linux
<x[LGWs4x4i]uG2N0> laughinggirls.jpg
<greghaynes> thoughtpolice: details prease?
<dforsyth> laughingelfman.jpg
<greghaynes> that sounds like a fun bug
<thoughtpolice> i'm sorry, i can't hear you over your freebsd sucking so bad
<thoughtpolice> and not supporting any hardware
<thoughtpolice> also did i mention sucking?

<Two9A> Why do you think JS writers go insane?
<Two9A> (Because they're writing lisp and don't know it, but that's by the by)
<hfaafb_> what?
<Two9A> What?
<hfaafb_> js === lisp?
<sixthgear> haha
<sixthgear> the === is a brilliant touch

<mux> it'd be easier to teach a duck to sing than to have any kind of meaningful
      "static typing" in php

<null__> @quote kisses
<lambdabot> reppie says: would you say kisses mark the minutes on the dial of love

<lambdabot> Erik-Naggum says: You have failed to consider the ramifications of
            the solutions and pose a problem that simply would not exist if you
            did. This taxes my patience, which is already legendary in its
            general absence.

<Vellos> Sometimes nature fucks over your double-helix a bit, and the region of
         your brain dedicated to handling all that social bullshit in realtime
         got re-purposed to analytical thinking

jorts69: [The worst thing about becming an adult is] that feeling that you spent
         your whole life tapping the world on the shoulder and when it finally
         turned around you forgot what you had to say.

mattijle: "I am an energetic self starter with a deep respect for engaging
          resources and getting deep buy-in. I'm always interested in being a
          key player and helping you gain synergistic value on your
          mission-critical project. It's important to think outside of the
          box, shift paradigms and push the envelope when bringing user-centric
          change to an organization. When I'm involved it's about about the
          value add and creating a win-win. In this  economic climate you have
          to rightsize but remain scalable and always deliver service that is
          world class"

<sirpengi> git gets easier once you get the basic idea that branches are
           homeomorphic endofunctors mapping submanifolds of a Hilbert space

<rcr> i never have a hard time with merge conflicts because my code is always
      the most correct

<saml> when you pull retina display, that means it's yolo webscale

<saml> stack solves everything. that's why mongodb is a stack

<rcr> i think we should implement a new rule here
<rcr> for all new commits, git show | grep pdb
<rcr> if it returns anything you dont get paid this week

<thoughtpolice> today, at least, if someone tries an active MITM on a popular,
                high-value site, it will almost certainly be detected by Google
                or Mozilla as it auto reports this. they then remove the CA from
                their trusted list, effectively taking that CA out back, putting
                it on its knees and letting it grovel a bit, and then shooting
                them in the head

<centrinia> Go is fail.
<shoenig> @slap centrinia 
<lambdabot> Come on, let's all slap centrinia
<hfaafb> @slap centrinia
* lambdabot pulls centrinia through the Evil Mangler
<centrinia> :(
<centrinia> @slap Go
* lambdabot would never hurt Go!
<shoenig> exactly.
<centrinia> What the fuck?

nandland go board linux toolchain

Sun, 25 Jan 2026 00:00:00 +0000

I picked up a nandland Go Board to start learning about FPGAs. Just for fun, nothing too serious. I got the “Getting Started with FPGAs” book along with the development board, for just over $100. Good deal.

The only snag is that book suggests using the iCEcube2 toolchain directly from Lattice Semiconductor. Although that official toolchain is free if you request a license for personal use, the Linux support is terrible. They only provide a 32-bit binary from 2020 and making it work on a modern Linux system is a nightmare.

making a new website

Fri, 20 Jun 2025 00:00:00 +0000

Since early January I have been working on something - a new website.

It’s a stock discussion forum called TickerFeed.

https://tickerfeed.net

The past few months have been a waterfall of changes and new beginnings. After a fulfilling chapter at HashiCorp, the news of the IBM acquisition served as a clear signal for me to pivot and embark on a new adventure. It felt like it was time to step away from the corporate world and build something meaningful.

use adguard dns

Wed, 04 Sep 2024 00:00:00 +0000

AdGuard offers DNS ad-blocking but they make it difficult to find the IP addresses.

They are

94.140.14.14
94.140.15.15

and for IPv6

2a10:50c0::ad1:ff
2a10:50c0::ad2:ff

These are listed at the bottom of https://adguard-dns.io/en/public-dns.html

vlans on mikrotik

Tue, 03 Sep 2024 00:00:00 +0000

Introduction

I purchased a MikroTik Hex-S a few years ago to handle routing for my house. It sits between my cable modem and 4 ethernet cables that bring internet to other rooms.

Homelab

In my real example I wanted to enable a VLAN on the Hex-S to isolate two pairs of the physical network interfaces (ports). In my case these are two computers I use for work. I want them to be isolated from other devices on my home network, such as the Nest thermostats, media center PC, and the XBOX. Call me paranoid but I’d bet those IoT devices like to snoop on your network.

setup tailscale on freebsd

Sun, 17 Dec 2023 00:00:00 +0000

This blog post describes how to install and setup the tailscale agent on a FreeBSD machine. Luckily a tailscale port is already packaged and contains the necessary daemon and CLI components.

sudo pkg install tailscale

From here though we need to edit /etc/rc.conf to enable the tailscaled agent daemon. Add the following line to the file.

tailscaled_enable="YES"

With this, we can activate the tailscaled agent.

sudo service tailscaled start

And with the tailscaled agent started, it is now possible to authenticate using the tailscale CLI tool.

build freebsd image using packer

Tue, 28 Nov 2023 00:00:00 +0000

Last time we covered booting a FreeBSD VM image using QEMU. This is great for setting up a VM by hand or just logging in to play around and experiment with the system. For production use though we want to automate configuration of the FreeBSD VM image so that we are not manually running install commands and editing configuration files, etc.

For this we use Packer from HashiCorp.

Once again these instructions assume an amd64 (x86_64) Ubuntu 22.04 host machine and using QEMU as the virtualization layer. The Linux KVM hardware accelerator can be used to improve performance.

run freebsd in qemu on linux

Sat, 25 Nov 2023 00:00:00 +0000

This blog post serves as a basic guide on how to run a FreeBSD 14 virtual machine in QEMU using a headless Ubuntu 22.04 linux host. The assumption here is both the host and guest operating systems are amd64 (a.k.a. x86_64).

Install QEMU packages

Install the qemu-system-x86 package. Do not install recommended packages as those will include a huge number of graphics packages we are not interested in for a headless server.

minimal consul acl connect setup

Thu, 20 Oct 2022 00:00:00 +0000

Often I need a simple Consul agent with ACLs and Connect enabled for a simple reproduction or demonstration. Although there is a “dev mode” Consul really does not make things easy for getting setup beyond a trivial, insecure agent.

This post is a step-by-step guide on how I start from nothing and end up with a single-node Consul cluster usable with ACLs and Connect enabled. This guide is not about creating a production - ready Consul cluster, for that take a look at the official Consul ACL Guide.

retro on a generic Go test assertions library

Sun, 16 Oct 2022 00:00:00 +0000

[Introduction]

About six months ago shoenig/test was created as a modern, generics based alternative to the popular stretchr/testify testing assertions library for Go. As a developer of Hashicorp’s Nomad project, I am lucky to work with a team willing to endure my experiments. This blog-post is a look back on how well Go generics have (and haven’t) served us in using the library on a real, very large Go project.

There were three guiding principles in designing the library.

go generics for the busy gopher

Sun, 06 Feb 2022 00:00:00 +0000

What is this?

Go 1.18 introduces generics to the language
Some may just want to skim Go generics by example in a blog post
This is that blog post

(!) caution: Language pedants may find the content triggering

new terms

generics: The idea that type information can be determined not during implementation, but later

type parameter: The thing that represents potential types during implementation

constraint: The thing that places restrictions on a type parameter

attempts to make python fast

Tue, 20 Oct 2020 00:00:00 +0000

Posted on Hacker News there was an Implementation Plan for making CPython (the official Python implementation) faster. The author claims a 5x speedup is possible for the low cost of $2 million USD.

The four step plan includes

creating an adaptive interpreter
improvements to internal types
creating a JIT compiler
extending the JIT compiler

We have witnessed other attempts at making Python fast, each achieving their own degree of success in terms of performance and compatibility. For posterity I started keeping a list of them here, in no particular order.

app-indicator output devices

Sun, 18 Oct 2020 00:00:00 +0000

With multiple sound output devices (headphones, speaker, etc.), selecting which one to use on Ubuntu 20.04 is kinda clumsy. Doing so involves opening Settings, navigating to Sound, selecting the Output device. Windows and macOS make it easy by having a selector on the taskbar. With a gnome extension Ubuntu can also have a sound output device selector on the task bar.

The best I’ve found is gse-sount-output-device-chooser.

Installation leaves a little to be desired, but it’s not too bad (and does not require root or adding custom PPAs).

disable mitigations on ubuntu

Sun, 24 May 2020 00:00:00 +0000

By default, most Linux distributions will now enable kernel level mitigations for CPU bugs such as the infamous meltdown and spectre. These mitigations are extremely important for the use of running untrusted code (e.g. cloud VMs), but are less relevant for personal computers, who generally control what is being executed.

The mitigations will impact performance, especially for workloads that incur context switching. In particular, developers are highly impacted, since our build tools are all reading and writing files to/from disk, for tasks from compilation to code auto-completion. There is room for noticeable performance improvement by disabling mitigations.

hclfmt as a service

Sat, 28 Mar 2020 00:00:00 +0000

As a developer working on Nomad, there is frequent need for copying and pasting snippets of HCL to and from GitHub issues, documentation, example files, etc. While the normal hclfmt command works great on files, it is not so convenient for web mediums. Pasted HCL snippets into issues and comments tend to be malformed.

Not anymore! sethops1.net/hclfmt is a simple HCL formatter for conveniently formatting those HCL snippets before pasting them into GitHub issues and the like. Under the hood it runs the hcl2 formatter as a library, so the output is guaranteed to be well formed. It does the same syntax validation the hclfmt command would do.

extract columns with fields

Sat, 29 Feb 2020 00:00:00 +0000

About a month ago there was an article on Hacker News, “Why Learn Awk? (2016)”. In the comments of that article was a thread started by jerf, which spoke directly to my own sentiments. Parsing columns of text should be simpler!

There are of course the timeless tools like awk and cut - which when combined with other arcane shell-isms can accomplish literally anything. But I wanted something that was simple, for a recurring use case that did not feel well served by the status quo.

rename pulseaudio sinks

Fri, 21 Feb 2020 00:00:00 +0000

One might notice the names of output devices (called “sinks”) that pulseaudio uses are not very pretty. Unless you’re a robot, something like pci-0000_0b_00.1.hdmi-stereo just doesn’t roll off the tongue quite like, Headphones. Fortunately, there is a way to reconfigure the names to whatever we want them to be.

First, get a list of the sinks pulseaudio is aware of …

$ pacmd list-sinks | grep name: | fields 2 | sed 's/<//g' | sed 's/>//g'
alsa_output.usb-0c76_USB_PnP_Audio_Device-00.analog-stereo
alsa_output.pci-0000_0b_00.1.hdmi-stereo
alsa_output.pci-0000_0d_00.4.analog-stereo

That is a list of the three output devices plugged into my computer. Actually, one of them happens to be a microphone that reports itself as both a sink and a source.

goroutine closure rule

Tue, 18 Feb 2020 00:00:00 +0000

One of the most insidious gotcha’s when coding in Go is the behavior of variable scoping in the context of goroutine closures. By now this foot-gun is well established. Consider the snippet …

for i := 0; i < 10; i++ {
	go func() {
		fmt.Println(i)
	}()
}

On first glance this should print the values [0,10), though perhaps in random order since the print statements are being launched in goroutines. However, the real output is not determinable. Why? Because the variable i inside the launched closure is not copied - it’s the same i that is being manipulated in the goroutine that the for-loop is executing in. Notice go vet captures this bug …

Seth Hoenig

Thu, 06 Feb 2020 00:00:00 +0000

History of employment

CattleCloud (2025 - present) – Founder
HashiCorp (2019 - 2025) – SWE
Indeed (2014 - 2019) – SWE
IBM (2013 - 2014) – SWE

History of education

University of Texas at Austin (2009-2012)
- B.S.C.S

Facial recognition

For fun

Coding in Go
Building things
Reading; top picks
- Cat’s Cradle by Kurt Vonnegut
- Apple In China by Patrick McGee
- Masters of Doom by David Kushner