Herouth Maoz wrote:
> However, your way would be much more logical, though I suspect a
> performance degredation if all the connections are done in the interpreted
> language itself.
I don't think so.
SQL conversation mean some small messages for establishing protocol then
sending the SQL command and after that receiving data. The most part of
the time is taken by data transferring over the network. I would bet on
a maximum 5% performance degrading.
Tcl/Tk 8.0 would be the only version that could deal with binary i/o
over files and sockets.
In my opinion, it wouldn't be so hard to rewrite pg interface directly
I would study that posibility and keep you informed.
FLEX Consulting Braila, ROMANIA
In response to
pgsql-interfaces by date
|Next:||From: Constantin Teodorescu||Date: 1998-01-13 12:56:07|
|Subject: Re: [INTERFACES] Printing from PgAccess|
|Previous:||From: Goran Thyni||Date: 1998-01-13 10:32:10|
|Subject: Re: [INTERFACES] DBD-pg error in current source tree|