Oliver Davies
e3ac4e5064
Create a hash of the repo path and store it as its own result, allowing for a different cached result per repo path.
42 lines
1.3 KiB
Bash
Executable file
42 lines
1.3 KiB
Bash
Executable file
#!/usr/bin/env bash
|
|
|
|
set -o errexit
|
|
set -o nounset
|
|
set -o pipefail
|
|
|
|
# Default to running in the current directory, but allow for overriding the
|
|
# repo path using an environment variable.
|
|
# e.g. `REPO=../somewhere-else ./git-commit-length-counter`
|
|
REPO="${REPO:-$(pwd)}"
|
|
|
|
# Return early if the repository directory doesn't exist.
|
|
[[ ! -d "${REPO}" ]] && exit 2
|
|
|
|
mkdir -p ./result
|
|
|
|
repo_hash="$(echo -n "${REPO}" | md5sum | cut -d ' ' -f1)"
|
|
result_file="./result/${repo_hash}"
|
|
|
|
echo "Result file: ${result_file}"
|
|
|
|
if [[ ! -f "${result_file}" ]]; then
|
|
for commit_id in $(git -C "${REPO}" rev-list --all --no-merges); do
|
|
echo "Processing commit ${commit_id}..."
|
|
# Calculate the legnth of the commit message.
|
|
commit_message=$(GIT_PAGER=cat git -C "${REPO}" show "${commit_id}" -s --format=%B)
|
|
commit_message_length=$(echo "${commit_message}" | wc -l)
|
|
|
|
# Store the commit IDs and the message length.
|
|
echo "${commit_message_length} ${commit_id}" >> "${result_file}"
|
|
done
|
|
fi
|
|
|
|
# Sort the result file so all files are sorted by the length of the message.
|
|
sort "${result_file}" --reverse --numeric-sort --output "${result_file}"
|
|
|
|
head "${result_file}" --lines 5
|
|
|
|
# Show the commit with the longest message.
|
|
commit_with_the_longest_message=$(head "${result_file}" --lines 1 | cut -d " " -f2)
|
|
git -C "${REPO}" show "${commit_with_the_longest_message}"
|