Assorted history questions

[Alex Kiernan]

Russ Allbery <rra at stanford.edu> writes:

> Alex Kiernan <alexk at demon.net> writes:
> 
> > I'm just going over what I've implemented and I think I've got it wrong
> > here... I can't imitate the deletion of the token when an article has
> > expired, but should still be remembered.
> 
> Oh!  Right, now I remember... that's what threshold was for.  If the entry
> was newer than threshold, its storage token would just be purged in the
> event of expiration.  If it was older, it would be removed from the
> history file entirely.
> 
> So everything gets passed to the callback, the callback returns true to
> keep and false to purge or something like that, and purge means either
> remove the storage API token but keep the entry or remove the entry
> entirely depending on the threshold value.  And with that, entries without
> storage API tokens could be processed based solely on the threshold value
> without ever calling the callback.
> 

OK, I think thats what I've implemented, but I've done the "process
entries without storage API tokens without invoking the callback"
approach; aside from not being able to count all the lines you process
in expire, I can't see any drawbacks?

[Apologies for not having a complete patch yet - I'm just trying to
get the legal formalities done here so that it can't rebound]

-- 
Alex Kiernan, Principal Engineer, Development, Thus PLC