Dimitri Fontaine <dimitri(at)2ndQuadrant(dot)fr> writes:
> This and removing module_pathname in the control files to just use
> $libdir/contrib in the .sql files. That would set a better example to
> people who want to make their own extensions, as the general case is
> that those will not get into contrib.
I'm not sure it's a better example. We considered doing that before,
and decided not to on the grounds that using MODULE_PATHNAME avoids
hard-wiring the name of the shared library into the SQL scripts. Also,
IIRC, there are a couple of contrib modules where there's an actual
problem in doing it like that; though I'm unable to recall details as
I'm fighting a head cold at the moment.
My original intention when I started working with the extensions patch
had in fact been to do what you suggest above, but I was convinced not
to. I don't think we should reverse that decision at the last minute.
regards, tom lane
In response to
pgsql-hackers by date
|Next:||From: Tom Lane||Date: 2011-04-02 19:12:32|
|Subject: Re: 9.0.3 SIGFAULT on FreeBSD with dtrace |
|Previous:||From: Joseph Adams||Date: 2011-04-02 18:56:09|
|Subject: Re: psql 9.1 alpha5: connection pointer is NULL|