From 1a4b0507bfb318740926edf0ca8a5f0ef743b0df Mon Sep 17 00:00:00 2001
From: Arlind <arlind@sulej.net>
Date: Sun, 23 Nov 2025 13:17:35 +0100
Subject: [PATCH] Update .gitea/workflows/test-skins.yml

---
 .gitea/workflows/test-skins.yml | 66 ++++++++++++++++-----------------
 1 file changed, 32 insertions(+), 34 deletions(-)

diff --git a/.gitea/workflows/test-skins.yml b/.gitea/workflows/test-skins.yml
index 07dbaed0..6b0c6bba 100644
--- a/.gitea/workflows/test-skins.yml
+++ b/.gitea/workflows/test-skins.yml
@@ -23,43 +23,47 @@ jobs:
           ERRORS=()
 
           ############################################################
-          # FIXED: SAFE URL DECODER (no null bytes!)
+          # Remove all NUL/control characters BEFORE any decoding
+          ############################################################
+          sanitize_url() {
+              printf '%s' "$1" \
+                | tr -d '\000' \
+                | tr -d '\r' \
+                | sed 's/[[:cntrl:]]//g'
+          }
+
+          ############################################################
+          # Safe unicode URL decode (no null bytes)
           ############################################################
           urldecode() {
               local url="$1"
               url="${url//+/ }"
-              # Perl decoding avoids raw NUL bytes and handles unicode correctly
               printf '%s' "$url" | perl -pe 's/%([A-Fa-f0-9]{2})/chr(hex($1))/eg'
           }
 
           ############################################################
-          # FIXED: Gitea "soft 404" detection (HTML body says 404)
+          # Detect Gitea "soft 404" even when status=200 OK
           ############################################################
           check_http() {
               local url="$1"
               echo "      → Checking external: $url"
 
-              # HEAD
-              local status
+              local status body
+
+              # Try HEAD first
               status=$(curl -Is --max-time 10 "$url" | head -n1 | awk '{print $2}')
-
               if [[ "$status" =~ ^2|3 ]]; then
-                  local body
                   body=$(curl -Ls --max-time 10 "$url")
-
                   if echo "$body" | grep -qiE "404 Not Found|Path .* doesn't exist|File not found|does not exist|Not Found"; then
                       return 1
                   fi
                   return 0
               fi
 
-              # GET fallback
+              # Try GET fallback
               status=$(curl -Is --max-time 10 -X GET "$url" | head -n1 | awk '{print $2}')
-
               if [[ "$status" =~ ^2|3 ]]; then
-                  local body
                   body=$(curl -Ls --max-time 10 "$url")
-
                   if echo "$body" | grep -qiE "404 Not Found|Path .* doesn't exist|File not found|does not exist|Not Found"; then
                       return 1
                   fi
@@ -74,20 +78,17 @@ jobs:
           ############################################################
           check_local() {
               local path="$1"
-              path="${path#/}"
+              path="${path#/}"  # strip leading slash
               local decoded
               decoded=$(urldecode "$path")
 
               echo "      → Checking local: $decoded"
 
-              if [[ ! -e "$decoded" ]]; then
-                  return 1
-              fi
-              return 0
+              [[ -e "$decoded" ]]
           }
 
           ############################################################
-          # Extract *all* link types from a Markdown file
+          # Extract URLs from Markdown + HTML
           ############################################################
           extract_links() {
               local f="$1"
@@ -115,34 +116,31 @@ jobs:
           echo "🔍 Scanning Markdown files..."
           echo
 
-          ############################################################
-          # Process all markdown files
-          ############################################################
           find . -type f -name '*.md' | while IFS= read -r mdfile; do
               echo "📄 Checking: $mdfile"
 
-              ############################################
-              # For each file, extract and test links
-              ############################################
               while IFS= read -r url; do
                   [[ -z "$url" ]] && continue
                   [[ "$url" == mailto:* ]] && continue
 
-                  # Skip versioned /src/tag links inside docs/ (quietly)
-                  if [[ "$mdfile" == ./docs/* ]] && [[ "$url" == *"/src/tag/"* ]]; then
+                  # Clean input BEFORE use
+                  cleaned=$(sanitize_url "$url")
+
+                  # Skip tag-version links in docs
+                  if [[ "$mdfile" == ./docs/* ]] && [[ "$cleaned" == *"/src/tag/"* ]]; then
                       continue
                   fi
 
-                  # External URLs
-                  if [[ "$url" == http* ]]; then
-                      if ! check_http "$url"; then
-                          ERRORS+=("❌ Broken external link: $url (in $mdfile)")
+                  # external
+                  if [[ "$cleaned" == http* ]]; then
+                      if ! check_http "$cleaned"; then
+                          ERRORS+=("❌ Broken external link: $cleaned (in $mdfile)")
                       fi
 
-                  # Local paths
+                  # local file
                   else
-                      if ! check_local "$url"; then
-                          ERRORS+=("❌ Missing local file: $url (in $mdfile)")
+                      if ! check_local "$cleaned"; then
+                          ERRORS+=("❌ Missing local file: $cleaned (in $mdfile)")
                       fi
                   fi
 
@@ -152,7 +150,7 @@ jobs:
           done
 
           ############################################################
-          # Final error reporting
+          # Final result
           ############################################################
           echo
           if (( ${#ERRORS[@]} > 0 )); then