On Fri, Aug 05, 2016 at 02:56:40PM +0200, Simon Ruderich wrote:
> I think it's a good argument in this case because I expect to
> always get a copy of all sent mails. And if I store it before I
> send it (like it's now) this is true, otherwise it isn't. 

False; it is POSSIBLE that it may not be true--but if Mutt had the
alternatives we've discussed in this thread, it is unbelievably
unlikely.  It can only happen if BOTH the message fails to send, for
reasons unrelated to a system failure, AND, *that* failure is
immediately followed by a system failure that prevents you from
reacting to the first failure.

The odds of this are astronomical.  The options Vincent and I have
suggested are both more flexible than what Mutt currently does, more
accurately reflect the state of the message, and practically speaking
also sufficient to prevent you from ever losing mail in this fashion
in your lifetime, probablistically speaking.

You're making a mountain out of an ant hill.  Is there a type of hill
that's smaller than an ant hill?  Then that.

> >> (mutt stores mails in /tmp which is either a tmpfs or emptied on
> >> reboot).
> >
> > False:  [snip]
> 
> I was talking about the defaults. 

So was I, not that it particularly matters.  None of my systems behave
the way you describe, and as far as I'm aware I haven't done anything
to make that so.  But if your system does behave that way, and you
don't want that behavior, then you have to think about it, plain and
simple.  

You may not want to think about it, but you can't always get what you
want, and Mutt does not control how your system gets installed or how
YOU choose to configure it (or don't).  You are the system
administrator of your system.  You are responsible for how it is
configured.  Being uninterested in configuring your system properly
does not excuse you from the responsibility.

But this is only relevant to one of about a half-dozen alternatives
for dealing with this problem.

> >> I admit the current situation is not perfect, but at least I
> >> won't lose the message no matter what happens once I've pressed
> >> send.
> >
> > False. [snip]
> 
> That was not my point. My point was that I'll always have a copy
> of the mail _if_ the mail is sent. Of course the computer can
> crash at any time, but if it does either I have a copy and the
> mail was or was not sent, or the mail wasn't sent.

It doesn't matter if it was your point, you're wrong.  If your system
crashes you have no guarantee that ANYTHING will be on your disk when
you reboot it, least of all an e-mail message that got written to
BUFFER CACHE right before your system died--it may not even have ever
been committed to disk.

Suffice it to say, data loss is data loss.

> > If your computer does crash, and the message got sent, you can just
> > ask the recipient to forward it to you.
> 
> As was pointed out in another thread, that's only true if the
> mail was correctly delivered. 

In other words, it wasn't sent? :)  So in that case the message should
NOT have been recorded.

> And still if it was, that's not the situation I'd like to be in when
> using mutt.

It's a worst case scenario, in an extremely unlikely combination of
events.  You obviously don't want to be in that situation, but if
losing a single mail message under this perfect storm is the worst
thing that happens to you that day, you should consider yourself very
lucky.  

Most users will assume that if the message got recorded, it got sent.
If they go looking for a message their boss said he was waiting for,
and don't find it in the sent folder, then there will be no confusion
that it didn't get sent--or at least that the sending process, in
whole, failed.

-- 
Derek D. Martin    http://www.pizzashack.org/   GPG Key ID: 0xDFBEAD02
-=-=-=-=-
This message is posted from an invalid address.  Replying to it will result in
undeliverable mail due to spam prevention.  Sorry for the inconvenience.

Attachment: pgpDbO1OQSWsW.pgp
Description: PGP signature

Reply via email to