Re: Unsafe coding in ReorderBufferCommit()

Andres Freund <andres(at)2ndquadrant(dot)com> writes:
> On 2015-01-23 16:47:30 -0500, Tom Lane wrote:
>> There are at least two bugs in reorderbuffer.c's ReorderBufferCommit():

> Thanks for fixing these!

> Unfortunately there's more - we'll currently do bad things if
> transaction commit fails. At the very least the (sub-)transaction begin
> commands need to be moved out of the exception block as they can
> fail... :(. E.g. because this is the 2^32-1 subxact or similar...

> I actually also want to strip the CATCH block of most of it's contents -
> there's really no need anymore for most of what it does.

No objection here. I was just doing a mechanical transform of the
function, not based on any deep understanding of what it does.

The less you need to do in a CATCH block, the better.

regards, tom lane