Andrew Dunstan <andrew@dunslane.net> writes:
> On 04/29/2014 02:56 AM, Heikki Linnakangas wrote:
>> On 04/28/2014 10:32 PM, Tom Lane wrote:
>>> Meh. I still think it's a bad idea to have CATALOG_VERSION_NO getting
>>> compiled into libpgcommon.a, where there will be no way to cross-check
>>> that it matches anything. But I guess I'm losing this argument.
>> FWIW, I agree it's a bad idea. I just have no better ideas (and
>> haven't given it much thought anyway).
> Sure sounds like a bad idea.
One idea would be to have relpath.h/.c expose a function (not a #define)
that returns the value of CATALOG_VERSION_NO compiled into it. Then,
if pg_rewind were to replace all its direct references to
CATALOG_VERSION_NO (including the value it checks against pg_control)
with calls of that function, consistency would be ensured.
A notational problem is that if pg_rewind or similar program is directly
using the TABLESPACE_VERSION_DIRECTORY macro anywhere, this wouldn't
work. But we could perhaps expose a function to return that string too.
regards, tom lane