Re: [svn:DBD-Pg] r9728 - in DBD-Pg/trunk: . t

-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160


David Wheeler wrote:
> It looks like you include the docs for async queries and savepoints
> about three times.

Heh - but it passed Test::Pod, honest! :) Fixed, thanks.

> Shouldn't we have the async constants start with "PG_" like all the
> other constants, rather than "DBDPG_"? I realize that they're
> different than the sorts of constants directly supported by  
> PostgreSQL, but surely we can control for that and keep things consistent.

Yeah, that's a good point, we're using pg_ for the methods after all.

> Do we need "OLDQUERY" in the cancel and wait constants? Seems kind
> of unnecessary to me.

Not sure - just a plain "CANCEL" or "WAIT" might be a little confusing, no?
These are kind of syntactic sugar left over from earlier development anyway, 
as the (I suppose) canonical way to do it will be to cancel or wait yourself 
with pg_cancel and pg_result.

> I really like the API. Thanks for doing this!

Glad you like it, it's been a struggle to fit into the DBI way of doing 
things. I'd really like it if execute() took attribs, so we could do 
an async per execute, not prepare. Other possible options include a custom 
execute method (pg_execute) or flipping a global attrib before the 
execute ($dbh->{pg_async}). I'm open to any and all API changes and suggestions.

- --
Greg Sabino Mullane greg-738XdyZ4GzZWk0Htik3J/[email protected]
End Point Corporation
PGP Key: 0x14964AC8 200707121400
http://biglumber.com/x/web?pk=2529DF6AB8F79407E94445B4BC9B906714964AC8
-----BEGIN PGP SIGNATURE-----

iD8DBQFGlm73vJuQZxSWSsgRA1a4AJ4n/4Mbg1MPrjGNPMk/1ME1LiZACwCeK3MD
klf3dZ9VyQVm8hCjG0tqpgo=
=mOfl
-----END PGP SIGNATURE-----