From: | "David E(dot) Wheeler" <david(at)kineticode(dot)com> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | Robert Haas <robertmhaas(at)gmail(dot)com>, Dimitri Fontaine <dimitri(at)2ndquadrant(dot)fr>, pgsql-hackers Hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: Quick Extensions Question |
Date: | 2011-03-03 18:57:21 |
Message-ID: | 901B2744-A3A7-4125-A429-C9E681F6632C@kineticode.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Mar 3, 2011, at 10:54 AM, Tom Lane wrote:
>> Which is why my suggestion is pretty much free from any design. Just a list of dependencies, with only a server version number. No other syntax at all. It can be added later.
>
> I basically agree with Robert that "requires = 9.1" is entirely useless.
> There's next to no scenario where an extension author wouldn't really
> be wanting to write "requires >= 9.1" instead. And to do that, we have
> to solve the whole version-number-comparison problem that we worked so
> hard to dodge before. So this all looks to me like something that needs
> considerably more thought than we can devote to it for 9.1.
I'm saying that
core_requires = 9.1.0, libxml, plpgsql
Means >= 9.1.0. That = is an assignment operator, not comparison. And this is the *only* version number I'd specify, the core version number, of which the core has perfect control of how things are compared (pg_version()).
Best,
David
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2011-03-03 19:09:39 | Re: Quick Extensions Question |
Previous Message | Tom Lane | 2011-03-03 18:54:05 | Re: Quick Extensions Question |