Skip to content

Commit

Permalink
.github: add shallow checkpatch to Github Actions
Browse files Browse the repository at this point in the history
Let's try to run checkpatch in an open-source and fast way while
leveraging Github's very good user interface.

Run checkpatch both with and without --strict; this has been a source of
confusion in the past, see
thesofproject/linux#1988

Note this attempt does not rely on git merge-bases which has been found
to be not compatible with shallow cloning:
thesofproject/linux#2556

Signed-off-by: Marc Herbert <marc.herbert@intel.com>
  • Loading branch information
marc-hb committed Mar 4, 2021
1 parent 26223f0 commit ea76703
Show file tree
Hide file tree
Showing 2 changed files with 124 additions and 0 deletions.
55 changes: 55 additions & 0 deletions .github/workflows/shallow_checkpatch.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,55 @@
---
# SPDX-License-Identifier: BSD-3-Clause
# Tools that can save round-trips to github and a lot of time:
#
# yamllint -f parsable this.yml
# pip3 install ruamel.yaml.cmd
# yaml merge-expand this.yml exp.yml && diff -w -u this.yml exp.yml
#
# github.com also has a powerful web editor that can be used without
# committing.

name: checkpatch

# yamllint disable-line rule:truthy
on: [pull_request]

jobs:
checkpatch:
runs-on: ubuntu-20.04
env:
PR_NUM: ${{github.event.number}}
CHK_CMD: ./scripts/checkpatch.pl --codespell --codespellfile
/usr/lib/python3/dist-packages/codespell_lib/data/dictionary.txt
steps:
# depth 2 so:
# ^1. we can show the Subject of the current target branch tip
# ^2. we reconnect/graft to the later fetch pull/1234/head,
- uses: actions/checkout@v2
with: {fetch-depth: 2}

- name: yamllint ourselves
run: yamllint .github/workflows/shallow_checkpatch.yml

- name: install codespell
run: sudo apt-get -y install codespell && dpkg -L codespell | grep dict

# See shallowness issue https://github.com/thesofproject/linux/issues/2556
- name: fetch PR commits
run: |
.github/workflows/shallowfetchPRcommits.sh \
${GITHUB_REPOSITORY} "$PR_NUM"
# show what we got
git --no-pager log --oneline --graph --decorate --max-count=50
- name: checkpatch
run: while read -r sha ; do
printf '\n -------------- \n\n';
${CHK_CMD} -g $sha;
done < PR_SHAs.txt

- name: checkpatch --strict
run: while read -r sha ; do
printf '\n -------------- \n\n';
${CHK_CMD} --strict -g $sha;
done < PR_SHAs.txt
69 changes: 69 additions & 0 deletions .github/workflows/shallowfetchPRcommits.sh
Original file line number Diff line number Diff line change
@@ -0,0 +1,69 @@
#!/bin/sh
# SPDX-License-Identifier: BSD-3-Clause

set -e

# This script fetches Pull Request commits missing from a shallow clone
# and creates a PR_SHAs.txt file. This script has a limit of 500 commits but the
# github API used has a lower limit of 250 commits.

# It does not rely on git merge-bases which basically don't work with
# shallow clones:
# https://github.com/thesofproject/linux/issues/2556

# Design goals:
#
# - Keep the code short and as simple as possible. No one is interested
# in maintaining this sort of script.
#
# - Fast and accurate for small Pull Requests
#
# - For large Pull Requests _with merges_ the only objective is to
# complete in a reasonable time; say less than 10 minutes. It's very
# unlikely will look at 250 checkpatch results and time optimizations
# should not make this script more complex.


# Sample usage:
# $0 thesoftproject/linux 2772
main()
{
local gh_project="$1"
local pr_number="$2"

printf '%s: fetching PR %d for project %s\n' "$0" "$pr_number" "$gh_project"

# As of March 2021, Github's documented limit is 250 commits
# Let's have another cap a 500.
# https://docs.github.com/en/rest/reference/pulls#list-commits-on-a-pull-request
local pagelen PRlen=0
for i in 1 2 3 4 5; do
curl -H 'Accept: application/vnd.github.v3+json' \
"https://api.github.com/repos/$gh_project/pulls/$pr_number/commits?per_page=100&page=$i" \
> commits_"$i".json
pagelen=$(jq length < commits_$i.json)
if [ "$pagelen" -eq 0 ]; then
break
fi
PRlen=$((PRlen + pagelen))
done

printf 'Found %d commits, SHA1 list is in PR_SHAs.txt\n' "$PRlen"

# 'cut' removes double quotes
cat commits_?.json |
jq '.[] | .sha' |
cut -b2-41 > PR_SHAs.txt

# PRlen+1 gets us the merge base for simple, linear histories. For
# pull requests with merges, depth=PRLEN goes already much further
# than needed and +1 makes little difference. It's not clear when
# and for what sort of PRs git fetching individual commits would be
# faster so keep a single and simple fetch for now.

set -x # this command may take a while so show it
git fetch --depth "$((PRlen+1))" "http://github.com/$gh_project" "pull/$pr_number/head"

}

main "$@"

0 comments on commit ea76703

Please sign in to comment.