#!/usr/bin/env bash
set -euo pipefail
cd "$(dirname "$0")/.."

device_id="<your-device-id>"   # find with: xcodebuild -showdestinations ...
bundle_id="<your-bundle-id>"
derived_data="/tmp/lexitaire-device-build"

xcodebuild build -project Lexitaire.xcodeproj -scheme Lexitaire \
  -destination "id=$device_id" -allowProvisioningUpdates \
  -derivedDataPath "$derived_data"

app_path="$derived_data/Build/Products/Debug-iphoneos/Lexitaire.app"
xcrun devicectl device install app --device "$device_id" "$app_path"
xcrun devicectl device process launch --device "$device_id" "$bundle_id" \
  || echo "Installed, but couldn't launch it automatically (often just a locked screen)."

Two small decisions worth calling out:

• Pin the device by ID, not by name. A device name can contain an apostrophe (a shell-quoting headache) and can change if you rename the phone; the ID survives both. Find it with xcodebuild -showdestinations -project Lexitaire.xcodeproj -scheme Lexitaire.
• Don’t treat a failed launch as a failed run. devicectl fails to launch on a locked screen, which isn’t a broken build — the install already succeeded by that point, so that step just fails soft with a reminder to unlock and tap the icon.

Now it’s just tools/install-to-iphone.sh and the build’s on my phone.

sudo dscacheutil -flushcache; sudo killall -HUP mDNSResponder

It’s really two separate flushes chained together, because macOS caches DNS answers in two different places depending on the OS version:

• dscacheutil -flushcache clears the Directory Service cache — the layer that backs dscacheutil -q host lookups and, on older macOS releases, most of what actually cached your DNS answers.
• killall -HUP mDNSResponder sends a HUP signal to mDNSResponder, the daemon that does the actual DNS (and Bonjour/mDNS) resolution on modern macOS. A HUP tells it to reload rather than crash, which in practice clears its internal cache.

Which one actually matters has shifted across macOS versions, and Apple doesn’t document it cleanly — so rather than remember which release needs which command, running both covers you regardless of what you’re on. The semicolon (rather than &&) matters a little here too: it runs the second command whether or not the first “succeeded,” which is what you want, since dscacheutil -flushcache doesn’t reliably signal failure even when there’s nothing to flush.

You’ll be prompted for your password (possibly twice, if sudo‘s credential cache has already expired between the two commands). There’s no output on success — it just quietly works.

Make it a one-word command

If you find yourself reaching for this more than once, drop a function in your ~/.zshrc (or ~/.bashrc):

flush-dns() {
    sudo dscacheutil -flushcache
    sudo killall -HUP mDNSResponder
    echo "DNS cache flushed."
}

Reload your shell config, and flush-dns is a real command from then on.

$ flush-dns
DNS cache flushed.

Оказывается, это крошечное отклонение порождает удивительное количество трений.

Усталость от «поиска и замены»

Конечно, я мог бы вручную искать и заменять кавычки. Но делать это раз за разом в разных файлах утомительно. Тогда я попробовал использовать Cursor, но это принесло целый новый набор проблем: большой расход токенов, медленную обработку и необходимость подталкивать его на каждом шагу, как упрямого робота-ассистента.

Это сбивало мой рабочий процесс.

На сцену выходит Unify (и uv)

У меня уже был установлен uv, так что подключить Unify было проще простого:

uv tool install --python=python3.10 unify

Этот параметр --python=python3.10 — ключевой: иначе uv по умолчанию берёт Python 3.13, что вызывает проблемы с lib2to3 (да, это всё ещё актуально!).

Traceback (most recent call last):
  File "/Users/abramowi/.local/bin/unify", line 10, in <module>
    sys.exit(main())
             ~~~~^^
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 230, in main
    return _main(sys.argv,
                 standard_out=sys.stdout,
                 standard_error=sys.stderr)
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 210, in _main
    if format_file(name, args=args, standard_out=standard_out):
       ~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 144, in format_file
    encoding = detect_encoding(filename)
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 125, in detect_encoding
    from lib2to3.pgen2 import tokenize as lib2to3_tokenize
ModuleNotFoundError: No module named 'lib2to3'

Unify зависит от него, так что более старая версия Python спасает положение.

Однострочная команда, которая всё исправила

Вот и всё, что нужно, чтобы переключиться на одинарные кавычки прямо в файле:

unify --quote "'" --in-place /path/to/file.py

И вот так, без лишней возни. Никакой траты токенов. Никакого подталкивания ИИ-инструментов к послушанию.

До и после

# Before
print("Hello, world!")

# After
print('Hello, world!')

Вывод

Иногда лучшее решение — точечное. Unify не просто привёл в порядок мой код — он вернул мне время и концентрацию. Если вы застряли в чистилище стилей кавычек, эта однострочная команда может стать вашим спасением.

结果这个小小的偏差却带来了惊人的麻烦。

「查找和替换」疲劳

当然，我可以手动查找并替换引号。但在多个文件中反复这样做很繁琐。后来我尝试用 Cursor 来处理，结果却带来了一整套新的麻烦：大量的 token 消耗、缓慢的处理速度，还需要每隔几步就推它一下，就像哄一个固执的机器人助手。

这打乱了我的工作节奏。

Unify（和 uv）登场

我已经安装了 uv ，所以引入 Unify 非常顺畅：

uv tool install --python=python3.10 unify

这个 --python=python3.10 的部分很关键——否则 uv 会默认使用 Python 3.13，这会导致 lib2to3 出问题（没错，它现在依然相关！）。

Traceback (most recent call last):
  File "/Users/abramowi/.local/bin/unify", line 10, in <module>
    sys.exit(main())
             ~~~~^^
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 230, in main
    return _main(sys.argv,
                 standard_out=sys.stdout,
                 standard_error=sys.stderr)
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 210, in _main
    if format_file(name, args=args, standard_out=standard_out):
       ~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 144, in format_file
    encoding = detect_encoding(filename)
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 125, in detect_encoding
    from lib2to3.pgen2 import tokenize as lib2to3_tokenize
ModuleNotFoundError: No module named 'lib2to3'

Unify 依赖它，所以用旧版本的 Python 就能解决问题。

解决一切的单行命令

只需要这一步，就能原地切换为单引号：

unify --quote "'" --in-place /path/to/file.py

就这样，不再需要折腾。不再浪费 token。也不用再去哄着 AI 工具听话了。

前后对比

# Before
print("Hello, world!")

# After
print('Hello, world!')

总结

有时候，最好的解决方案就是精准的一刀。Unify 不仅整理了我的代码——还让我重新夺回了时间和专注力。如果你也困在引号风格的炼狱中，这一行命令或许就是你的逃生通道。

ところが、このほんの小さな逸脱が驚くほどの摩擦を生むことが分かりました。

「検索と置換」疲れ

もちろん、引用符を手動で検索・置換することもできます。しかし、それを複数のファイルに渡って繰り返すのは面倒です。そこでCursorを使ってみましたが、それはそれで新たな頭痛の種を生みました。大量のトークン消費、遅い処理、そして頑固なロボットアシスタントのように何ステップかごとに突く必要があること。

それは私の作業の流れを乱していました。

Unify（とuv）の登場

すでに uv をインストールしていたので、Unifyを導入するのはシームレスでした：

uv tool install --python=python3.10 unify

この --python=python3.10 という部分が重要です——そうしないと uv はデフォルトでPython 3.13を使ってしまい、 lib2to3 に問題が生じます（そう、今でも関係あるんです！）。

Traceback (most recent call last):
  File "/Users/abramowi/.local/bin/unify", line 10, in <module>
    sys.exit(main())
             ~~~~^^
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 230, in main
    return _main(sys.argv,
                 standard_out=sys.stdout,
                 standard_error=sys.stderr)
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 210, in _main
    if format_file(name, args=args, standard_out=standard_out):
       ~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 144, in format_file
    encoding = detect_encoding(filename)
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 125, in detect_encoding
    from lib2to3.pgen2 import tokenize as lib2to3_tokenize
ModuleNotFoundError: No module named 'lib2to3'

Unifyはこれに依存しているので、古めのPythonバージョンが救いになります。

すべてを解決した一行コマンド

その場で単一引用符に切り替えるのに必要なのはこれだけです：

unify --quote "'" --in-place /path/to/file.py

それだけで、もう手間はかかりません。トークンの浪費もありません。AIツールを従わせるために突く必要もありません。

変更前と変更後

# Before
print("Hello, world!")

# After
print('Hello, world!')

まとめ

時には、最善の解決策は外科的なものです。Unifyは単にコードを整えただけでなく、時間と集中力を取り戻させてくれました。もしあなたが引用符スタイルの煉獄で立ち往生しているなら、この一行コマンドが脱出口になるかもしれません。

Acontece que esse pequeno desvio gera uma quantidade surpreendente de atrito.

A fadiga do “localizar e substituir”

Claro, eu poderia buscar e substituir as aspas manualmente. Mas fazer isso repetidamente em vários arquivos é cansativo. Então tentei usar o Cursor para cuidar disso, mas isso trouxe um conjunto totalmente novo de dores de cabeça: alto consumo de tokens, processamento lento, e a necessidade de cutucá-lo a cada poucos passos como um assistente robô teimoso.

Isso estava atrapalhando meu fluxo de trabalho.

Chega o Unify (e o uv)

Eu já tinha o uv instalado, então incorporar o Unify foi tranquilo:

uv tool install --python=python3.10 unify

Esse detalhe do --python=python3.10 é essencial—caso contrário o uv pega o Python 3.13 por padrão, o que causa problemas com o lib2to3 (é, ainda relevante!).

Traceback (most recent call last):
  File "/Users/abramowi/.local/bin/unify", line 10, in <module>
    sys.exit(main())
             ~~~~^^
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 230, in main
    return _main(sys.argv,
                 standard_out=sys.stdout,
                 standard_error=sys.stderr)
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 210, in _main
    if format_file(name, args=args, standard_out=standard_out):
       ~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 144, in format_file
    encoding = detect_encoding(filename)
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 125, in detect_encoding
    from lib2to3.pgen2 import tokenize as lib2to3_tokenize
ModuleNotFoundError: No module named 'lib2to3'

O Unify depende dele, então uma versão mais antiga do Python salva o dia.

A linha de comando que resolveu tudo

É só isso que é preciso para trocar para aspas simples direto no arquivo:

unify --quote "'" --in-place /path/to/file.py

E pronto, sem mais ajustes manuais. Sem mais gasto de tokens. Sem mais precisar empurrar as ferramentas de IA para obedecerem.

Antes e depois

# Before
print("Hello, world!")

# After
print('Hello, world!')

Conclusão

Às vezes a melhor solução é uma intervenção cirúrgica. O Unify não só limpou meu código—ele me permitiu recuperar tempo e foco. Se você está preso no purgatório do estilo de aspas, essa linha de comando pode ser sua rota de fuga.

Es stellt sich heraus, dass diese winzige Abweichung überraschend viel Reibung erzeugt.

Die „Suchen und Ersetzen“-Erschöpfung

Klar, ich könnte Anführungszeichen manuell suchen und ersetzen. Aber das wiederholt über mehrere Dateien hinweg zu tun, ist mühsam. Dann versuchte ich es mit Cursor, aber das brachte eine ganz neue Reihe von Kopfschmerzen mit sich: hoher Token-Verbrauch, langsame Verarbeitung, und die Notwendigkeit, es alle paar Schritte anzustupsen wie einen sturen Roboter-Assistenten.

Es hat meinen Arbeitsfluss gestört.

Unify (und uv) betritt die Bühne

Ich hatte bereits uv installiert, also war die Einbindung von Unify nahtlos:

uv tool install --python=python3.10 unify

Das --python=python3.10-Detail ist entscheidend—sonst greift uv standardmäßig auf Python 3.13 zurück, was Probleme mit lib2to3 verursacht (ja, immer noch relevant!).

Traceback (most recent call last):
  File "/Users/abramowi/.local/bin/unify", line 10, in <module>
    sys.exit(main())
             ~~~~^^
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 230, in main
    return _main(sys.argv,
                 standard_out=sys.stdout,
                 standard_error=sys.stderr)
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 210, in _main
    if format_file(name, args=args, standard_out=standard_out):
       ~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 144, in format_file
    encoding = detect_encoding(filename)
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 125, in detect_encoding
    from lib2to3.pgen2 import tokenize as lib2to3_tokenize
ModuleNotFoundError: No module named 'lib2to3'

Unify hängt davon ab, also rettet eine ältere Python-Version den Tag.

Der Einzeiler, der alles gelöst hat

Das ist alles, was nötig ist, um direkt in der Datei auf einfache Anführungszeichen umzustellen:

unify --quote "'" --in-place /path/to/file.py

Und schon war Schluss mit dem Herumfummeln. Schluss mit Token-Verschwendung. Schluss damit, KI-Tools zur Fügsamkeit zu drängen.

Vorher und nachher

# Before
print("Hello, world!")

# After
print('Hello, world!')

Fazit

Manchmal ist die beste Lösung ein chirurgischer Eingriff. Unify hat nicht nur meinen Code bereinigt—es hat mir Zeit und Konzentration zurückgegeben. Wenn du im Anführungszeichen-Fegefeuer feststeckst, könnte dieser Einzeiler genau dein Fluchtweg sein.

Il s’avère que cette petite entorse engendre une friction surprenante.

La lassitude du « Rechercher et remplacer »

Bien sûr, je pourrais rechercher et remplacer les guillemets manuellement. Mais le faire à répétition sur plusieurs fichiers est fastidieux. J’ai alors essayé d’utiliser Cursor pour s’en charger, mais cela a introduit toute une nouvelle série de tracas : une grosse consommation de tokens, un traitement lent, et le besoin de le relancer à chaque étape comme un assistant robot têtu.

Cela perturbait mon flux de travail.

Entrée en scène de Unify (et uv)

J’avais déjà uv installé, donc intégrer Unify a été sans accroc :

uv tool install --python=python3.10 unify

Ce détail --python=python3.10 est essentiel—sinon uv récupère Python 3.13 par défaut, ce qui pose des problèmes avec lib2to3 (eh oui, toujours d’actualité !).

Traceback (most recent call last):
  File "/Users/abramowi/.local/bin/unify", line 10, in <module>
    sys.exit(main())
             ~~~~^^
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 230, in main
    return _main(sys.argv,
                 standard_out=sys.stdout,
                 standard_error=sys.stderr)
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 210, in _main
    if format_file(name, args=args, standard_out=standard_out):
       ~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 144, in format_file
    encoding = detect_encoding(filename)
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 125, in detect_encoding
    from lib2to3.pgen2 import tokenize as lib2to3_tokenize
ModuleNotFoundError: No module named 'lib2to3'

Unify en dépend, donc une version plus ancienne de Python sauve la mise.

La ligne de commande qui a tout réglé

Voici tout ce qu’il faut pour passer aux guillemets simples directement dans le fichier :

unify --quote "'" --in-place /path/to/file.py

Et comme ça, plus de bricolage. Plus de gaspillage de tokens. Plus besoin de forcer les outils d’IA à obéir.

Avant et après

# Before
print("Hello, world!")

# After
print('Hello, world!')

Ce qu’il faut retenir

Parfois, la meilleure solution est chirurgicale. Unify n’a pas seulement nettoyé mon code—il m’a permis de récupérer du temps et de la concentration. Si vous êtes coincé dans le purgatoire du style des guillemets, cette ligne de commande pourrait bien être votre porte de sortie.

Resulta que esta pequeña desviación genera una cantidad sorprendente de fricción.

La fatiga del “buscar y reemplazar”

Claro, podría buscar y reemplazar comillas manualmente. Pero hacerlo repetidamente en varios archivos es tedioso. Luego probé usar Cursor para encargarse de eso, pero eso trajo un conjunto totalmente nuevo de dolores de cabeza: mucho consumo de tokens, procesamiento lento, y tener que empujarlo cada pocos pasos como a un asistente robot testarudo.

Estaba arruinando mi flujo de trabajo.

Llega Unify (y uv)

Ya tenía uv instalado, así que incorporar Unify fue muy sencillo:

uv tool install --python=python3.10 unify

Ese detalle de --python=python3.10 es clave—de lo contrario uv toma Python 3.13 por defecto, lo cual causa problemas con lib2to3 (sí, ¡todavía relevante!).

Traceback (most recent call last):
  File "/Users/abramowi/.local/bin/unify", line 10, in <module>
    sys.exit(main())
             ~~~~^^
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 230, in main
    return _main(sys.argv,
                 standard_out=sys.stdout,
                 standard_error=sys.stderr)
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 210, in _main
    if format_file(name, args=args, standard_out=standard_out):
       ~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 144, in format_file
    encoding = detect_encoding(filename)
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 125, in detect_encoding
    from lib2to3.pgen2 import tokenize as lib2to3_tokenize
ModuleNotFoundError: No module named 'lib2to3'

Unify depende de él, así que una versión más antigua de Python salva el día.

La línea de código que lo arregló todo

Esto es todo lo que se necesita para cambiar a comillas simples en el mismo archivo:

unify --quote "'" --in-place /path/to/file.py

Y así, sin más, se acabaron los ajustes manuales. Se acabó el consumo de tokens. Se acabó tener que empujar a las herramientas de IA para que obedezcan.

Antes y después

# Before
print("Hello, world!")

# After
print('Hello, world!')

Conclusión

A veces la mejor solución es una intervención quirúrgica. Unify no solo limpió mi código—me permitió recuperar tiempo y concentración. Si estás atrapado en el purgatorio del estilo de comillas, este pequeño comando podría ser tu vía de escape.

It turns out this one tiny deviation spawns a surprising amount of friction.

The “Find and Replace” Fatigue

Sure, I could manually search and replace quotes. But doing it repeatedly across files is tedious. Then I tried using Cursor to handle it, but that introduced a whole new set of headaches: large token usage, slow processing, and needing to nudge it every few steps like a stubborn robot assistant.

It was messing with my flow.

Enter Unify (and uv)

I already had uv installed, so bringing in Unify was seamless:

uv tool install --python=python3.10 unify

That --python=python3.10 bit is key—otherwise uv grabs Python 3.13 by default, which causes problems with lib2to3 (yep, still relevant!).

Traceback (most recent call last):
  File "/Users/abramowi/.local/bin/unify", line 10, in <module>
    sys.exit(main())
             ~~~~^^
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 230, in main
    return _main(sys.argv,
                 standard_out=sys.stdout,
                 standard_error=sys.stderr)
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 210, in _main
    if format_file(name, args=args, standard_out=standard_out):
       ~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 144, in format_file
    encoding = detect_encoding(filename)
  File "/Users/abramowi/.local/share/uv/tools/unify/lib/python3.13/site-packages/unify.py", line 125, in detect_encoding
    from lib2to3.pgen2 import tokenize as lib2to3_tokenize
ModuleNotFoundError: No module named 'lib2to3'

Unify depends on it, so an older Python version saves the day.

The One-Liner That Fixed Everything

Here’s all it takes to switch to single quotes in-place:

unify --quote "'" --in-place /path/to/file.py

And just like that, no more fiddling. No more token burn. No more nudging AI tools into submission.

Before & After

# Before
print("Hello, world!")

# After
print('Hello, world!')

Takeaway

Sometimes the best fix is a surgical one. Unify didn’t just clean up my code—it let me reclaim time and focus. If you’re stuck in quote-style purgatory, this one-liner might just be your escape route.