Skip to content

Conversation

@foca
Copy link
Collaborator

@foca foca commented Aug 7, 2015

Instead of depending on Celluloid, which is complex and way too much for what we actually nead, this sets up a ConnectionPool that checks out nothing (i.e. the connections are just true), so that we can abuse it as a thread pool.

This would always make the workers threaded (set DISC_CONCURRENCY=1 if you want a single thread...).

Having a single code path is good, though. It means it's simpler to test features / bugfixes as we don't need to run every test by duplicate on threaded/nonthreaded mode.

Oh, and speaking about tests, did I mention this is just a WIP? :P

cc @pote

Instead of depending on Celluloid, this would _always_ make the workers
threaded (set DISC_CONCURRENCY=1 if you want a single thread...)

Having a single code path is good, though. It means it's simpler to
test features / bugfixes as we don't need to run every test by duplicte
on threaded/nonthreaded mode.

Oh, and speaking about tests, did I mention this is just a WIP? :P
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Not entirely sure why Disc.disque_timeout had a default different from DISQUE_TIMEOUT. Was this on purpose?

Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Probably not.

@pote
Copy link
Owner

pote commented Aug 10, 2015

I think dropping Celluloid and having a single code path are good things, however I'd like to move to something that won't break the non-threaded mode at least, what are your thoughts on the stability and safety of using ConnectionPool? Should we keep this as a WIP for the time being until we cook up a more definitive solution?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants