Re: WIP: a way forward on bootstrap data

Robert Haas wrote:
> On Thu, Dec 21, 2017 at 5:32 PM, Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org> wrote:
> > Hmm, patch 0008 removes data lines from the .h but leaves the dependent
> > OID define lines around:
>
> Just a question here -- do we actually have consensus on doing the
> stuff that these patches do? Because I'm not sure we do.

My reading of the old threads (linked provided by John in his initial
email in this thread) is that we have a consensus that we want to get
rid of the old data representation, because it causes endless amount of
pain. The proposed format seems to satisfy the constraints that we all
discussed, namely

1. be easier to modify than the current format,
2. in particular, allow for default values on certain columns
3. not cause git merge problems because of too similar lines in every
record
4. not require onerous Perl modules

The one thing we seem to lack is a tool to edit the data files, as you
suggested[1]. Stephen Frost mentioned[2] that we could do this by
allowing the .data files be loaded in a database table, have the changes
made via SQL, then have a way to create an updated .data file. Tom
said[3] he didn't like that particular choice.

So we already have Catalog.pm that (after these patches) knows how to
load .data files; we could use that as a basis to enable easy oneliners
to do whatever editing is needed.

Also, the CPP symbols remaining in the pg_*.h that I commented yesterday
was already mentioned[4] before.

[1] https://www.postgresql.org/message-id/CA%2BTgmoa4%3D5oz7wSMcLNLh8h6cXzPoxxNJKckkdSQA%2BzpUG0%2B0A%40mail.gmail.com
[2] https://www.postgresql.org/message-id/20150304150712.GV29780%40tamriel.snowman.net
[3] https://www.postgresql.org/message-id/24766.1478821303%40sss.pgh.pa.us
[4] https://www.postgresql.org/message-id/15697.1479161432@sss.pgh.pa.us

--
Álvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services