The Day One File Looked Different on Two Devices
Last Thursday, one of our team bots on the Mac mini opened a card news draft file, and every bit of Korean text came out as garbled characters. The exact same file opened fine on the Windows device. At first we thought the file was corrupted and tried re-saving it, but the result was the same.
Tracing the cause, we found that the script on the Windows side was saving files using the system default encoding, cp949. The Mac mini reads files with UTF-8 as default, so Korean characters showed up broken while English letters and numbers stayed intact. That explained why only the Korean text was affected.
Instead of fixing it right away, we asked Robo first: how many other scripts might be saving files without specifying an encoding. Turned out there were three, the card news drafts, the translation comparison sheet, and the publishing log. All three had the exact same issue.
The fix itself was simple. We added one rule across every file-saving script, explicitly setting `encoding="utf-8"`. It was basically telling the code that default values differ by device. After that, files read the same whether they were saved from the Windows device or the Mac mini.
It was a small problem, but what struck us was that we had never treated it as obvious that the same file could look different depending on the device. If we had only checked one script, we might have missed the other two, but checking all three at once caught the issue before it spread further.
For any team running both a Windows device and a Mac mini together, this one rule turned out to be pretty useful.