Re: WIP: a way forward on bootstrap data

On 12/13/17, Peter Eisentraut <peter(dot)eisentraut(at)2ndquadrant(dot)com> wrote:
> On 12/13/17 04:06, John Naylor wrote:
>> There doesn't seem to be any interest in bootstrap data at the moment,
>> but rather than give up just yet, I've added a couple features to make
>> a data migration more compelling:
>
> I took a brief look at your patches, and there appear to be a number of
> good cleanups in there at least. But could you please send patches in
> git format-patch format with commit messages, so we don't have to guess
> what each patch does?

Thanks for taking a look and for pointing me to git format-patch.
That's much nicer than trying to keep emails straight. I've attached a
new patchset.

Note that 4-7 and 9-10 are units as far as the build is concerned.
Meaning, once 4 is applied, the build is broken until 7 is applied.
Also, postgres.bki won't diff 100% clean with the master branch
because of some useless quotes in the latter.

One thing that occured to me while looking over patch 0004 again: It's
now a bit uglier to handle indexing.h and toasting.h. I think it might
be cleaner to keep those statements in the header of the catalog they
refer to. That has the additional benefit of making the headers the
Single Point of Truth for a catalog schema.

TODO:
-Docs and README
-Finish SQL generation patch
-Consider generating pg_type #defines

-John Naylor