Re: Roadmap for FE/BE protocol redesign

> -----Original Message-----
> From: Merlin Moncure [mailto:merlin(dot)moncure(at)rcsonline(dot)com]
> Sent: 12 March 2003 13:35
> To: Dave Page
> Cc: pgsql-hackers(at)postgresql(dot)org
> Subject: RE: [HACKERS] Roadmap for FE/BE protocol redesign
>
>
> > values. That code is just plain nasty in VB. In pgAdmin III we've
> already
> > mentioned stealing bits of the PostgreSQL parser.
>
> Do not be swayed by the dark side. In my spare time I threw
> together a
> 'proof of concept' replacement for the technology used in
> pgAdmin. It
> is written in C++. In three seconds I can query a table and
> put 100000 records in a grid. I plan to finish it and
> release it. The main reason not to use it is that it relies
> on commercial tools to build.
>
> If you or anybody else is intereted, let me know and I'll
> send it your way.

You do realise I'm the pgAdmin project lead? Anyway, the 'technology' in
pgAdmin II is ADO/ODBC which I quite agree doesn't play well with that
many records - but then, not many humans do either.

In pgAdmin III we use libpq directly from C++. If you can speed up that
then I'm sure there will be more people than just me that are interested
:-)

The code I'm actually referring too in this thread, is that which
decides whether and how the results from a given query can be updated.
That would be vastly simplified is I could easily locate the
pg_attribute row for each column.

Regards, Dave.