Re: Extensions, this time with a patch

Tom Lane <tgl@sss.pgh.pa.us>

From: Tom Lane <tgl@sss.pgh.pa.us>
To: Dimitri Fontaine <dimitri@2ndQuadrant.fr>
Cc: Itagaki Takahiro <itagaki.takahiro@gmail.com>, Alvaro Herrera <alvherre@commandprompt.com>, PostgreSQL Hackers <pgsql-hackers@postgresql.org>
Date: 2010-10-20T15:21:56Z
Lists: pgsql-hackers
Dimitri Fontaine <dimitri@2ndQuadrant.fr> writes:
> Tom Lane <tgl@sss.pgh.pa.us> writes:
>> I don't think that "no changes to the makefiles" is a requirement,
>> or even a wish-list item, for this.  I think it's perfectly reasonable
>> for the makefile to have to specify the module name; far better that
>> than that we get the name by some "magic" or other.

> It seemed easy to get a reasonable approach requiring very few edits in
> contribs so I favoured that. Now, it's still entirely possible to hand
> adjust. Determining the extension name automatically from DATA_built or
> DATA is only done where EXTENSION has not been provided,

That is simply a horrid idea.  Just make it specify EXTENSION.

> and guessing
> the CONTROL file name from the EXTENSION name only occurs when CONTROL
> has not been provided.

Here, on the other hand, I'm wondering why have two variables at all.
Is there any sane use-case for the control file to not be named the same
as the extension?  It seems like that would accomplish little except to
sow confusion.

			regards, tom lane