Drop multiprocessing in rendervariations command

codingjoe · codingjoe · commit d4bb2f20124f · 2021-03-04T18:43:53.000+01:00
It is profered to use asynchronous renders andyways and to do
image rendering on optimized works. Therefore, using multiprocessing
to simply queue tasks adds more errors than it does good.
diff --git a/README.md b/README.md
@@ -12,7 +12,6 @@ Django Field that implement the following features:
 * Access thumbnails on model level, no template tags required
 * Preserves original image
 * Asynchronous rendering (Celery & Co)
-* Multi threading and processing for optimum performance
 * Restrict accepted image dimensions
 * Rename files to a standardized name (using a callable upload_to)
 
@@ -195,12 +194,3 @@ The `replace` option will replace all existing files.
 The `ignore-missing` option will suspend missing source file errors and keep
 rendering variations for other files. Othervise command will stop on first
 missing file.
-
-### Multi processing
-Since version 2 stdImage supports multiprocessing.
-Every image is rendered in separate process.
-It not only increased performance but the garbage collection
-and therefore the huge memory footprint from previous versions.
-
-**Note:** PyPy seems to have some problems regarding multiprocessing,
-for that matter all multiprocessing is disabled in PyPy.
diff --git a/stdimage/management/commands/rendervariations.py b/stdimage/management/commands/rendervariations.py
@@ -1,6 +1,3 @@
-from concurrent.futures import ProcessPoolExecutor
-from multiprocessing import cpu_count
-
 import progressbar
 from django.apps import apps
 from django.core.files.storage import get_storage_class
@@ -75,14 +72,12 @@ def render(field, images, count, replace, ignore_missing, do_render):
         )
         with progressbar.ProgressBar(max_value=count, widgets=(
             progressbar.RotatingMarker(),
-            ' | CPUs: {}'.format(cpu_count()),
             ' | ', progressbar.AdaptiveETA(),
             ' | ', progressbar.Percentage(),
             ' ', progressbar.Bar(),
         )) as bar:
-            with ProcessPoolExecutor() as executor:
-                for _ in executor.map(render_field_variations, kwargs_list):
-                    bar += 1
+            for _ in map(render_field_variations, kwargs_list):
+                bar += 1
 
 
 def render_field_variations(kwargs):
