Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> What was the rationale for changing the assignments of existing
> 2PC IDs? So far as I can tell, that breaks pg_upgrade (if there
> are any open prepared transactions) for no redeeming social
> benefit. Is there a reason why TWOPHASE_RM_PREDICATELOCK_ID has
> to be 2 and not at the end?
I'm sure that Dan will jump in if this guess is wrong, but since the
predicate lock code is largely derived from the heavyweight locking
code, it probably seemed to have a minor cosmetic benefit to put it
adjacent to that. It didn't occur to me when the SSI 2PC code went
in, but I can see the problem now that you point it out. The new
entry should be moved to the end for compatibility. Would you like
me to submit a patch to fix this, or should I stay out of your way?
-Kevin