Ensure events send out after being forked

[travisjeffery]

This ensures that the threads that the event requests are on run after their process has been forked.

For people using forking servers like unicorn and passenger, they no longer have to setup the sdk again with:

• https://segment.io/docs/libraries/ruby/#unicorn
• https://segment.io/docs/libraries/ruby/#passenger

[calvinfo]

I think we still need this so the thread gets properly GC'd: #27

[calvinfo]

niiiiiiice! :D this looks really sweet.

only concern that I have is if they do something like:

analytics.init('secret')
analytics.track(...)

== FORK ==

analytics.track(...)

since the thread might still be "alive" but not actually synced with the right "queue" properly when the memory is copied.

if that case works, lgtm!

[jtippett]

So to clarify - does this mean unicorn, et al users no longer have to re-init analytics-ruby after forking?

[toothrot]

This is pretty amazing. You should blog about this solution. I think it's a great pattern for lots of background threading.