defrex / django-after-response Goto Github PK

The function_queue variable is shared between all the threads on a process. If thread 1 adds something to the queue and then does some more processing and thread 2 comes along and finishes, even if it didn't add anything to the queue, it will process thread 1's function while thread 1 is still running.

You should replace function queue with a method that will pull the list from the threading.local and create it if it does not exist. Do not just put the variable in threading.local and leave it as a global variable. Then you'll just be sharing the thread local with all the other processes.

It should be something like this:

import threading
threadlocal = threading.local()    

def function_queue():
  queue = getattr(threadlocal, 'function_queue', None)
  if v is None:
    queue = []
    setattr(threadlocal, 'function_queue', queue)
  return queue

Would be cool if the decorator could override the async option, so we can optimise for both safe and unsafe tasks.

I don't know how exactly this error occurs, but in my case i get the following error:

No handlers could be found for logger "after_response.signals"

More disturbing however is the fact that the database update I placed within the decorated function doesn't happen either.
I understand the function isn't getting executed but I have no idea why.
Any idea why this is happening?

I'm running django version 1.11.13 on Python 2.7.

$ pip install django-after-response==0.2.2
Collecting django-after-response==0.2.2
  Could not find a version that satisfies the requirement django-after-response==0.2.2 (from versions: 0.1.0, 0.1.1, 0.2.0, 0.2.1)
  No distributions matching the version for django-after-response==0.2.2

I had to install from git to get 0.2.2:

$ pip install git+git://github.com/defrex/django-after-response@98e00c766b5160b529a4ef810412eb9d2e6fc9a3

For me this is magic. I have no clue how I can execute code after the http response was sent to the client with django.

Could you please elaborate the README and explain how the magic works.

This would be very kind.

Regards,
Thomas Güttler

Hello,

Starting at version 0.2.0 I have many test fails because my tests expect synchronous execution of after-response callbacks.

I could add timers, but I think synchronous tests are more reliable.

Celery provides a CELERY_ALWAYS_EAGER option to disable asynchronous calls. Could you add something similar ?

Thank you,
Julien

I need to import a model into the after response, but I can not.

import after_response

@after_response.enable
def example():
from automacao.models import Automacao
Automacao.objects.count()
...

{ProgrammingError}relation "automacao" does not exist
LINE 1: SELECT COUNT(*) FROM "automacao" WHERE "automacao"."is_delet...

Hey there

Is there any way this could be adapted to create a progress bar?

for example, if one were to have a large function running, after_response would send a value out for jquery or ajax call to ingest at every iteration of a loop?

Cheers

I haven't been able to get any threading.Thread to work with uwsgi, including when using the django-after-response decorators. Has anyone experienced the same problem? I've actually written a more detailed explanation on stackoverflow: http://stackoverflow.com/questions/37930268/differences-between-uwsgi-and-gunicorn-when-running-async-calls

Thanks,

Because the setting is loaded to local variable, these test decodators do not work. I do not any problem to load the setting from the settings every time?

Now I have to import those and them override it manyally like: signals.AFTER_RESPONSE_RUN_ASYNC = False