Move to COMPRESS_OFFLINE for multiple replicas

This generates a compress-manifest.json which is used to
determine which file hashes are in use as apparently they
are not consistent.

Author: frankwiles

docker-compose.yml

@@ -15,21 +15,21 @@ services:
       context: .
       target: release
     tty: true
-    working_dir: /nuremberg
-    command: python manage.py runserver 0.0.0.0:8000
+    working_dir: /code
+#    command: python manage.py runserver 0.0.0.0:8000
     ports:
       - "127.0.0.1:8000:8000"
     volumes:
       # BIND MOUNTS
-      - ./web:/nuremberg
+      - ./web:/code
       - ./media:/media
     environment:
       # let Django load Docker-specific settings conditionally
       - DOCKERIZED=True
-      - DEBUG=True
+      - DEBUG=False
       - SECRET_KEY=top-secret-1234
-      - LOCAL_DEVELOPMENT=True
-      - DJANGO_VITE_DEV_MODE=True
+      - LOCAL_DEVELOPMENT=False
+      - DJANGO_VITE_DEV_MODE=False
     depends_on:
       - solr
     stop_signal: SIGKILL

web/nuremberg/settings.py

@@ -145,7 +145,9 @@
 #
 COMPRESS_PRECOMPILERS = (('text/less', 'lessc {infile} {outfile}'),)
 
-COMPRESS_STORAGE = 'compressor.storage.GzipCompressorFileStorage'
+COMPRESS_STORAGE = 'compressor.storage.OfflineManifestFileStorage'
+COMPRESS_OFFLINE = True
+COMPRESS_OFFLINE_MANIFEST = 'compress-manifest.json'
 
 # whitenoise settings
 # https://warehouse.python.org/project/whitenoise/
@@ -244,6 +246,7 @@
     SECRET_KEY = 'supersecret'
     DEBUG = True
     COMPRESS_ENABLED = True
+    COMPRESS_FORCE = False
 
     CACHES = {
         'default': {'BACKEND': 'django.core.cache.backends.dummy.DummyCache'}