Re: [HACKERS] Our feature change policy - Mailing list pgsql-hackers

From David G. Johnston
Subject Re: [HACKERS] Our feature change policy
Date
Msg-id CAKFQuwavVzmn2_760dXErH0rXNk297ZTcyLw6e4D7PkACgiC7g@mail.gmail.com
Whole thread Raw
In response to Re: [HACKERS] Our feature change policy  (Stephen Frost <sfrost@snowman.net>)
List pgsql-hackers
On Mon, Mar 20, 2017 at 8:57 AM, Stephen Frost <sfrost@snowman.net> wrote:
Tom,

* Tom Lane (tgl@sss.pgh.pa.us) wrote:
> Stephen Frost <sfrost@snowman.net> writes:
> > * Bruce Momjian (bruce@momjian.us) wrote:
> >> 1.  make the change now and mention it in the release notes
> >> 2.  #1, but also provide backward compatibility for 5+ years
> >> 3.  mark the feature as deprecated and remove/change it in 5+ years
> >> 4.  #3, but issue a warning for deprecated usage
>
 
I was imagining the changes to pg_stat_activity as possibly falling
under #3/change, meaning that we'd wait for 5 years before actually
renaming those columns, which is part of why I was objecting to that
idea.

Which ends up boiling down to:

A. Make a change and document it in the release notes

B. If applicable, and desired, provide a 5 year backward​ compatibility deprecation period (i.e., 3 =>[implies] 2). Generally 2 => 3 but the deprecation period is undefined.

C. Optionally, if deprecating, provide explicit warnings when the deprecated feature is used

Guidelines for when to desire the 5 year period would be helpful.  And also acknowledge that we may wish to choose a shorter period of time, or institute immediate removal, at our discretion.

Nothing says we cannot go longer than 5 years but given our support policy I would say that we'd rarely desired to do so intentionally - the burden of proof falling to those who would want to keep something longer.

David J.

pgsql-hackers by date:

Previous
From: Bruce Momjian
Date:
Subject: Re: [HACKERS] Our feature change policy
Next
From: Robert Haas
Date:
Subject: Re: [HACKERS] Partition-wise join for join between (declaratively)partitioned tables