Re: registry vs. environment (was re:binary
- From: "Magnus Hagander" <mha(at)sollentuna(dot)net>
- To: "Claudio Natoli" <claudio(dot)natoli(at)memetrics(dot)com>, "Andrew Dunstan" <andrew(at)dunslane(dot)net>, <pgsql-hackers-win32(at)postgresql(dot)org>
- Subject: Re: registry vs. environment (was re:binary
- Date: Fri, 13 Feb 2004 13:11:01 +0100
- Message-id: <6BCB9D8A16AC4241919521715F4D8BCE34B208(at)algol(dot)sollentuna(dot)se>
> > I'm not clear *which* settings it is proposed to set in the
> registry,
> > nor how possible it will be to override the settings, and
> with which
> > mechanism.
>
> I pretty much agree with Andrew here. Frankly, I think this
> whole discussion "registry vs env" is a bit pointless.
>
> What *exactly* does anyone think will need to be put in the registry?
>
> Surely, our Win32 service wrapper can just allow something like this:
>
> postgres /regserver servicename <command line args go here>
>
> So, if I want to have multiple servers registered, I could do:
> postgres /regserver MyPGServer_1 -D c:/data1
> postgres /regserver MyPGServer_2 -D c:/data2
> and so on.
Where do you plan to store this data, if not in the registry?
Can *all* env variable settings be set on the commandline as well? If
so, we can just use a single registry key to store a commandline for the
service. If not, then there need to be some way of changing them
per-instance.
> And of course, you can always run another instance from the
> command line, or have a single instance registered that takes
> all its args from the environment, or ...
Yes, certainly.
//Magnus
Home |
Main Index |
Thread Index