Re: libpq object hooks

Tom Lane wrote:
> Andrew Chernow <ac@esilo.com> writes:
>> 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.
>
> What?  I thought what you wanted back was the void * pointer that had
> been registered with a particular callback function.  So you use that
> callback function.  If it's not actually registered, you get a NULL.
>
>> This is what is passed to PQaddObjectHooks, along with a conn:
>
> This is all wrong IMHO, not least because it creates ABI problems if you
> want to add another hook type later.  Register each hook separately, eg
>
> typedef void (*PGCRHook) (PGconn *conn, void *passthrough);
>
> void PQregisterConnResetHook(PGconn *conn, PQCRHook func, void *passthrough);
>
> ... repeat for each possible hook ...
>
>             regards, tom lane
>
>

One can make a case to break apart the obj hooks structure into
individual register functions, but I think you have a different idea in
your head than what is being proposed.  For starters, there is no
passthru pointer to register with a callback (there could be but that is
different than hook data...your register looks more like a user_ptr).
The passthru pointer, what we call hookData, is allocated with a PGconn
(not provided by the user).  This is the point of the initHookData callback.

typedef void *(*PGinitHookData)(const PGconn *conn);

PQregisterInitHookData((PGconn *)NULL, (PGinitHookData)func);
PQregisterConnResetHook((PGconn *)NULL, (PGCRHook)func);
//etc...
conn = PQconnectdb();

When connectdb returns, initHookData has already been called.  So, a
call to PQhookData(conn, ????) will work.  BUT, what is still missing
from the equation is how to uniquely reference hookData on a conn.

What I was previously suggesting was to use the address of initHookData,
since w/o this address there wouldn't be any hook data to get.  Seemed
like a logical choice.

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