patchwork: loses Acks if they arrive before the patchmail

[Jeremy Kerr]

Hi Peter,

> Well, you wouldn't have to keep them indefinitely, and you only
> need to keep emails which have references/in-reply-to pointing to
> an email you haven't seen yet, so that's not very much data.
> 
> If you can't trust patchwork to actually retain all the relevant
> traffic about a patch, I think its utility is significantly
> reduced.

True. I'll have a go at implementing this and see where it gets us.

However, we'll never get 100% retention of patch activity; there's too many 
variations in possible inputs for that :)

Cheers,


Jeremy