Re: Frustrating issue with PGXS - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Frustrating issue with PGXS
Date
Msg-id 15995.1182953405@sss.pgh.pa.us
Whole thread Raw
In response to Re: Frustrating issue with PGXS  (Fabien COELHO <coelho@cri.ensmp.fr>)
List pgsql-hackers
Fabien COELHO <coelho@cri.ensmp.fr> writes:
>> What was the problem with just making all uses of pg_config in 
>> Makefile.global use a hardcoded bindir directly?

> Because bindir is given by pg_config:-)

> ISTM that the underlying issue, which was not foreseen in the initial pgxs 
> and fixed later, is that some distributions use a different installation 
> prefix at compile time and once the software is actually installed.

Right, the installation tree is supposed to be relocatable.  Otherwise
we would not need to use pg_config to find the paths at all; we'd just
hardwire all of them when constructing Makefile.global.
        regards, tom lane


pgsql-hackers by date:

Previous
From: Michael Enke
Date:
Subject: self defined data type "with limit"?
Next
From: Tom Lane
Date:
Subject: Re: Frustrating issue with PGXS