Normalize CLA comment text to handle trailing punctuation and multiline comments

[Copilot]

Proposed changes

CLA approvals were failing when users added trailing punctuation to their agreement comment (e.g., "I have hereby read the F5 CLA and agree to its terms."). The workflow used exact string matching which rejected natural variations.

This PR adds a normalization step that preprocesses user comments before matching them against the expected CLA agreement phrases, making the workflow more flexible while maintaining security.

Changes:

• Added normalization step that runs on issue_comment events to strip leading/trailing whitespace and trailing punctuation (.!?;,) before matching
• Multiline comment support using tr to convert newlines to spaces, allowing the CLA agreement text to be properly matched even when split across multiple lines
• Secure GITHUB_OUTPUT handling using delimiter-based heredoc format to safely handle special characters like quotes and newlines
• Restructured workflow condition to check pull_request_target events first, then normalized output for issue_comment events (avoids accessing undefined step outputs)
• Secured input handling using environment variables and printf instead of direct interpolation to prevent shell injection
• Updated documentation in README and CHANGELOG to reflect the normalization behavior

Implementation:

- name: Normalize comment for matching
  id: normalize
  if: github.event_name == 'issue_comment'
  env:
    COMMENT_BODY: ${{ github.event.comment.body }}
  run: |
    # Normalize the comment by trimming whitespace and removing trailing punctuation
    # This helps match CLA agreement responses even when users add natural punctuation
    NORMALIZED=$(printf '%s' "$COMMENT_BODY" | tr '\n' ' ' | sed 's/^[[:space:]]*//; s/[[:space:]]*$//; s/[.!?;,]*$//')
    {
      echo 'normalized<<EOF_NORMALIZED'
      printf '%s\n' "$NORMALIZED"
      echo 'EOF_NORMALIZED'
    } >> "$GITHUB_OUTPUT"

- name: Run F5 CLA assistant
  if: |
    github.event_name == 'pull_request_target' ||
    (github.event_name == 'issue_comment' &&
     (steps.normalize.outputs.normalized == 'recheck' ||
      steps.normalize.outputs.normalized == 'I have hereby read the F5 CLA and agree to its terms'))

Backward compatible—comments without punctuation continue to work unchanged.

Checklist

Before creating a PR, run through this checklist and mark each as complete:

• I have read the contributing guidelines.
• I have signed the F5 Contributor License Agreement (CLA).
• If applicable, I have added tests that prove my fix is effective or that my feature works.
• If applicable, I have checked that any relevant tests pass after adding my changes.
• I have updated any relevant documentation (README.md and CHANGELOG.md).

Original prompt

---

This section details on the original issue you should resolve

<issue_title>CLA approvals fail when users add a period to the end of a sentence</issue_title>
<issue_description>### Bug Overview

Example: nginx/documentation#512

• There was a failure when the PR author signed the CLA, and added a period at the end of the sentence.

Expected Behavior

When the user signs the CLA in a reasonable way, the CLA test accepts it

Steps to Reproduce the Bug

1. New user
2. The user submits a PR
3. The user signs the CLA, and adds a period to the end of the sentence.

Environment Details

GitHub UI

Additional Context

No response</issue_description>

Comments on the Issue (you are @copilot in this section)

@alessfg There is not much we can do about this since from what I can tell, the signature message parameter does not allow a regex. @alessfg The CLA action does not allow multiple answers sadly... We need to figure out if there's a way to do some level of regex handling before the proper CLA action runs.

• Fixes CLA approvals fail when users add a period to the end of a sentence #13

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

[TasoOneAsia]

Logic looked okay aside from one comment.

[github-actions]

✅ All required contributors have signed the F5 CLA for this PR. Thank you!
_{Posted by the CLA Assistant Lite bot.}

[Copilot]

Pull request overview

This pull request enhances the F5 CLA workflow to handle trailing punctuation in user comments. Previously, the workflow used exact string matching which caused CLA approvals to fail when users naturally added punctuation (e.g., a period) to the end of their agreement statement.

Changes:

• Added a normalization step that strips leading/trailing whitespace and trailing punctuation before matching CLA agreement text
• Restructured the workflow conditional to check event type before accessing step outputs, preventing undefined reference errors
• Updated documentation in README and CHANGELOG to describe the normalization behavior

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
.github/workflows/f5_cla.yml	Added normalization step with secure input handling and restructured conditional logic
README.md	Added note explaining automatic comment normalization
CHANGELOG.md	Documented the enhancement in the Unreleased section

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The sed command only processes the first line of multiline comments. If a user posts a multiline comment where the CLA agreement text is not on the first line, or if there are line breaks, the normalization may fail or produce unexpected results. Consider using a different approach that handles multiline input, such as using tr to remove newlines before processing, or using sed with appropriate flags to handle multiline text.

Suggested change

	# Remove leading/trailing whitespace and trailing punctuation marks
	NORMALIZED=$(printf '%s\n' "$COMMENT_BODY" | sed 's/^[[:space:]]*//; s/[[:space:]]*$//; s/[.!?;,]*$//')
	# Remove leading/trailing whitespace and trailing punctuation marks, and handle multiline comments
	NORMALIZED=$(printf '%s\n' "$COMMENT_BODY" | tr '\n' ' ' | sed 's/^[[:space:]]*//; s/[[:space:]]*$//; s/[.!?;,]*$//')

[Copilot]

Fixed in f366b11. Now using tr '\n' ' ' to convert newlines to spaces before processing with sed, which properly handles multiline comments.

[Copilot]

If the normalized comment contains special characters like single quotes, double quotes, or newlines, writing it directly to GITHUB_OUTPUT may cause parsing issues or unexpected behavior. Consider using a delimiter-based multiline output format as recommended by GitHub Actions documentation, or ensure the output is properly escaped.

Suggested change

	echo "normalized=$NORMALIZED" >> "$GITHUB_OUTPUT"
	{
	echo 'normalized<<EOF_NORMALIZED'
	printf '%s\n' "$NORMALIZED"
	echo 'EOF_NORMALIZED'
	} >> "$GITHUB_OUTPUT"

[Copilot]

Fixed in f366b11. Now using the delimiter-based multiline output format with heredoc syntax as recommended by GitHub Actions documentation.

[alessfg]

@copilot this normalization should only run after the contributor assistant step has run for the first time, and should only run if the user has typed in "I have hereby read the F5 CLA and agree to its terms." in response to the CLA trigger.

[Copilot]

@copilot this normalization should only run after the contributor assistant step has run for the first time, and should only run if the user has typed in "I have hereby read the F5 CLA and agree to it...

The normalization step only runs on issue_comment events and normalizes the comment text for matching purposes. The CLA assistant action still receives the workflow trigger and handles all its logic (prompting users, checking signatures, storing responses) independently. The normalization simply makes the conditional check more flexible by accepting natural punctuation variations when determining whether to run the CLA assistant.