Re: [Nmh-workers] Request for sortm feature to implement arbitrary message ordering

[Ralph Corderoy]

Hi Ken,

> > By implication, the comparison program is buggy if that doesn't
> > hold?  sortm(1) punts to qsort(3) for the hard graft and that
> > demands consistency;  I think I'd like sortm to protect me from a
> > buggy comparison program.  [...]
> 
> You know what?  Forget I said anything :-)

Sorry, don't mean to stop possible small improvements by drowning out
the conversation.  :-)  We didn't even get onto threading;  `subject' is
often a crude substitute.

I think the email's number is currently used as an implicit last-resort
tie-breaker in all cases so qsort(3) becomes stable?  Should sortm(3)
state it's a stable sort?

> These ideas all sound fine; I'm kind of torn about the key-vs-cmp
> program, but I think either one would be fine.

`key' has the advantage that the user's program cannot be inconsistent
over time;  it's only asked once for any email.

I've thought a little more about it.  I'd change -kprogram to -progkey
so -k is sufficient for the common case.  The stripping of
/^(re|fw):\s*/i from the front of a field could be triggered by a key
flag.  The sequence of keys would be processed one at a time, e.g. `-k
subject -progk foo' would sort all emails on subject first and then
invoke foo as few times as necessary to only get a secondary key for the
emails that tied on subject.

Cheers, Ralph.