If messages end up queued in Outbox, they are delivered in the wrong order. Always.
Could you even say they're "queued" in that case? Or would the right term be something liked "piled"? "Stuffed" maybe?
Stacked. They are presumably being stacked. That's the most likely arrangement that 1. Someone might actually build 2. You would consistently notice getting it wrong even for just two emails.
Or the emails are assigned some id based on a hash (time+$$? header? content?) when being temporarily saved. Then delivered in that sorted hash order.
Or instead of guessing you could just, you know, try it. They are always delivered backwards and exactly, precisely wrong: newest first.
This is actually the first time I've ever heard someone claim the outbox is a queue. IAC, it's probably delivering on multiple threads to speed up delivery.
Since the progress bar clearly says "delivering 1 of 3"... no.
Since an account can only have one outbound mail server... no.
I'm pretty sure smtpd can handle simultaneous connections, even from a single client. Though I'm also pretty sure that (speaking only of postfix) the queue is cleared by a single process.
...I add, contributing nothing...
Though, are they in an order?
Unlike most, or at least many clients, Mail.app operates based on the timestamp of the file on the filesystem, and seems to ignore anything in the header. For instance, I moved a bunch of files between IMAP servers, and they got a new FS datestamp, Mail.app is the only client I use which saw them as all having been delivered at the time of the copy. Makes sorting suck until you fix the timestamps of the files on the IMAP server.
Maybe it's something to do with that?
I like the idea that there's some kind of twisted desert-island logic to this, because BUT IT'S A PHONE and if you're one of those people who drives off into the mountains in winter with nothing but socks to eat maybe getting your last message through (THE BEAR IS ABOUT TO EAT ME) first before the battery dies is going to be important.
Except they could just change the rule when the battery is below 10%. But that wouldn't account for loss-of-service, I guess.