Adding link checker script + GitHub CI

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

Author: aireilly

.github/workflows/check-links.yml

@@ -0,0 +1,46 @@
+name: Check Links
+
+on:
+  pull_request:
+    types: [opened, reopened]
+
+jobs:
+  check-links-global:
+    name: Check All Links
+    runs-on: ubuntu-latest
+    container:
+      image: asciidoctor/docker-asciidoctor
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Make script executable
+        run: chmod +x scripts/check-links.sh
+
+      - name: Run link checker (global)
+        run: scripts/check-links.sh --global
+  check-links-modified:
+    name: Check Modified Links
+    runs-on: ubuntu-latest
+    container:
+      image: asciidoctor/docker-asciidoctor
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0  # required for git diff to work with branches
+
+      - name: Make script executable
+        run: chmod +x scripts/check-links.sh
+
+      - name: Run link checker on modified files
+        shell: bash
+        run: |
+          MODIFIED_FILES=$(git diff --name-only origin/${{ github.base_ref }}...HEAD)
+          CHECK_FILES=()
+
+          while IFS= read -r file; do
+            [[ "$file" =~ ^assemblies/[^/]+\.adoc$ ]] && CHECK_FILES+=("$file")
+          done <<< "$MODIFIED_FILES"
+
+          for f in "${CHECK_FILES[@]}"; do
+            ./scripts/check-links.sh "$f"
+          done

scripts/check-links.sh

@@ -0,0 +1,86 @@
+#!/bin/bash
+#
+# Checks for 404 links using Asciidoctor and curl
+# Checks all master.adoc files by using --global
+
+usage() {
+    echo "Usage: $0 [--global] [<adoc_file>]"
+    exit 1
+}
+
+# Parse arguments
+GLOBAL_MODE=0
+INPUT_FILE=""
+
+if [ "$1" == "--global" ]; then
+    GLOBAL_MODE=1
+elif [ "$#" -eq 1 ]; then
+    INPUT_FILE="$1"
+else
+    usage
+fi
+
+# Create temp file for flagging broken links
+TMP_FILE=$(mktemp)
+echo "0" > "$TMP_FILE"
+
+# Ignore patterns
+IGNORE_PATTERNS=(
+    ".*docs\.google\.com.*"
+    ".*google\.com.*"
+    ".*issues\.redhat\.com.*"
+    ".*0\.0\.0\.0.*"
+    ".*localhost.*"
+    ".*registry\.redhat\.io.*"
+    ".*example\.org.*"
+    ".*github.com/example/myrepo\.git"
+)
+
+PATTERNS_DECL=$(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" =~ $PATTERN ]]; then
+            exit 0
+        fi
+    done
+
+    STATUS=$(curl -Ls -o /dev/null -w "%{http_code}" --max-time 5 --connect-timeout 2 "$URL")
+
+    if [[ "$STATUS" == "404" ]]; then
+        echo -e "Invalid URL (HTTP status $STATUS): \n\033[31m$URL\033[0m"
+        echo "1" > "$TMP_FILE"
+    fi
+}
+
+# Export variables and functions for xargs subshells
+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 '(http|https)://[a-zA-Z0-9./?=%_-]*' | \
+        sort -u | \
+        xargs -P 10 -n 1 bash -c "$PATTERNS_DECL; check_url \"\$0\""
+}
+
+if [ "$GLOBAL_MODE" -eq 1 ]; then
+    while IFS= read -r file; do
+        run_url_checks "$file"
+    done < <(find assemblies -maxdepth 1 -type f -name "*.adoc")
+else
+    run_url_checks "$INPUT_FILE"
+fi
+
+if [ "$(cat "$TMP_FILE")" -eq 1 ]; then
+    echo "Errors found"
+    exit 1
+fi