From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Alvaro Herrera <alvherre(at)commandprompt(dot)com> |
Cc: | Hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: fixing dllist? |
Date: | 2007-03-21 23:37:12 |
Message-ID: | 23088.1174520232@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Alvaro Herrera <alvherre(at)commandprompt(dot)com> writes:
> Hi,
> While coding the autovacuum stuff I noticed that the dllist.c doubly
> linked list infrastructure is using malloc(). And the failure cases are
> handled in #ifdef FRONTEND exit(1) #else elog(ERROR) #endif.
> This seems a bit ugly, but more importantly, it doesn't let me free the
> whole list by simply resetting a context. Would anybody be too upset if
> I wholesaledly changed malloc() to palloc() and get rid of the #ifdef
> FRONTEND? AFAICS, no frontend code in our tree uses it.
IIRC, libpq once used it to manage the pending-notify list, but that was
a long time ago. I have no objection to getting rid of the FRONTEND
support.
If you change to plain palloc you'll need to check that existing callers
call it in a suitable context --- they may be assuming that the lists
will be durable regardless of CurrentMemoryContext.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Jeremy Drake | 2007-03-22 00:09:22 | Re: patch adding new regexp functions |
Previous Message | Grzegorz Jaskiewicz | 2007-03-21 23:28:01 | Re: relation 71478240 deleted while still in use on 8.1 |