[postgis-devel] libpq 7.2 support

[Ron Mayer]

Markus Schaber wrote:
> Jim C. Nasby wrote:
>>> Do we still want to support building against 7.2? Or is it time to
>>> follow postgresql.org who don't support 7.2 any more, and are very
>>> likely to drop support for 7.3 shortly after 8.2 is released, which is
>>> in Beta now.
>> Actually, Tom Lane is on the hook to support 7.3 until Red Hat EOL's it,
>> which IIRC will be 5 years from when it was released (well, actually, 5
>> years from when the last Red Hat version with 7.3 was released). I think
>> it's got another year or two.
> 
> Thanks for the info.
> 
> Then we should support 7.3 at least as long, too.

IMHO there are two separate questions.

(1) For which versions is free community support available.

    For that, I think a fine answer to be "whatever is most
    convenient for the developer community" so they can stay
    focused on moving the project forward.  If this happens
    to include a lot of very old releases, great.  But I don't
    think it should be a rule cast in stone.

and

(2) For which versions is there commercial support available.

    Corporate users would like it if some support vendor
    would clearly state that they'll (for pay, of course)
    provide support for postgresql versions matching those on the
    major commercially supported Linux distros (Red Hat,
    SuSE, and remembering that Debian is supported by HP, etc).

    Perhaps Refractions could state this, and I wonder if any
    of the other postgresql support organizations like
    Fujitsu support PostGIS.

It may be that since the developers are likely to be the
same guys that do commercial support the answers might
be the same - but even so there's a subtle difference.

If a CIO were to ask "how long is X supported", the answer
of "what versions does can a support contract I can buy cover"
may be more important than "what did the community say
on a mailing list".   OTOH, if I'm a developer, a mailing
list comment like "looks like postgresql version X is
getting hard to maintain because Y so support will be
dropped in Postgresql release Z" is mostly fine because
it gives me ample warning that I should stay with version
Z-1 if I don't want to upgrade both.