I started running trail ultras again in 2019. I finished every race I started, building up from 50K to 100M. At the 100M, I ate more frequently and more calories than ever before. I remember telling people that the thought of dropping out never entered my mind, not even at 3am when I was falling asleep on my feet.

After that race, we decided that more carbs and calories were better. As a result, in 2022 I switched to using Spring Energy, specifically Awesome Sauce, as my primary fuel. On long training runs and in races, I would eat one hourly, and rotate other fuels on the half hours. Here are my results from 2019 until now in all of the trail ultras I’ve started:

To me, it’s pretty clear what happened here. I went from a reliable finisher to not being able to find a finish line beyond 50K when I switched to Spring. I thought I was hitting these high carbs and calories, because I trusted nutrition labels to be at least close to accurate. In all of my inner reflection about why I couldn’t finish a race, questioning the labels never occurred to me.

Of course, I take some of the blame here. This is one of those, “If it’s too good to be true, it probably is,” situations. My training wasn’t great in 2022, so that was the obvious thing to focus on and fix. I got a coach for 2023 to try to fix it, but I still couldn’t finish.

2024 and Beyond

It’s time to get back to my pre-Spring finishing ways. But I’m in a much better position now than I was in 2022.

Now I have a coach. We’re doing cycling to get “free fitness” on days I’m not running. We’re doing strength & mobility to focus on my injury-prone areas. We’re doing hill repeats and speed workouts as opposed to only slow stuff. I’m more athletic than I was in 2022.

And I’m going to use those lab tests that people are getting on other nutrition products to make sure I’m actually fueling properly. My ceiling is so much higher than where I ended 2022.

I’m off for a run…

HStack {
    Image(systemName: "globe")
    Text("Hello, world!")
}

Label("Hello, world!", systemImage: "globe")

In addition to more concise and expressive, the Label is also better for accessibility. It applies the .accessibilityElement(children: .combine) modifier to group its subviews.

This allows accessibility tools like VoiceOver and Switch Control to have larger, more semantic hit areas with more meaningful feedback. Here is what gets selected as you go through these subviews with VoiceOver:

View	Image	Text	Label
Screenshot
Accouncement	Globe, image. Globe.	Hello, world!	Hello, world!

By using the Label, you’ve created a better accessibility UX. While you could use the HStack and the .accessibilityElement(children: .combine) modifier, using a Label is simpler and will get you platform native behavior when it’s time to use the code on different platforms, as well.

// Naive implementation -- probably not what you want to do
var body: some View {
    Text("Awesome Sauce")
        .onAppear {
            Task {
                // Async task that is NOT cancelled when view disappears.
            }
        }
}

Each time the View appears, a new task will be created. If this is in a NavigationStack where views can be pushed and popped often, you may end up with many tasks running for the same view or doing background work for a view that no longer exists.

SwiftUI supports this common pattern with the .task view modifier. The primary benefit of this over the naive approach is support for cancellation when the view disappears.

// Improved implementation -- supports task cancellation
var body: some View {
    Text("Awesome Sauce")
        .task {
            // Async task that is cancelled when the View disappears.
        }
}

Supporting Task Cancellation

Just because a task is cancelled doesn’t mean that it’s terminated. That would result in undefined behavior. Instead, it is up to the developer to check if a task is cancelled and handle it appropriately for the workload that’s behing performed. From the Swift documentation:

Depending on the work you’re doing, that usually means one of the following:

• Throwing an error like CancellationError
• Returning nil or an empty collection
• Returning the partially completed work

There are two primary ways of checking if the task has been cancelled:

• Calling Task.checkCancellation() will throw a CancellationError if the task has been cancelled.
• Checking Task.isCancelled and handling it appropriately.

There are static and instance versions of these. If inside the task, use the static versions and it will report on the currently running task. If you have an instance of a task, use the instance version to check on that specific task’s status.

When I changed from a NavigationView to a NavigationStack, it started working. A good reminder to follow the guide to migrate to new navigation types. As with most things Apple does, move quickly with the direction they’re heading and your life will be easier.

I ended up wasting a bunch of time because I named the file time.dat. I tried the following set of actions and it kept returning the text time:

1. File - I set the correct path to time.dat
2. Split Text by New Lines
3. Get Last Item from Split Text

This should have returned the last line from the file, but I kept getting time. Finally, I stumbled upon this Reddit post that had the key: the file must have a .txt extension.

I didn’t want to rename the actual file on disk, but the post says you can use the “Set Name” action to rename it within the Shortcut. Now my Shortcut has these actions and it works like a charm:

1. File - I set the correct path to time.dat
2. Set name of File to time.txt
3. Split Text by New Lines
4. Get Last Item from Split Text

I’m writing this up in hopes of saving someone else from wasting as much time as I did, most likely my future self.

I’ll be using .example domains throughout to avoid referencing any real emails or GitHub resources.

# ~/.config/git/config
[user]
    name = Branton Boehm
    email = branton@personal.example

[includeIf "hasconfig:remote.*.url:git@github.example:Work/**"]
    path = config-work

# ~/.config/git/config-work
[user]
    email = branton@work.example

I can verify this is working by checking the email from within a git clone:

~/src/website $ git remote -v
origin  git@github.example:branton/website.git (fetch)
origin  git@github.example:branton/website.git (push)

~/src/website $ git config --get user.email
branton@personal.example

~/src/api $ git remote -v
origin  git@github.example:Work/api.git (fetch)
origin  git@github.example:Work/api.git (push)

~/src/api $ git config --get user.email
branton@work.example

If this isn’t working as expected, the git config --list --show-origin command will display the config file each setting is coming from.

Similarly to macOS, you can now remap modifier keys on hardware keyboards in iPadOS. Find it in Settings > General > Keyboard > Hardware Keyboard > Modifier Keys.

I found that I rarely had a good use for the Globe key, so that’s what I’ve remapped to Escape. If you do this, you can still cycle through your keyboards (including emojis) by using Control+Space.

Remappings are on a per-hardware keyboard basis. I use both the Smart Keyboard Folio and also the Kinesis Advantage2. iPadOS keeps remappings separate between them.

Escape Works in More than Vim

If you’re used to using Escape in macOS, it “just works” in a lot of places in iPadOS, too. For example, if a dialog has a Cancel button then pressing Escape will dismiss it. It’s one more reason to, as an iOS/iPadOS developer, use native controls instead of building custom UI. Supporting the Escape key is unlikely to be a high priority for your product manager, so you can sneak it in by using native controls. Nobody is going to complain that they got better keyboard functionality for free!

Escape in Blink Before iPadOS

Blink itself has a solution for the lack of an Escape key. It allows you to remap keys and I had remapped Caps Lock. It was done in software locally within Blink, so when switching to another app, TEXT ALWAYS STARTED IN ALL CAPS. I got used to it, but it wasn’t ideal.

Looking back at Blink’s settings now, there were probably better options available to me. I rarely use the right Option key, for example, but it’s also uncomfortable to hit. Remapping the Globe key works far better for me.

direnv is an extension for your shell. It augments existing shells with a new feature that can load and unload environment variables depending on the current directory.

In my specific scenario, I have the following .envrc file in the working directory for my ledger project:

# .envrc
export LEDGER_FILE=transactions.dat
export LEDGER_DATE_FORMAT=%Y/%m/%d

When I cd into the directory, direnv loads the file:

$ cd finances
direnv: loading .envrc
direnv: export +LEDGER_DATE_FORMAT +LEDGER_FILE
$ echo $LEDGER_FILE
transactions.dat

And when I cd back out of the directory, direnv unloads:

$ cd ..
direnv: unloading
$ echo $LEDGER_FILE

This allows me to keep project-specific configuration committed to the same git repository as the project itself. It’s available to install in most Unix-like distributions. I have it in my Brewfile for macOS and bootstrap.sh for Raspbian, and have added eval "$(direnv hook zsh)" to my .zsh_profile.

It’s done by writing to the user defaults store from the Terminal:

$ defaults write com.apple.dock persistent-apps -array-add '{"tile-type"="small-spacer-tile";}'
$ killall Dock

This will add a small spacer to the end of the main section of your dock, which can then be dragged to wherever you’d like it. Add as many as you need to create your groups. To remove one, drag it out of the dock and release. A larger spacer-tile is also avaiable if you prefer.

Here on the site, I finally posted my race recap from winning the innagural Tuscazoar 50 miler last August. I also archived last year’s training and created some training charts to compare data from each year.

This is the first of what I hope to be monthly updates, especially since I plan on racing less often this year. Not a whole lot of interesting stuff this month. The biggest takeaway was that I got a consistent month of winter training in, putting me way ahead of where I was last year at this time.

Tempo Runs

In week 5 of the training plan, speedwork is introduced on Thursdays. I’m going to stick to tempo runs in this training block and add in long hill repeats in my primary training block.

The first tempo run, I didn’t know what to expect. I knew I was running faster than this time last year, but I’ve also been sticking to pavement whereas last year I was trying to get onto trails as much as possible. After a post-work nap, I headed out. I did a 20 minute warmup, an 18 minute tempo run, and then cooled down for the rest of the 7 miles. I was very happy with the pace, averaging 6:25/mile with my last rolling mile at 6:05.

A week later, I had the same workout and took the same post-work nap. This time, I covered a bit more ground and averaged 6:19/mile. (My initial math was wrong and I thought I averaged 6:05/mile… nope.)

It’s a lot of fun pushing the pace like this. It’s sparking some imagination, too. I’m wondering what I could do in a 5K on the track this summer.

Friday Run

The training plans I use are from Relentless Forward Progress: A guide to Running Ultramarathons. They are 5 days/week, with Mondays and Fridays as rest days. This year, I was planning on adding recovery runs on Fridays in my primary training block. My legs have been feeling great, so I decided to introduce them earlier and got my first one in on the last day of the month.

It was my 23rd run of the month, topping my best months from last year with 22 runs in them. I just ran up to the local park and ran around the soccer field for a while, not worried about pace or distance.

Ideally, I’ll build up to ~5 miles during my primary training block, but I’ll play it by ear and do what feels best for my legs.