From 4fa8d18621b123cdb6182150c5895d35c04fa5ab Mon Sep 17 00:00:00 2001
From: Philip Sargent <philip.sargent@klebos.com>
Date: Thu, 23 Jun 2022 19:01:25 +0300
Subject: [PATCH] utf8 encoding checks when reading files in the view

---
 core/views/expo.py | 24 ++++++++++++++++++++----
 1 file changed, 20 insertions(+), 4 deletions(-)

diff --git a/core/views/expo.py b/core/views/expo.py
index 66dffd0..6256aab 100644
--- a/core/views/expo.py
+++ b/core/views/expo.py
@@ -1,6 +1,9 @@
 import os
 import re
 import subprocess
+
+from sys import getfilesystemencoding as sys_getfilesystemencoding
+
 from pathlib import Path
 from urllib.parse import urljoin, unquote as urlunquote
 from urllib.request import urlopen
@@ -125,7 +128,12 @@ def expowebpage(request, expowebpath, path):
         # Should not get here if the path has suffix "_edit"
         print(f' - 404 error in expowebpage() {path}')
         return render(request, 'pagenotfound.html', {'path': path}, status="404")
-        
+    
+    # print(f' - {sys_getfilesystemencoding()=}')
+    if (sys_getfilesystemencoding() != "utf-8"):
+        return HttpResponse(default_head +  '<h3>UTF-8 Parsing Failure:<br>Default file encoding on this Troggle installation is not UTF-8:<br>failure detected in expowebpage in views.expo.py</h3> Please Please reconfigure Debian/Apache/Django to fix this, i.e. contact Wookey. </body' )
+
+    # This next bit can be drastically simplified now that we know that the system encoding actually is utf-8
     try:
         with open(expowebpath / path, "r", encoding='utf-8') as o:   
             html = o.read()
@@ -220,9 +228,13 @@ def expopage(request, path):
         # do not redirect to a file path without the slash as we may get in a loop. Let the user fix it:
         return render(request, 'dirnotfound.html', {'path': path, 'subpath': path[0:-1]})
     
-    # So it must be a file in  /expoweb/ but not .htm or .html probably an image
+    # So it must be a file in  /expoweb/ but not .htm or .html probably an image, maybe a txt file
     filetobeopened = expowebpath / path
-    
+
+    # print(f' - {sys_getfilesystemencoding()=}')
+    if (sys_getfilesystemencoding() != "utf-8"):
+        return HttpResponse(default_head +  '<h3>UTF-8 Parsing Failure:<br>Default file encoding on this Troggle installation is not UTF-8:<br>failure detected in expowebpage in views.expo.py</h3> Please Please reconfigure Debian/Apache/Django to fix this, i.e. contact Wookey. </body' )
+
     try:
         content = open(filetobeopened, "rb")
         content_type=getmimetype(path)
@@ -278,6 +290,10 @@ def editexpopage(request, path):
     except Cave.DoesNotExist:
         pass
 
+    print(f' - {sys_getfilesystemencoding()=}')
+    if (sys_getfilesystemencoding() != "utf-8"):
+        return HttpResponse(default_head +  '<h3>UTF-8 Parsing Failure:<br>Default file encoding on this Troggle installation is not UTF-8:<br>failure detected in expowebpage in views.expo.py</h3> Please Please reconfigure Debian/Apache/Django to fix this, i.e. contact Wookey. </body' )
+
     try:
         filepath = Path(settings.EXPOWEB) / path
         o = open(filepath, "r", encoding="utf8")
@@ -361,7 +377,7 @@ def write_and_commit(filepath, content):
             cwd = filepath.parent
             filename = filepath.name
             git = settings.GIT
-            # GIT see also core/models/cave.py writetrogglefile()
+            # GIT see also core/models/caves.py writetrogglefile()
             # GIT see also core/views/uploads.py dwgupload()
             
             try: