Ken Hornstein <kenh(_at_)pobox(_dot_)com> writes:
But then you say (in another message) that you want nmh programs to not
deadlock under our hypothetical nmhlock program
If I said something that amounted to that, it's not what I meant. I don't know
what I might have said that led you to believe that's what I meant. Which is
to say that, in my confusion, I said nothing that amounted to that.
I guess I was thinking that based on this message:
I mean, why would you want nmh programs to not deadlock under mhlock unless
you wanted to run them? That's where things get tricky.
Yep. I said that. But I no longer know why, if I ever did know.
I also admit that I don't understand why nmh locking has to be so complicated,
but I leave that issue to my betters.
Well, it's just as complicated as it needs to be. The issue is that the
sequences files and context files get modified a lot, so those need to
be locked. And (generally) sequence files are per-folder. So I'm not
sure how you could make locking work and not make any simpler than
it is now (unless you wanted to do the equivalent of a biglock; we don't
want that, do we?).
I don't know why we don't want that. But there is no reason why I need to know.
At this point I retire from this locking discussion with the hope that I don't
gag on all the feet in my mouth.
Nmh-workers mailing list