Index of every page on this site: llms.txt
All skills
wado-lang avatar

/git-upstream-sync

@b208ba0
by Wadowado-lang/wado121 stars
2
Run it yourself

The only way to merge origin/main into a branch, conflicts or not. Use it every time main comes in — a PR reported conflicting or DIRTY, a branch behind origin/main, a CI failure to reproduce on the merged tree. Never run `git merge origin/main` or `git pull origin main` by hand.

  • 1 file
  • 3.5 KB
  • Updated 2 weeks ago
  • GitHub
Use this Skill: https://skilld.dev/gh/wado-lang/wado/git-upstream-sync

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SKILL.md

≈75 tokens for metadata: the name and description. ≈824 when used: this file.

Description uses 3.8% of example budget

Before choosing a Skill, your Agent reads its name and description. All available Skills share that space.

  • A shorter description leaves more room for other Skills. This entry exceeds our 1% size suggestion.

In our Claude Code example, all Skill names and descriptions share 8,000 characters. This Skill uses ≈300 characters, or 3.8%.

The 1% threshold is a size suggestion. Longer descriptions can still fit.

Your model, settings, and other Skills decide how much text your Agent can read.

Example settings and source

The example uses a 200k-token context and default Claude Code settings. The count includes the name, description, separators, and when_to_use when present. Codex also counts local file paths.

Skit's source and limits: Codex 0.160.1, Claude Code 2.1.292.

Overview

Merge origin/main into the current branch. Follow every step whether or not the merge conflicts: a clean merge still ends with the sanity check, except as step 4 says.

Procedure

1. Fetch and merge with zdiff3

If the branch is on origin, pull it first: CI's tidy job may have pushed chore: tidy onto it. If that pull conflicts, take it through steps 2 and 3, then start again here, so origin/main is merged too. Then move the checked-out submodules to the pointers the merge recorded, so the next git add -A does not stage a stale checkout.

b=$(git branch --show-current)
if git ls-remote --exit-code --heads origin "$b" > /dev/null; then
  git pull --no-rebase origin "$b" || exit 1   # a conflict stops here
fi
git fetch origin main
git -c merge.conflictstyle=zdiff3 merge origin/main
git submodule update --recommend-shallow

2. If conflicts exist, commit them as-is

If the merge produces conflicts, commit the conflict markers without resolving them first. This records the raw conflict state in a dedicated commit, separate from the resolution.

git add -A
git diff --cached origin/main -- vendor
git commit -m "merge origin/main (conflicts unresolved)"

The vendor diff prints nothing unless the branch bumps a submodule itself. Anything else is a stale checkout that git add -A staged, reverting main's bump. Restore main's pointer for each such path before committing, and check again after every git add -A, step 3's included:

git update-index --cacheinfo 160000,"$(git rev-parse origin/main:vendor/wasmtime)",vendor/wasmtime

3. Resolve conflicts

After committing the unresolved state:

  1. Read each conflicted file and understand both sides of the conflict
  2. Resolve conflicts in code files (remove conflict markers, choose correct code)
  3. For generated files (golden fixtures, generated parsers, …) that conflict, regenerate them by re-running their generator and commit the fresh output — do not hand-resolve the markers or defer them (see "Generated files" below)
  4. Stage and commit the resolution:
git add -A
git diff --cached origin/main -- vendor   # as in step 2
git commit -m "resolve merge conflicts"

4. Run mise run test for sanity check

CI applies clippy and format, so a quick sanity check is sufficient.

The one exception is a merge whose only conflict is benchmark/fuel.json. Regenerate it with mise run update-baselines, commit that as the resolution, and skip the tests.

Important

  • Always use -c merge.conflictstyle=zdiff3 so the merge base is visible in conflict markers (with zealous zdiff3 reducing noise)
  • Always commit the unresolved conflicts first, then resolve in a separate commit — this preserves a clear record of what the conflicts looked like vs how they were resolved
  • Do NOT squash the two commits together

Generated files

  • Always regenerate golden / generated files and commit. Re-run the generator (do not hand-resolve the markers) and commit its output. Do not skip them or defer them to on-task-done / CI.
  • Never silently discard generated output (e.g. git restore to clean the tree or quiet a hook). Commit it, or ask. Discarding generated work without asking is always wrong.

Source: SKILL.md on GitHub

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No third-party reports yet.

Signed by skilld at b208ba0. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub yesterday.

Activeupdated 2 weeks ago

README badge

README badge for wado-lang/wado/git-upstream-sync