Adding link checker script + GitHub CI

Signed-off-by: Aidan Reilly <[email protected]>

Author: aireilly

.github/workflows/check-links.yml

@@ -0,0 +1,38 @@
+name: Check links in AsciiDoc
+
+on: [push, pull_request]
+
+jobs:
+  check-links:
+    name: Check links in modified files
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Install Asciidoctor
+        run: |
+          sudo apt-get update
+          sudo apt-get install -y asciidoctor
+
+      - name: Make script executable
+        run: chmod +x scripts/check-links.sh
+
+      - name: Check links in modified files
+        shell: bash
+        run: |
+          MODIFIED_FILES=$(git diff --name-only origin/main...HEAD --diff-filter=d -- "*.adoc")
+          CHECK_FILES=()
+
+          while IFS= read -r file; do
+            [[ "$file" =~ ^assemblies/[^/]+\.adoc$ ]] && CHECK_FILES+=("$file")
+          done <<< "$MODIFIED_FILES"
+
+          if [ ${#CHECK_FILES[@]} -eq 0 ]; then
+            echo "No modified .adoc files found in assemblies/. Skipping link check."
+            exit 0
+          fi
+
+          for f in "${CHECK_FILES[@]}"; do
+            ./scripts/check-links.sh "$f"
+          done

scripts/check-links.sh

@@ -0,0 +1,79 @@
+#!/bin/bash
+#
+# Checks for 404 links using Asciidoctor and curl
+
+usage() {
+    echo "Usage: $0 [<adoc_file>]"
+    exit 1
+}
+
+# Parse arguments
+INPUT_FILE=""
+
+# Check dependencies
+if ! asciidoctor -v >/dev/null 2>&1; then
+    echo "Error: Asciidoctor is not installed" >&2
+    exit 1
+fi
+
+if [ $# -eq 0 ]; then
+    usage
+fi
+
+INPUT_FILE="$1"
+
+# Create temp file for flagging broken links
+TMP_FILE=$(mktemp)
+trap 'rm -f "$TMP_FILE"' EXIT
+echo "0" > "$TMP_FILE"
+
+# Load ignore patterns from external file
+IGNORE_FILE="$(dirname "$0")/links.ignore"
+
+if [ ! -f "$IGNORE_FILE" ]; then
+    echo "Error: Missing ignore patterns file: $IGNORE_FILE" >&2
+    exit 1
+fi
+
+mapfile -t IGNORE_PATTERNS < "$IGNORE_FILE"
+PATTERNS_DECL=$(declare -p IGNORE_PATTERNS)
+
+check_url() {
+    local URL=$1
+    eval "$PATTERNS_DECL"
+
+    # Strip trailing punctuation
+    URL=${URL%[.,;:?!\]\)]}
+
+    for PATTERN in "${IGNORE_PATTERNS[@]}"; do
+        if [[ "$URL" =~ $PATTERN ]]; then
+            return 0
+        fi
+    done
+
+    STATUS=$(curl -Ls -o /dev/null -w "%{http_code}" --max-time 5 --connect-timeout 2 "$URL")
+
+    if [[ "$STATUS" != "000" && "$STATUS" != "403" && ! "$STATUS" =~ ^(2|3)[0-9]{2}$ ]]; then
+        echo -e "Invalid URL (HTTP status $STATUS): \n\033[31m$URL\033[0m"
+        echo "1" > "$TMP_FILE"
+    fi
+}
+
+export TMP_FILE
+export -f check_url
+
+run_url_checks() {
+    local FILE="$1"
+    echo -e "\033[32mChecking: $FILE\033[0m"
+    asciidoctor "$FILE" -o - | \
+        grep -Eo 'https?://[^")>\]\s]+' | \
+        sort -u | \
+        xargs -P 10 -n 1 bash -c "$PATTERNS_DECL; check_url \"\$0\""
+}
+
+run_url_checks "$INPUT_FILE"
+
+if [ "$(cat "$TMP_FILE")" -eq 1 ]; then
+    echo "Errors found"
+    exit 1
+fi

scripts/links.ignore

@@ -0,0 +1,15 @@
+# Add ignore link regexes one per line
+.*docs\.google\.com.*
+.*google\.com.*
+.*issues\.redhat\.com.*
+.*0\.0\.0\.0.*
+.*localhost.*
+.*registry\.redhat\.io.*
+.*example\.org.*
+.*github.com/example/myrepo\.git
+.*fonts\.googleapis\.com.*
+.*mixtral-my-project.apps\.my-cluster\.com.*
+.*openshiftapps\.com.*
+.*minio-cluster\.local.*
+.*codeflare-operator-webhook-service\.redhat-ods-applications\.svc
+.*example.com.*