Subject: Re: Updating databases/libpqxx
To: None <firstname.lastname@example.org>
From: Jan Danielsson <email@example.com>
Date: 05/23/2007 02:01:50
-----BEGIN PGP SIGNED MESSAGE-----
Joerg Sonnenberger wrote:
> On Wed, May 23, 2007 at 01:32:52AM +0200, Jan Danielsson wrote:
>> "make" works, but "make install" fails horribly because all the
>> "install" commands for the headers fails. I assume the fix is easy --
>> but what is it?
> Check that the directories it wants to install to actually exist. But it
> is hard to tell if you don't show us the error.
Yes, the nonexistence of the target directory is the problem. I'm
trying to track down when the "install" commands are executed. Digging
through build systems is not something I've done much before.
I'm having some trouble understanding where the line is drawn between
pkgsrc and the packages own build system. I understand that pkgsrc does
some preparations, then modifies the package's build system as required,
then kicks it off. Once it's done, it returns to pkgsrc for installation
or ever build target has been chosen. But how does it know which actions
to take? Does it always assume that it should run:
..unless anything else has been specified?
Are the files/directories (apart from the unpackaged source
directory) directly under work/ all a part of pkgsrc?
>> Also -- what is PLIST actually used for? Is it only used for
>> pkg_delete and checking an installed package's integrity?
> It determines the content of the package. It is used to create binary
> packages, for pkg_delete and integrity checks.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (NetBSD)
-----END PGP SIGNATURE-----