Re: WAL dirty-buffer management bug

[Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>]

Simon Riggs <simon(at)2ndquadrant(dot)com> writes:
> Problem for indexes only. heap xlrecs don't specify exact insert points

Sure they do.  They had better, else (for example) the associated index
insertions will be wrong.

> Accesses to local buffers don't need to be critical sections either.

True, but in most places it would uglify the code quite a bit to make it
like that, because START/END_CRIT_SECTION are bracketing code that is
shared between both cases.  And I'm not seeing where we'd get any
particular reliability gain from it: if the code has any significant
risk of elog'ing during the critical section, it's broken anyway ...

			regards, tom lane