git-compare-cherried

#!/bin/bash

# Usage: ./compare_branches.sh <upstream_branch> <head_branch>
# Example: ./compare_branches.sh main feature/my-branch

if [ $# -lt 2 ]; then
  echo "Usage: $0 <upstream_branch> <head_branch>"
  exit 1
fi

UPSTREAM=$1
HEAD=$2

echo "Analyzing commits between $UPSTREAM and $HEAD..."

# Get commits that are in HEAD but not in UPSTREAM (similar to git cherry)
UNIQUE_COMMITS=$(git rev-list "$UPSTREAM".."$HEAD")

if [ -z "$UNIQUE_COMMITS" ]; then
  echo "No unique commits found in $HEAD that aren't in $UPSTREAM."
  exit 0
fi

# Normalize a commit message by stripping all PR number suffixes like (#123)
normalize_msg() {
  echo "$1" | sed 's/ (#[0-9]*)\b//g; s/(#[0-9]*)//g' | sed 's/[[:space:]]*$//'
}

# Get all normalized commit messages from upstream for comparison
UPSTREAM_MSGS=$(git log --format="%s" "$UPSTREAM" | while IFS= read -r msg; do normalize_msg "$msg"; done)

echo "Commits in $HEAD that aren't in $UPSTREAM based on commit message:"
echo "--------------------------------------------------------------"

while IFS= read -r commit; do
  # Get the commit message for this commit
  COMMIT_MSG=$(git log -n 1 --format="%s" "$commit")
  NORMALIZED_MSG=$(normalize_msg "$COMMIT_MSG")

  # Check if this normalized commit message exists in upstream (fixed-string match to handle special regex chars)
  if ! echo "$UPSTREAM_MSGS" | grep -qF "$NORMALIZED_MSG"; then
    # This commit message doesn't exist in upstream, so show it
    git show --no-patch --format="%C(yellow)%h%C(reset) - %s %C(green)(%cr) %C(bold blue)<%an>%C(reset)" "$commit"
  fi
done <<< "$UNIQUE_COMMITS"