<<< Date Index >>>     <<< Thread Index >>>

Re: hcache broken (slightly)



Hello,

> Okay, I prefer whitelisting while you prefer blacklisting. Preparing a
> patch in either flavour isn't difficult, I guess.

I don't prefer black over whitelisting or vice versa. I prefer
readiblity and ease of maintenance over duplicate code and bad habits.

> But: things will almost certainly break either way in the beginning as
> we need to find out which members we don't want first.

Ideally we want all members. Or the members don't belong in the struct
HEADER but somewhere else.

> If we were to extend the hcache in a way that enables it storing
> updated messages, we also need to ignore things like:

>   header->offset
>   header->index
>   header->score

> and not just pointers as these depend on the runtime environment
> (including user config).

Not necessary. But we have to invalidate them at the right moment.

> I guess it'll take some time until we really will have found out which
> members need to be ignored to not fool mutt upon restoring them.

Do that. Because I will not take an active role in it. I said what I
think, but I won't provide any code or play an active role in this. So
it is the hands who get active: In yours.

        Thomas