| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Hitoshi Harada <umi(dot)tanuki(at)gmail(dot)com>, Greg Stark <gsstark(at)mit(dot)edu>, pgsql-hackers(at)postgresql(dot)org, Chris Spotts <rfusca(at)gmail(dot)com> |
| Subject: | Re: Aggregate-function space leakage |
| Date: | 2009-07-23 19:34:05 |
| Message-ID: | 25309.1248377645@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
I wrote:
> Anyway, I'll go take a look at exactly what would be involved in the
> first choice.
Actually, it seems this way results in a net *savings* of code, because
we can simply remove the code that was responsible for retail pfree'ing
of the transition values. I suppose that code must have predated the
introduction of MemoryContextReset, or it would have occurred to us to
do it like this to begin with.
I think that WindowAgg does not need any changes because it already does
MemoryContextResetAndDeleteChildren(winstate->wincontext) at partition
boundaries. Hitoshi, do you agree?
regards, tom lane
| Attachment | Content-Type | Size |
|---|---|---|
| unknown_filename | text/plain | 4.0 KB |
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Jaime Casanova | 2009-07-23 20:47:24 | Re: Determining client_encoding from client locale |
| Previous Message | Emanuel Calvo Franco | 2009-07-23 19:10:28 | uuid contrib don't compile in OpenSolaris |