On Thu, Jun 2, 2016 at 9:03 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> I wrote:
>> Jeff Janes <jeff.janes@gmail.com> writes:
>>> My biggest gripe with it at the moment is that the signature size should be
>>> expressed in bits, and then internally rounded up to a multiple of 16,
>>> rather than having it be expressed in 'uint16'.
>>> If that were done it would be easier to fix the documentation to be more
>>> understandable.
>
>> +1 ... that sort of definition seems much more future-proof, too.
>> IMO it's not too late to change this. (We probably don't want to change
>> the on-disk representation of the reloptions, but we could convert from
>> bits to words in bloptions().)
>
> There were no objections to this, but also no action. Attached is a draft
> patch ... any complaints?
One thing from the commit-message:
"On-disk, we can still store it in words, so as to not break on-disk
compatibility with beta1."
Hasn't that ship already sailed?
from beta1 to HEAD:
The database cluster was initialized with CATALOG_VERSION_NO
201605051, but the server was compiled with CATALOG_VERSION_NO
201605191.
Or is the concern about intra-version pg_upgrade rather than direct
on-disk compatibility?
Cheers,
Jeff