From c5bfcb7d2c018c3518010508403778ca2a881347 Mon Sep 17 00:00:00 2001
From: Marian Steinbach <marian@giantswarm.io>
Date: Mon, 16 Dec 2024 11:36:40 +0100
Subject: [PATCH] Check content dir for for file and folder name validity
 (#2427)

---
 .github/workflows/validate.yaml | 40 +++++++++++++++++++++++++++++++++
 1 file changed, 40 insertions(+)

diff --git a/.github/workflows/validate.yaml b/.github/workflows/validate.yaml
index 8c3e8bd1d4..5de5e58a48 100644
--- a/.github/workflows/validate.yaml
+++ b/.github/workflows/validate.yaml
@@ -77,6 +77,46 @@ jobs:
             exit 1
           fi
 
+      - name: Check file and folder names in content
+        shell: python {0}
+        run: |
+          # Walk content folder, check file names to use only allowed characters
+          from pathlib import Path
+          import re
+          import os
+          import sys
+
+          allowed_chars = re.compile(r'^[a-z0-9-\.+]+$')
+          allowed_chars_static = re.compile(r'^[a-z0-9-\._]+$') # images etc. may use the underscore
+
+          content_dir = 'src/content'
+          errors = []
+          for root, dirs, files in os.walk(content_dir):
+            for file in files:
+              full_path = Path(root, file)
+              if full_path.name == '_index.md':
+                continue
+              if full_path.name == '_template.md.tpl':
+                continue
+              
+              if full_path.suffix.lower() == '.md':
+                if not allowed_chars.match(full_path.stem):
+                  errors.append(f"- FILE {full_path}")
+              else:
+                # static files
+                if not allowed_chars_static.match(full_path.name):
+                  errors.append(f"- FILE {full_path}")
+
+            for dir in dirs:
+              if not allowed_chars.match(dir):
+                errors.append(f"- DIR {Path(root, dir)}")
+          
+          if len(errors) > 0:
+            sys.stderr.write("The following file/folder names use invalid characters. Only lowercase letters, digits, hyphens and period are allowed.\n")
+            for error in errors:
+              sys.stderr.write(error + "\n")
+            sys.exit(1)
+
       - name: Check for moved or deleted files
         run: |
           git --no-pager diff --name-status --diff-filter=RD "refs/heads/${GITHUB_BASE_REF}" -- . | tee files.txt