> The few postings I have noticed with users running 7.4 has been with a
> release several less than the newest. ...
> Supporting old versions is a great and noble thing but there comes a
> time when it is a waste of resources because the effort goes unused.
Yeah, that's a really good point. An example is that Red Hat is still
shipping/supporting 7.4.x in RHEL 4, but it's been quite a long time
since I've been able to persuade them to push a 7.4.x update that didn't
involve a security issue. (They're currently shipping 7.4.19, and I'm
not even going to bother suggesting an update to .22.)
Probably everyone has got their own slightly different set of hot-button
considerations for whether it's worth updating to a new minor release,
but it's really unclear that there's going to be any uptake at all for
7.4.22 as constituted, because the bugs it fixes are so minor.
The suggestion I started this thread with amounted to not bothering with
pushing 7.4.x updates in update cycles where we'd made no "serious" bug
fixes in it; which is a very long way from desupport. Maybe an
appropriate compromise is to announce now that 7.4 is in maintenance
mode and will receive only really critical bug fixes (which are the only
ones that 7.4.x users are going to pay attention to anyway, so nothing
is lost); and that actual desupport will occur a year from now.
regards, tom lane