Re: libpq object hooks

From: Andrew Chernow <ac(at)esilo(dot)com>
To: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: Merlin Moncure <mmoncure(at)gmail(dot)com>, Andrew Dunstan <andrew(at)dunslane(dot)net>, Alvaro Herrera <alvherre(at)commandprompt(dot)com>, "Patches (PostgreSQL)" <pgsql-patches(at)postgresql(dot)org>
Subject: Re: libpq object hooks
Date: 2008-05-15 18:36:42
Message-ID: 482C82BA.3090100@esilo.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers pgsql-patches

Tom Lane wrote:
> Andrew Chernow <ac(at)esilo(dot)com> writes:
>> There can be cases to use the same callbacks, although unlikely. To
>> completely avoid collisions, the below would work:
>
> Still looks like overdesign to me. If we use the hook function address
> we solve the problem with no extra notation and no extra storage.
>
> Note that if you want N fixed keys, you can just have N hook functions
> (all calling a shared workhorse routine, no doubt). So your proposal
> adds no functionality whatever if the usage involves a fixed number of
> static handles. Now it could possibly allow a variable-at-runtime
> number of handles, but I'd want to see a worked-out use case before
> designing for that much complexity. In particular, it seems to me that
> the problem would then shift to how do you know which handle to use for
> the lookup, thus you've just introduced another layer of complexity
> without buying anything.
>
> I think the typical use case is just that you need to distinguish "your"
> hook from anyone else's hooks, so the function address is plenty
> sufficient.
>
> It should also be noted that this whole problem *can* be solved without
> any PQhookData at all: as long as you have hooks to get control at
> creation and destruction of PGconns and PGresults, you can maintain your
> own index data structure. I'm willing to grant some amount of extra
> API-stuff to save users having to do that in simple cases, but I don't
> think we need to try to support infinitely complex cases.
>
> regards, tom lane
>
>

Okay. No problem over here.

Which callback do we use as the key? Currently, none are required (only
the name was required). We have to choose one callback that must be
provided. Maybe initHookData() must be provided? If the end-user
doesn't need it they can just return NULL.

This is what is passed to PQaddObjectHooks, along with a conn:

typedef struct
{
//char *name; REMOVED

void *data;

/* Invoked when PQsetObjectHook is called. The pointer returned
* by the hook implementation is stored in the private storage of
* the PGconn, accessible via PQhookData(PGconn*). If no
* storage is needed, return NULL or set this hook to NULL.
*/
void *(*initHookData)(const PGconn *conn);

/* Invoked on PQreset and PQresetPoll */
void (*connReset)(const PGconn *conn);

/* Invoked on PQfinish. */
void (*connDestroy)(const PGconn *conn);

/* Invoked on PQgetResult, internally called by all exec funcs */
void *(*resultCreate)(const PGconn *conn, const PGresult *result);

/* Invoked on PQcopyResult */
void *(*resultCopy)(PGresult *dest, const PGresult *src);

/* Invoked when PQclear is called */
void (*resultDestroy)(const PGresult *result);
} PGobjectHooks;

--
Andrew Chernow
eSilo, LLC
every bit counts
http://www.esilo.com/

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Tom Lane 2008-05-15 18:44:14 Re: libpq object hooks
Previous Message Tom Lane 2008-05-15 18:19:16 Re: libpq object hooks

Browse pgsql-patches by date

  From Date Subject
Next Message Tom Lane 2008-05-15 18:37:14 Re: Patch to change psql default banner v6
Previous Message Bruce Momjian 2008-05-15 18:31:32 Re: Patch to change psql default banner v6