Re: [Nmh-workers] Locking (specifically, sequences)

[chad]

On 09 Mar 2013, at 08:57, Ken Hornstein <address@hidden> wrote:

>> That sounds like a problem to me.  Could an inc have to
>> wait a half hour for a refile/folder -pack/or any of a
>> large number of other operations to finish?
> 
> That's what I was trying to explain.  In the inc case, the actual fiddling
> time is small.  Re-reading the sequence file after you're done incorporating
> the messages solves that problem.  I cannot envision a scenario where
> that does not solve the problem, but if you can come up with one I'd
> gladly rethink that.
> 
> Refile generally doesn't take a huge amount of time, if it can do hard links.
> folder -pack is the big one, but when I looked at that there are actually
> a bunch of things that does that really expect no one is doing anything
> else to the folder during that time period, so I would argue that any other
> operations should block during that time.

Spitballing for a moment: I wonder if we can detect the case where
the sequence file changes out from under an operation.  If so, it
might be possible to save the original data, compute a delta of
some sort, and use that to repair the sequence, or punt to the user?

I used to see this problem (rarely) with a periodic fetchmail+slocal
setup, and 'folder -pack' was almost always enough to make the world
right. On the other end, things like rsync, hg, and git have advanced
the state of the art in version-skew management a fair bit since
those days.

Just an idea; I hope it helps.
~Chad