Reza Lotun wrote:

Hi Igor,

...

I need to download a videofile, a thumbnail to it, process that videofile, and after the processing is finished, process the thumbnail with the result of processing the video.

[SNIP]

url = 'http://..../video.flv' filename = 'video.flv' video_deffered = downloadPage(url, filename) video_deffered.addCallback(process_video, filename) => process_video returns video_id

url = 'http://..../thumbnail.jpg' filename = 'thumbnail.jpg' image_deffered = downloadPage(url, filename)

image_deffered.chainDeffered(video_deffered) image_deffered.addCallback(process_image, filename) =< video_id needed here

As far as I understand you want to: 1. Get the videofile - represented by video_deferred 2. Process the videofile after it's downloaded (the function process_video) 3. Get the thumbnail in parallel - image_deferred 4. Do a final processing *after* step 2 and step 3 (process_image)

What you want to do is fire off video_deferred and image_deferred as you do immediately, but put the two deferreds in a DeferredList. What this allows you to do is attach a callback that will only fire when *both* jobs have been completed. The addCallback on that deferred list will be process_image.

To run jobs in threads, please take a look a twisted.internet.threads.deferToThread.

Hope that helps.

Reza

Thank you, Reza, for a quick and helpful answer! That worked perfectly. And to make my existing callbacks non-blocking conveniently, I wrote a decorator for deferToThread, looks like this: def defer_to_thread_decorator(func): def wrap(*args, **kwargs): return deferToThread(func, *args, **kwargs) return wrap